阅读说明：本技术 一种对焦方法及系统 (Focusing method and system ) 是由 张珺 梅春明 狄媛媛 于 2020-06-24 设计创作，主要内容包括：本发明属于对焦技术领域,涉及一种对焦方法及系统。将图像采集系统采集到的样品图像进行图像清晰度评价以及目标内容含量的评价；将高于阈值的目标内容含量评价得分对应的图像视野继续评价；以找到图像清晰度评价曲线的峰为目的,根据图像清晰度曲线的走向趋势确定下一个采样位移,重复图像采集和评价,直至找到各视野下图像对焦曲线的峰,根据峰对应的采样位移确定各视野下当前对焦精度下的对焦位置。该对焦方法能够在确保较大的搜索范围、较高的对焦准确度和对焦位置精确度的情况下迅速对焦。(The invention belongs to the technical field of focusing, and relates to a focusing method and a focusing system. Evaluating the image definition and the content of target content of a sample image acquired by an image acquisition system; continuously evaluating the image view field corresponding to the target content evaluation score higher than the threshold value; and determining the next sampling displacement according to the trend of the image definition curve, repeating the image acquisition and evaluation until the peak of the image focusing curve under each visual field is found, and determining the focusing position under the current focusing precision under each visual field according to the sampling displacement corresponding to the peak. The focusing method can quickly focus under the condition of ensuring larger search range, higher focusing accuracy and focusing position accuracy.)

1. A focusing method is characterized by comprising the following steps:

(1) evaluating the image definition and the content of target content in the image of sample images collected by an image collection system under different visual fields and at different sampling displacement positions to obtain the image definition and the content evaluation score of the target content in each image;

(2) comparing the maximum value of the image target content evaluation scores of the sample images at different sampling displacements in the same visual field with a given threshold value, reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation, and executing the step (3);

(3) determining the next sampling displacement according to the trend of the image definition evaluation curve for the purpose of finding the peak of the image definition evaluation curve under the target focusing precision for any visual field which is continuously evaluated, wherein the image definition evaluation curve is a curve formed by a visual field downsampling displacement data set and a definition evaluation score data set of an image corresponding to each sampling displacement;

(4) acquiring an image of the visual field and evaluating the image definition according to the determined next sampling displacement by the method in the step (1), supplementing the acquired sampling displacement data and the acquired image definition evaluation score data to the sampling displacement data set and the evaluation score data set, and determining the next sampling displacement according to a new curve trend;

(5) and (5) repeating the step (4) until a peak of the visual field definition evaluation curve under the target focusing precision is found, and determining the focusing position under the focusing precision under the visual field according to the sampling displacement corresponding to the peak.

2. The focusing method according to claim 1, wherein the image sharpness of the image is evaluated and the content of the object content in the image is evaluated separately by an evaluation method;

the evaluation method for evaluating the image definition has the advantages that the higher the evaluation score is, the clearer the image is;

the evaluation method for evaluating the content of the target content in the image has the advantage that the higher the evaluation score is, the higher the content of the target content in the image is.

3. The focusing method according to claim 1, wherein the same image has one or more image sharpness evaluation scores.

4. The focusing method according to claim 1, wherein the step (3) determines the next sampling displacement according to the trend of the image sharpness evaluation curve, specifically:

with the sampling displacement of image is the abscissa to the definition evaluation score of the image that each sampling displacement corresponds is the ordinate, and the positive direction definition of the coordinate axis that corresponds with the abscissa is for the direction from a left side to the right side, sees from a left side to the right side:

if no peak exists on the image definition evaluation curve and the curve trend is continuously upward, the peak of the curve is shown on the right side, and the abscissa corresponding to the next sampling displacement is positioned on the right side of the curve; and the sampling step is the current step;

if the image definition evaluation curve has no peak and the curve trend is continuously downward, the peak of the curve is shown on the left side, and the abscissa corresponding to the next sampling displacement is positioned on the left side of the curve; and the sampling step is the current step;

and if the image definition evaluation curve has a peak, taking the sampling displacement corresponding to the peak as a central sampling displacement, sampling at two sides of the central sampling displacement, and reducing the sampling step compared with the last sampling step.

5. The focusing method according to claim 4, wherein 2 to 4 samples are respectively sampled in smaller steps on both sides of the peak of the image sharpness evaluation curve that has been found to find the peak position in smaller steps and at higher focusing accuracy; and then sampling displacement by taking the sampling displacement corresponding to the newly found peak as central sampling displacement, sampling 2-4 samples at two sides of the central sampling displacement in a smaller sampling step mode to obtain the peak position under the new focusing precision, and repeating the steps until the focusing peak and the focusing position under the target focusing precision are found.

6. The focusing method according to claim 1, wherein the sample is a biological tissue slide, a chip or a solid surface of materials.

7. The focusing method according to claim 1, wherein the steps (3) to (5) are repeated for other fields of view for which evaluation is continued, and a set of focusing positions for all fields of view for which evaluation is continued is obtained.

8. A focusing control system using the focusing method as claimed in any one of claims 1 to 7, comprising an image acquisition module, an image evaluation module, a displacement platform and a focusing control module;

the image acquisition module is used for acquiring images at different sampling displacement positions under different visual fields of the surface of the sample on the displacement platform;

the image evaluation module is used for evaluating the image definition and the content of target content of the image acquired by the image acquisition system; the system is used for comparing the maximum value of the image target content evaluation scores of the sample images corresponding to different sampling displacements in the same visual field with a given threshold value, and reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation;

the focusing control module is used for driving the displacement platform or the image acquisition system to move to reach a specified sampling displacement, driving the image acquisition module to sample an image at the sampling displacement, and driving the image evaluation module to evaluate the acquired image;

the displacement platform is used for adjusting the relative distance between the displacement platform and the image acquisition system according to the processing result of the focusing control module, so that the image acquisition module can conveniently acquire the target view image of the sample on the displacement platform again under the new sampling displacement;

when the system works, the image evaluation module evaluates an image in a certain view field acquired by the image acquisition module, and the focusing control module controls the displacement platform or the image acquisition system to adjust to the specified sampling displacement according to the evaluation result of the image evaluation module; the image acquisition module acquires the image again under the new sampling displacement, and the image evaluation module evaluates the image acquired again by the image acquisition module; and the focusing control module adjusts the sampling displacement according to the evaluation result of the image evaluation module again, drives the image acquisition module to acquire, drives the image evaluation module to evaluate, and then adjusts, acquires, evaluates and controls again, and the process is circulated until the focusing control module judges that the current sampling displacement is the focusing position in the current view according to the evaluation result of the image evaluation module.

9. A slide scanner comprising the focus control system of claim 8.

10. A focusing method for biological tissue slide scanning, comprising the steps of:

(1) evaluating the image definition and the content of target contents in the images of the biological tissue slide sample under different visual fields and at different sampling displacement positions acquired by an image acquisition system to obtain the image definition of each image and the evaluation score of the content of the target contents in the images; the target content of the image is the content of effective biological tissues in the image;

(2) comparing the maximum value of the image target content evaluation scores of the sample images at different sampling displacements in the same visual field with a given threshold value, reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation, and executing the step (3);

(3) determining the next sampling displacement according to the trend of the image definition evaluation curve for the purpose of finding the peak of the image definition evaluation curve under the target focusing precision for any visual field which is continuously evaluated, wherein the image definition evaluation curve is a curve formed by a visual field downsampling displacement data set and a definition evaluation score data set of an image corresponding to each sampling displacement;

(4) acquiring an image of the visual field and evaluating the image definition according to the determined next sampling displacement by the method in the step (1), supplementing the acquired sampling displacement data and the acquired image definition evaluation score data to the sampling displacement data set and the evaluation score data set, and determining the next sampling displacement according to a new curve trend;

(5) and (5) repeating the step (4) until a peak of the visual field definition evaluation curve under the target focusing precision is found, and determining the focusing position under the focusing precision under the visual field according to the sampling displacement corresponding to the peak.

Technical Field

The invention belongs to the technical field of focusing, and particularly relates to a focusing method and a focusing system.

Background

The hospital pathology department mainly diagnoses by means of biological slides made of patient tissues. Due to the need for department informatization, digitization and intelligence, the biological slide needs to be converted into a digital image, namely digital scanning. Fig. 2 is a schematic diagram of a conventional pathological slide structure and a relationship with an objective lens, wherein:

a: the slide glass is a tissue sample carrier, and the common slide glass specification is 76mm (length) X26mm (width), and the thickness is 1.1 +/-0.04 mm;

b: tissue samples, tissue slices removed from the sample in the production phase. For a conventional pathological tissue slide, a tissue sample is manually cut from a tissue block by using a microtome, and the thickness is controlled to be about 2 um; for other types of slides, the thickness variation is large, ranging from 1um (small cell or tissue mass, evenly distributed across the section) to 40um (large cell or tissue mass forming clumps). The tissue sample is dyed in a dyeing link so as to be convenient for observation under a microscope;

c: a coverslipping piece, a transparent gel for sealing the tissue sample and adhering the slide to the coverslip, a common coverslipping piece being a neutral gum. The thickness of the tablet is not required precisely because the dosage of the tablet cannot be controlled precisely in tabletting; after tabletting is finished, the biological tissue is fixed in the tabletting agent;

d: a cover glass, a thin glass slide covering the tissue, for protecting the underlying tissue from damage, the thickness of the cover glass being commonly about 0.17 ± 0.02 mm; for some types of slides (e.g., bone marrow smear, peripheral blood smear), the slides are not coverslipped at the production section due to diagnostic reading requirements.

E: microscope objective lenses, which are used for observing slides of different tissues according to pathological diagnosis requirements by using objective lenses of different multiples, and commonly used objective lenses comprise 10 times, 20 times, 40 times, 60 times and 100 times.

F: and accurately focusing the distance from the objective lens to the tissue. For a fixed objective lens, the distance from the objective lens to the tissue after accurate focusing is a fixed value. The image imaged by the CCD camera at one time is called the field of view, and a 40 x field of view of the objective lens images a rectangular area of about 500umX500um on the slide. For any field, in the case of a fixed objective, due to the difference in tissue height, the stage (platform on the microscope for carrying the slide) needs to be adjusted to move up and down to ensure that the distance from the objective to the tissue to be observed is the distance F, which is the focusing process. Due to the large tissue on the slide, the slide is not flat on the micrometer scale and requires focus for each field of view when viewed with a manual microscope.

Digital scanners are typically comprised of a microscope translation stage (i.e., stage), a CCD camera, a focus control algorithm, and a scan control algorithm. The focusing control algorithm controls the slide on the microscope displacement platform to move relative to the objective lens, and controls the CCD camera to perform image imaging sampling until a focusing position is found for a certain visual field. And (4) the focusing positions found by the visual fields are used by a scanning control algorithm, and the scanning control algorithm realizes digital scanning of the slide. The focus control algorithm is the basis for the digital scanning of the slide.

In the practical process of focusing a slide, the following major problems are encountered:

error exists in the thickness of the slide prepared by non-standardization of the slide preparation link (inherent error of the thickness of the slide, uneven thickness of the tissue sample slice and the sealing tablet and inherent error of the thickness of the cover glass). The specific performance of the focusing link is that the focusing positions of different slides are different, and the variation range is large (the focusing positions of some two slides can be different by 500 um); the focusing positions of different fields of the same slide are different.

Pathological slides of different tissue types and different slide preparation methods have obvious difference in image morphological expression, such as cervical cell slides, conventional tissue slides, immunohistochemical slides, bone marrow smears, peripheral blood smears and fluorescent slides, and a focusing control algorithm is required to adapt to the difference.

Different fields, the tissue samples of which are randomly distributed, and the condition that only a small number of regions have the tissue samples exists in the whole field, for the fields, the field which can be focused needs to be accurately selected by a focusing control algorithm, so that the focusing inaccuracy is avoided.

Different types of pathological slides require different objective lens multiples used in digital scanning during clinical diagnosis, and require a focusing control algorithm to support all objective lens specifications.

Depending on the light source used for microscopic observation, the pathological slides can be further classified into bright field slides (i.e., white light source imaging, such as cervical cell slides, conventional tissue slides, immunohistochemical slides, bone marrow smears, peripheral blood smears) and dark field slides (i.e., fluorescent light source imaging, such as tissue fluorescence slides, blood fluorescence slides). In clinical diagnosis, the data word is required to scan a focusing scanning bright field slide and a scanning dark field slide.

Because the slide generally has a certain gradient (caused by the addition of factors such as perpendicularity error of the objective lens and the objective lens, flatness error of the objective lens, thickness error of the slide, thickness of sample tissue and the like, the gradient is generally about 5 per thousand and cannot be eliminated), if each focusing point cannot be focused accurately, a good integral fitting plane cannot be obtained, and multiple times of focusing are needed in the scanning process to ensure the definition of most images. And because the difference of the multiple focusing positions leads to the obvious difference of the edges of the images, obvious cracks exist during image splicing.

For high power objective lenses, such as 60 power objective lenses and 100 power objective lenses, focusing can be started from a far distance in order to avoid the accidental contact between the objective lenses and the glass slide (the accidental contact can damage the objective lenses and the glass slide), but after actual focusing, the distance between the objective lenses and the glass slide is very small (for example, about 120um when the objective lenses are 100 times), how to ensure that the objective lenses cannot mistakenly contact the glass slide in the rapid focusing scanning process is also a great challenge to the focusing control algorithm.

Disclosure of Invention

In order to overcome the defects or the improvement requirements in the prior art, the invention provides a focusing method and a focusing system, wherein a search control method is adopted, a sample on a displacement platform is sampled from any position based on the existing displacement platform and an image acquisition system, the sampling result is evaluated, the next sampling position is controlled according to the trend of a curve formed by the evaluation result, the process is continued until the focusing position is searched in the minimum step, and the technical problems of multiple sampling times, low focusing speed, small search range, low adaptability and the like in the focusing method in the prior art are solved.

To achieve the above object, according to one aspect of the present invention, there is provided a focusing method including the steps of:

(1) evaluating the image definition and the content of target content in the image of sample images collected by an image collection system under different visual fields and at different sampling displacement positions to obtain the image definition and the content evaluation score of the target content in each image;

(2) comparing the maximum value of the image target content evaluation scores of the sample images at different sampling displacements in the same visual field with a given threshold value, reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation, and executing the step (3);

(3) determining the next sampling displacement according to the trend of the image definition evaluation curve for the purpose of finding the peak of the image definition evaluation curve under the target focusing precision for any visual field which is continuously evaluated, wherein the image definition evaluation curve is a curve formed by a visual field downsampling displacement data set and a definition evaluation score data set of an image corresponding to each sampling displacement;

(4) acquiring an image of the visual field and evaluating the image definition according to the determined next sampling displacement by the method in the step (1), supplementing the acquired sampling displacement data and the acquired image definition evaluation score data to the sampling displacement data set and the evaluation score data set, and determining the next sampling displacement according to a new curve trend;

(5) and (5) repeating the step (4) until a peak of the visual field definition evaluation curve under the target focusing precision is found, and determining the focusing position under the focusing precision under the visual field according to the sampling displacement corresponding to the peak.

Preferably, the sample images at different sampling displacements in step (1) are at least two sample images acquired at two different sampling displacements.

Preferably, an evaluation method is adopted to evaluate the image definition of the image and respectively evaluate the content of the target content in the image;

the evaluation method for evaluating the image definition has the advantages that the higher the evaluation score is, the clearer the image is;

the evaluation method for evaluating the content of the target content in the image has the advantage that the higher the evaluation score is, the higher the content of the target content in the image is.

Preferably, the image definition evaluation scores of the same image are one or more.

Preferably, the step (3) determines the next sampling displacement according to the trend of the image sharpness evaluation curve, specifically:

with the sampling displacement of image is the abscissa to the definition evaluation score of the image that each sampling displacement corresponds is the ordinate, and the positive direction definition of the coordinate axis that corresponds with the abscissa is for the direction from a left side to the right side, sees from a left side to the right side:

if no peak exists on the image definition evaluation curve and the curve trend is continuously upward, the peak of the curve is shown on the right side, and the abscissa corresponding to the next sampling displacement is positioned on the right side of the curve; and the sampling step is the current step;

if the image definition evaluation curve has no peak and the curve trend is continuously downward, the peak of the curve is shown on the left side, and the abscissa corresponding to the next sampling displacement is positioned on the left side of the curve; and the sampling step is the current step;

and if the image definition evaluation curve has a peak, taking the sampling displacement corresponding to the peak as a central sampling displacement, sampling at two sides of the central sampling displacement, and reducing the sampling step compared with the last sampling step.

Preferably, 2-4 samples are respectively sampled in smaller steps on two sides of the peak of the found image definition evaluation curve to find the peak position under smaller steps and higher focusing accuracy; and then sampling displacement by taking the sampling displacement corresponding to the newly found peak as central sampling displacement, sampling 2-4 samples at two sides of the central sampling displacement in a smaller sampling step mode to obtain the peak position under the new focusing precision, and repeating the steps until the focusing peak and the focusing position under the target focusing precision are found.

Preferably, the sample is a biological tissue slide, chip or a solid surface of materials.

Preferably, the steps (3) to (5) are repeated for other fields of view to be continuously evaluated, and a set of in-focus positions of all fields of view to be continuously evaluated is obtained.

According to another aspect of the present invention, there is provided a focusing control system of the focusing method, comprising an image acquisition module, an image evaluation module, a displacement platform and a focusing control module;

the image acquisition module is used for acquiring images at different sampling displacement positions under different visual fields of the surface of the sample on the displacement platform;

the image evaluation module is used for evaluating the image definition and the content of target content of the image acquired by the image acquisition system; the system is used for comparing the maximum value of the image target content evaluation scores of the sample images corresponding to different sampling displacements in the same visual field with a given threshold value, and reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation;

the focusing control module is used for driving the displacement platform or the image acquisition system to move to reach a specified sampling displacement, driving the image acquisition module to sample an image at the sampling displacement, and driving the image evaluation module to evaluate the acquired image;

the displacement platform is used for adjusting the relative distance between the displacement platform and the image acquisition system according to the processing result of the focusing control module, so that the image acquisition module can conveniently acquire the target view image of the sample on the displacement platform again under the new sampling displacement;

when the system works, the image evaluation module evaluates an image in a certain view field acquired by the image acquisition module, and the focusing control module controls the displacement platform or the image acquisition system to adjust to the specified sampling displacement according to the evaluation result of the image evaluation module; the image acquisition module acquires the image again under the new sampling displacement, and the image evaluation module evaluates the image acquired again by the image acquisition module; and the focusing control module adjusts the sampling displacement according to the evaluation result of the image evaluation module again, drives the image acquisition module to acquire, drives the image evaluation module to evaluate, and then adjusts, acquires, evaluates and controls again, and the process is circulated until the focusing control module judges that the current sampling displacement is the focusing position in the current view according to the evaluation result of the image evaluation module.

According to another aspect of the invention, there is provided a slide scanner comprising the focus control system.

According to another aspect of the present invention, there is provided a focusing method for biological tissue slide scanning, comprising the steps of:

(1) evaluating the image definition and the content of target contents in the images of the biological tissue slide sample under different visual fields and at different sampling displacement positions acquired by an image acquisition system to obtain the image definition of each image and the evaluation score of the content of the target contents in the images; the target content of the image is the content of effective biological tissues in the image;

(2) comparing the maximum value of the image target content evaluation scores of the sample images at different sampling displacements in the same visual field with a given threshold value, reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation, and executing the step (3);

(3) determining the next sampling displacement according to the trend of the image definition evaluation curve for the purpose of finding the peak of the image definition evaluation curve under the target focusing precision for any visual field which is continuously evaluated, wherein the image definition evaluation curve is a curve formed by a visual field downsampling displacement data set and a definition evaluation score data set of an image corresponding to each sampling displacement;

(4) acquiring an image of the visual field and evaluating the image definition according to the determined next sampling displacement by the method in the step (1), supplementing the acquired sampling displacement data and the acquired image definition evaluation score data to the sampling displacement data set and the evaluation score data set, and determining the next sampling displacement according to a new curve trend;

(5) and (5) repeating the step (4) until a peak of the visual field definition evaluation curve under the target focusing precision is found, and determining the focusing position under the focusing precision under the visual field according to the sampling displacement corresponding to the peak.

In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:

(1) the focusing method provided by the invention gradually accumulates the sampling images in each focusing visual field in the focusing process to generate an evaluation curve, controls the next sampling action in real time through the evaluation curve, and can quickly focus because the number of points needing to be sampled is small under the conditions of ensuring a larger search range, ensuring the focusing accuracy and ensuring the focusing position accuracy.

(2) The focusing method of the present invention supports a sufficiently large focus search range (the search range can be expanded to 600um and above) to ensure good slide adaptability. The scanner realized based on the focusing method has no special requirements on the flatness and the slide production quality. The high adaptability of the scanner is ensured, the reliability of the follow-up high-flux scanner is ensured, and a foundation is laid for the real realization of digitization in departments.

(3) Accurate focusing results to ensure good slide scanning rate. The focusing method has good focusing accuracy because the invention adopts measures such as multiple evaluation curves, comprehensive evaluation and the like. The slide scanner based on the focusing method has a scanning excellent film rate of more than 99 percent. The cervical cell slide scanning analysis system of the ThinPrep type of the Hologic company in the United states clearly shows that the excellent slide scanning rate is only 92.9 percent and only supports the focusing scanning of the cervical cell slide.

(4) The focusing method provided by the preferred embodiment of the invention is based on the existing microscope displacement platform and CCD camera imaging for sampling, and is suitable for conventional slide bright field focusing and fluorescent slide dark field focusing; some scanners in the current market are equally divided into a bright field scanner model and a dark field scanner model, so that the procurement cost of departments is increased. The scanner realized based on the focusing method of the invention can simultaneously support bright field focusing scanning and dark field focusing scanning.

(5) Can adapt to different objective lens multiples. The focusing method can ensure that the focusing precision can be rapidly converged to any focusing precision on the premise of a wide focusing search range. The slide scanner realized based on the focusing control algorithm can support the focusing scanning of 10 times, 20 times, 40 times, 60 times and 100 times of objective lenses and really support the full-time scanning of slides. Wherein the focusing precision reaches 0.025um based on the fluorescence laminating scanning of 100 times of objective lens. The general scanners on the market today (such as the 3DHistech Pannoramic series pathological section scanner, hungarian) generally only support 20-fold and 40-fold objective lenses.

(6) Because the progressive approximation method is adopted to determine the focusing position, focusing from far to near ensures that the objective lens does not touch the cover glass or the tissue sample by mistake (the glass slide or the objective lens can be damaged). The distance from the objective lens to the cover slip or tissue sample when the 100 x objective lens starts to focus is about 600um, and only about 120um after focusing, i.e. only 120um oversampling is allowed.

(7) The invention ensures a larger search range, focusing accuracy and focusing position accuracy, and can quickly focus with fewer sampling points. Under the condition of 40 times of objective lens, in the range of 600um, under the condition that the focusing precision is 0.5um, only 40 times of sampling are required on average, and the focusing time of the visual field is reduced. Mathematically, for a 600um range to search to 0.5um accuracy, 13 samples are required, but considering that multiple dense samples are required near each peak to ensure accuracy, 40 samples are already close to the theoretical minimum number of samples.

(8) Because a larger search range is ensured, the focusing accuracy is ensured, and the focusing position accuracy is ensured, a good fitting plane can be obtained only by few points for the whole glass slide, and the common method of scanning and focusing by other manufacturers is avoided. On the premise of ensuring the definition of the image, the time for focusing is reduced to the maximum extent.

(9) The focusing method provided by the invention is not only suitable for focusing the biological tissue slide with or without a cover glass, but also can be applied to all systems for focusing and imaging by adjusting the relative distance between the imaging system and the imaged object in principle.

Drawings

FIG. 1 is a flow chart of a focusing method employed in the present invention;

FIG. 2 is a schematic view of the structure of a biological slide scanner used in example 1;

FIG. 3 is an image of a field of view 1 of a cervical cell slide taken at three different sampling displacements under a 20-fold objective lens in example 1 of the present invention;

FIG. 4 is an image of the field of view 2 of a cervical cell slide at three different sampling displacements under a 20-fold objective lens in example 1 of the present invention;

FIG. 5 is the focusing curve of the field 2 of cervical cell slide under 20 times objective lens in example 1 of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

The focusing method provided by the invention, as shown in fig. 1, comprises the following steps:

(1) evaluating the image definition and the content of target content in the image of sample images collected by an image collection system under different visual fields and at different sampling displacement positions to obtain the image definition and the content evaluation score of the target content in each image;

(2) comparing the maximum value of the image target content evaluation scores of the sample images at different sampling displacements in the same visual field with a given threshold value, reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation, and executing the step (3);

(3) determining the next sampling displacement according to the trend of the image definition evaluation curve for the purpose of finding the peak of the image definition evaluation curve under the target focusing precision for any visual field which is continuously evaluated, wherein the image definition evaluation curve is a curve formed by a visual field downsampling displacement data set and a definition evaluation score data set of an image corresponding to each sampling displacement;

(4) acquiring an image of the visual field and evaluating the image definition according to the determined next sampling displacement by the method in the step (1), supplementing the acquired sampling displacement data and the acquired image definition evaluation score data to the sampling displacement data set and the evaluation score data set, and determining the next sampling displacement according to a new curve trend;

(5) and (5) repeating the step (4) until a peak of the visual field definition evaluation curve under the target focusing precision is found, and determining the focusing position under the focusing precision under the visual field according to the sampling displacement corresponding to the peak.

In order to improve the focusing efficiency, in some embodiments of the present invention, the sample images at different sampling displacements in step (1) are at least two sample images acquired at two different sampling displacements.

The image acquisition system (also called image imaging system) comprises a digital camera, an optical lens and other auxiliary components. The displacement platform comprises a platform for carrying a sample. The moving part capable of regulating and controlling sampling displacement can be bound with an image imaging system and also can be bound with a displacement platform. The sampling displacement is the distance between a displacement platform where a sample is located and the image acquisition system when the image acquisition system acquires the sample image, and the adjustment of the sampling displacement is the displacement adjustment caused by the relative movement of the displacement platform and the image imaging system in the imaging direction of the image imaging system.

The sampling displacement of the invention is adjusted according to the installation position of the mechanical moving part. The displacement platform can move, and the image imaging system is fixed; or the displacement platform is fixed and the image imaging system moves.

According to the characteristics of the adopted image acquisition system and the current application scene, the initial search focusing range is determined, and then the initial sampling position in the step (1) is determined.

The focusing method of the invention can be applied to various situations of focusing imaging by adjusting the relative distance between an image imaging system and an imaged sample, wherein the sample can be various transparent or opaque samples to be imaged, including but not limited to focusing imaging of biological tissue slides (such as slices, dripping slices, smears and the like of animals and plants), focusing imaging of chip defect detection, material solid surface detection imaging and the like.

Biological tissue slides such as biological histopathology slides, including brightfield slides imaged with a white light source, e.g., cervical cytology slides, routine tissue sections, immunohistochemical sections, bone marrow smears, peripheral blood smears, water-based specimen slides, etc., and may also include darkfield slides imaged with a fluorescent light source, e.g., tissue fluorescence slides, blood fluorescence slides, etc. For the biological tissue sample, the content of the target content in the corresponding image is the effective tissue content in the image.

And chip defect detection imaging, such as image focusing imaging of the chip surface during or after the production of the semiconductor chip.

In some embodiments of the present invention, an evaluation method is used to evaluate the image sharpness of the image and the content of the target content in the image.

The evaluation method for evaluating the image definition has the advantages that the higher the evaluation score is, the clearer the image is; the evaluation method includes, but is not limited to, one or more of gray scale variance evaluation, gray scale gradient evaluation, gray scale information entropy evaluation, Brenner evaluation, laplace evaluation, energy variance evaluation, and the like, which are generally adopted in the field of image processing.

The evaluation method for evaluating the content of the target content in the image has the advantage that the higher the evaluation score is, the higher the content of the target content in the image is. In some embodiments, the content of the target content in the image is evaluated by evaluating a proportion of an area of non-background content in the image to a total area of the image. In some application scenarios, the non-background content is a non-background color, and the background color may be defined first, and then a portion of the image that does not belong to the background color is a non-background color region.

In some embodiments of the present invention, the given threshold value in step (2) is two percent, and the threshold value represents: the image portion corresponding to the target content occupies the area percentage of the whole image, and the threshold value can be further adjusted (can be appropriately reduced or increased) according to the average distribution of the effective image (target content image) in the whole image.

The step (2) of the invention regards the image corresponding to the target component content evaluation score which is lower than or equal to the given threshold value as the image with less target content, and excludes the corresponding visual field from the visual field needing focusing.

The curve in step (3) of the invention comprises that the sampling displacement of each image of the visual field for continuous evaluation is taken as an abscissa, and the definition evaluation score of the image corresponding to each sampling displacement is taken as an ordinate to obtain the definition curve of each image in the visual field.

In order to improve focusing accuracy, the same evaluation index of the same image in the same visual field, such as an image definition evaluation index or an image target content evaluation index, can be evaluated by adopting different evaluation methods, so that the image definition evaluation score of the same image can be one or more, and the target content evaluation score in the image can also be one or more. The invention can adopt a plurality of evaluation methods to evaluate the definition of the image, and each evaluation method can obtain a definition curve, so that the definition curves under the same visual field obtained by the steps of the method can be multiple.

For the evaluation of the image definition or the content of the target content, different evaluation methods can be mutually proved, for example, for the same evaluation method of the image definition, for the same image in the same visual field, if the evaluation scores obtained by evaluating the same image by the different evaluation methods are more consistent, the accuracy of the focusing evaluation score of the visual field image is higher. On the other hand, an evaluation method suitable for the current application scenario may be selected from a plurality of evaluation methods according to different application scenarios.

In some embodiments, the step (3) determines the next sampling displacement according to the trend of the image sharpness evaluation curve, specifically:

with the sampling displacement of image is the abscissa to the definition evaluation score of the image that each sampling displacement corresponds is the ordinate, and the positive direction definition of the coordinate axis that corresponds with the abscissa is for the direction from a left side to the right side, sees from a left side to the right side:

if no peak exists on the image definition evaluation curve and the curve trend is continuously upward, the peak of the curve is shown on the right side, and the abscissa corresponding to the next sampling displacement is positioned on the right side of the curve; and the sampling step is the current step;

if no peak exists on the image definition evaluation curve and the curve trend is continuously downward, the peak of the curve is shown on the left side, and the abscissa corresponding to the next sampling displacement is located on the left side of the curve; and the sampling step is the current step;

and if the image definition evaluation curve has a peak, taking the sampling displacement corresponding to the peak as a central sampling displacement, sampling at two sides of the central sampling displacement, and reducing the sampling step compared with the last sampling step.

In some embodiments, 2 to 4 samples are respectively sampled at smaller steps on both sides of the peak of the image sharpness evaluation curve that has been found to determine the peak position at smaller steps and higher focusing accuracy, and further, the sampling displacement corresponding to the newly found peak is taken as the central sampling displacement, 2 to 4 samples are sampled at smaller sampling steps on both sides of the central sampling displacement to obtain the peak position at new focusing accuracy, and the above steps are repeated until the focusing peak and the focusing position at the target focusing accuracy are found.

In the focusing method, the focusing precision is achieved by reducing the sampling step, and the focusing position of the target view field is obtained under the step of the focusing precision. The focusing accuracy depends on several aspects, including not only the resolution of the image acquisition system, but also the mechanical accuracy of the microscope displacement stage.

The focusing precision is achieved by reducing sampling steps, the step value of the next sampling displacement adjustment is determined according to the search range, the focusing position is determined by using a step-by-step approximation method, and large steps are used for sampling in a large search range; and (3) reducing the search range and using small step one-step search near the focusing position determined by large step sampling, and continuing the step until the focusing position is determined at the minimum step, namely the focusing precision, namely the precise focusing position of the image visual field.

The present invention determines the focus position using a successive approximation method. For example, in some embodiments, at the beginning of sampling, a large step is used to sample over a large search range (e.g., a 64um step is used when searching over a 600um range). Narrowing the search range further searches using small steps (e.g., 8um steps when searching in the 100um range) near the focus position determined by the large step sample. This step is continued until the focus position is determined at the minimum step (i.e., focus accuracy, such as 0.5um), i.e., the exact focus position for the field of view. In some embodiments of the invention, the tissue peak is found for the first time and then is densely sampled near the tissue peak to ensure focusing accuracy.

The method can focus the image visual fields at any position of the surface of the sample so as to obtain the focusing position sets of the image visual fields on the surface of the sample, and then focus, scan and image the sample by utilizing the focusing position sets so as to obtain the scanning image of the sample with high accuracy.

In some embodiments, the sample is a biological tissue pathological slide, and the focusing method provided by the present invention is a biological tissue pathological slide focusing method.

The invention also provides a focusing control system using the focusing method, which comprises an image acquisition module, an image evaluation module, a displacement platform and a focusing control module;

the image acquisition module is used for acquiring images at different sampling displacement positions under different visual fields of the surface of the sample on the displacement platform;

the image evaluation module is used for evaluating the image definition and the content of target content of the image acquired by the image acquisition system; the system is used for comparing the maximum value of the image target content evaluation scores of the sample images corresponding to different sampling displacements in the same visual field with a given threshold value, and reserving the visual field corresponding to the target content evaluation score higher than the threshold value as the visual field for continuous evaluation;

the focusing control module is used for driving the displacement platform or the image acquisition system to move to reach a specified sampling displacement, driving the image acquisition module to sample an image at the sampling displacement, and driving the image evaluation module to evaluate the acquired image;

the displacement platform is used for adjusting the relative distance between the displacement platform and the image acquisition system according to the processing result of the focusing control module, so that the image acquisition module can conveniently acquire the target view image of the sample on the displacement platform again under the new sampling displacement;

when the system works, the image evaluation module evaluates an image in a certain view field acquired by the image acquisition module, and the focusing control module controls the displacement platform or the image acquisition system to adjust to the specified sampling displacement according to the evaluation result of the image evaluation module; the image acquisition module acquires the image again under the new sampling displacement, and the image evaluation module evaluates the image acquired again by the image acquisition module; and the focusing control module adjusts the sampling displacement according to the evaluation result of the image evaluation module again, drives the image acquisition module to acquire, drives the image evaluation module to evaluate, and then adjusts, acquires, evaluates and controls again, and the process is circulated until the focusing control module judges that the current sampling displacement is the focusing position in the current view according to the evaluation result of the image evaluation module.

The present invention also provides a slide scanner comprising a focus control system as described above.

The following are examples:

the slide scanner of this embodiment, as shown in fig. 2, includes a focusing control system, wherein the image acquisition module is an image imaging system including a CCD camera, the focusing control module and the image processing module are integrated in the image imaging system through software, the displacement platform is a microscope displacement platform, and the sample is a cervical cell slide.

Based on the existing microscope displacement platform and CCD camera imaging for sampling, the cervical cell slide is placed on the microscope displacement platform, and the CCD camera is adopted to collect and focus the image by using a 20-time objective lens.

Fig. 3 and 4 are images acquired by two fields of view (field of view 1 and field of view 2) of the cervical cell slide under a 20-fold objective lens under three different sampling displacements, and the sharpness of each image is respectively evaluated by an energy variance evaluation method (energy variance), a gray variance evaluation method (gray variance) and a gray gradient evaluation method (gray tenegide) to obtain an image sharpness evaluation score; the content of the target content in the image is evaluated by evaluating the proportion of the area of the non-background content in the image to the total area of the image, wherein the non-background content is a non-background color, the part of the image which does not belong to the background color is a non-background color area, the target content evaluation score of each image is obtained, the threshold value of the target content evaluation score is set to be 2%, the highest target content evaluation score in the three images of the view 1 is lower than 2%, the highest target content evaluation score in the three images of the view 2 is higher than 2%, and the highest target content evaluation score in the image of the view 1 is lower than the threshold value, so that only the view 2 is focused.

The sampling displacement of each image in the visual field 2 is taken as an abscissa, the image definition evaluation score corresponding to each sampling displacement is taken as an ordinate, the positive direction of the coordinate axis corresponding to the abscissa is defined as the direction from left to right, when the image definition evaluation curve is seen from left to right, no peak exists on the image definition evaluation curve, the curve trend is continuously upward, the peak of the curve is shown on the right side, the abscissa corresponding to the next sampling displacement is positioned on the right side of the curve, and the sampling step is kept unchanged. If the image definition evaluation curve has no peak and the curve trend is continuously downward, the peak of the curve is shown on the left side, and the abscissa corresponding to the next sampling displacement is positioned on the left side of the curve; and the sampling steps remain unchanged. If the image definition evaluation curve has a peak, taking the sampling displacement corresponding to the peak as the central sampling displacement, sampling at two sides of the central sampling displacement, and reducing the sampling step compared with the last sampling step. The present embodiment determines the in-focus position using a successive approximation method. At the very beginning of the sampling, a large step is used for sampling over a large search range and a 64um step is used for searching over a 600um range. And (3) reducing the search range and using small step one-step search near the focusing position determined by large step sampling, and using 8um step when searching in the range of 100 um. This step is continued until the focus position is determined at the minimum step, i.e. the focus accuracy is 0.5um, i.e. the exact focus position of this field of view. The focus curve of field of view 2 obtained in this way is shown in fig. 5. It can be seen that the focusing positions obtained by respectively adopting the three image definition evaluation methods are basically overlapped, which shows that the focusing position of the focusing method of the invention is accurate, and the method densely samples near the tissue peak to ensure the focusing precision.

The slide scanner of this embodiment based on this focusing method has a scanning quality ratio of 99% or more.

It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.