阅读说明：本技术 一种大视场异物检测装置及方法 (Large-view-field foreign matter detection device and method ) 是由 屈玉福 张恺 于 2021-08-12 设计创作，主要内容包括：本发明公开了一种大视场异物检测装置及方法,通过利用安装于相机前的中空内部抛光反射镜,完成对装置前后场景的同时捕捉,从而实现对相机视场范围的扩展。本发明利用反射镜通过单个相机的单次拍摄完成了对视场的扩展,装置结构简单,成本低,具有广泛的应用前景。(The invention discloses a large-view-field foreign matter detection device and method, which are used for capturing scenes in front of and behind the device simultaneously by utilizing a hollow internal polishing mirror arranged in front of a camera, so that the field range of the camera is expanded. The invention completes the expansion of the visual field by single shooting of a single camera by utilizing the reflector, and has the advantages of simple structure, low cost and wide application prospect.)

1. The utility model provides a big visual field foreign matter detection device which characterized in that includes central processing unit, data storage module, communication module, drive module, power module, imaging module, central processing unit links to each other with data storage module, communication module, drive module, power module, imaging module respectively.

2. The wide-field foreign matter detection device of claim 1, wherein the foreign matter detection device is controlled by the driving module to stay at a designated position during operation, the imaging module captures an image, the captured image is compared with a non-foreign matter image stored in the data storage module and captured in advance, and the central processing unit analyzes the captured image to determine whether foreign matter exists, and then the determination result is transmitted to the communication module and transmitted to the control center by the communication module.

3. The wide-field foreign object detection device according to claim 1, wherein the foreign object detection device runs on a dedicated rail installed on a side wall, and travels and stays on the rail controlled by a driving module.

4. The wide field of view foreign object detection device of claim 1, wherein the communication module is connected to the central processing unit, receives information of processing results from the central processing unit, and transmits the information to the control center, and the control center determines further processing measures.

5. The large field of view foreign object detection apparatus of claim 1, wherein the imaging module comprises the following components:

a light source for generating light to illuminate a scene;

the reflecting mirror is of a central rotational symmetric structure, is hollow, is internally polished into a mirror surface and is used for reflecting light;

and the camera is used for receiving the light and imaging.

6. The wide field of view foreign object detection device of claim 5, wherein said reflector is formed by a circular truncated cone with a center rotationally symmetrical, and the top and bottom surfaces of the circular truncated cone are open, so that light from the side wall right in front of the device directly passes through the hollow portion of the reflector to be received by the camera photosurface, and light from the side wall on the same side as the device is reflected by the reflector to be received by the camera photosurface.

7. The wide field of view foreign object detection device of claim 5, wherein the lower bottom surface of the reflector is close to the camera and parallel to the photosensitive surface of the camera, the interior of the truncated cone is hollowed out, and the inner wall of the truncated cone is polished to serve as the reflector.

8. The large-field foreign object detection apparatus according to claim 5, wherein an axis of the mirror and an optical axis of the camera are coincident.

9. The wide field of view foreign object detection device of claim 6, wherein the reflector can be of a truncated cone shape or a truncated pyramid shape, with the top and bottom surfaces of the truncated pyramid open, the bottom surface being close to and parallel to the camera and polished inside.

10. The wide-field foreign object detection device of claim 6, wherein the reflector can be of a truncated cone or truncated pyramid structure, or a multi-plane mirror splicing structure, any of the plurality of plane mirrors can be selected to be placed around the optical axis as required, and the angle of a single plane mirror can be adjusted to meet the requirement.

11. A large-field foreign matter detection method, characterized by using the detection device of claim 1, comprising the following operation steps:

step 1: and selecting proper device parameters and installation positions according to the actual shooting environment and the shooting requirements. After the installation is finished, shooting the environment to be detected;

step 2: segmenting the shot image to obtain a front scene image and a rear scene image;

and step 3: and respectively processing the front scene image and the rear scene image, and restoring the images into a complete image under a normal visual angle.

12. The large-field foreign object detection method according to claim 11, wherein the image captured in step 2 is a combined image in which a front scene image and a rear scene image are superimposed, and it is necessary to perform division processing on the combined image.

13. The large-field foreign object detection method according to claim 11, wherein the step 2 of dividing processing includes the steps of:

step 2.1: when a truncated cone-shaped reflector is selected for shooting, a radial image is obtained, radial cutting is needed, the inner circle after cutting is the radial image of the front scene image, and the outer ring is the radial image of the rear scene image;

step 2.2: when a frustum-shaped reflector or a plane mirror spliced reflector is selected for shooting, a combined image of a front scene image and a rear scene image at different visual angles is obtained, the imaging positions of different positions of a scene on a final image are analyzed according to the imaging principle, and the image is reasonably segmented for further processing.

14. The large-field foreign object detection method according to claim 11, wherein the step 3 includes the steps of:

step 3.1: when a truncated cone-shaped reflector is selected for shooting, a radial image of a front scene image and a radial image of a rear scene image which are obtained after cutting need to be processed respectively, and conversion from a radial visual angle to a three-dimensional visual angle is achieved;

step 3.2: when a frustum-shaped reflector or a plane mirror splicing reflector is selected for shooting, planar image splicing processing needs to be performed on planar images obtained by cutting under different viewing angles so as to restore the planar images into complete scene images.

15. The foreign object detection method according to claim 14, wherein the radial image of step 3.1 needs to be processed to realize the conversion from the radial viewing angle to the stereoscopic viewing angle, and the method comprises the following operation steps:

step 3.1.1: continuously extracting radial lines of the radial images, dividing the radial lines into three parts which are respectively used for generating a left view, a central view and a right view, and overlapping the extracted images to generate three-view images;

step 3.1.2: blurring the center view picture to compensate for mirror aberration;

step 3.1.3: and taking the central view as a reference, and performing stereo matching by using the left view image and the right view image by using a normalized cross-correlation parallax matching method to obtain a final image.

Technical Field

The invention relates to a large-view-field foreign matter detection device and method, which are suitable for various fields needing large-range detection, such as subway tunnels and the like, and belong to the technical field of rail transit safety.

Background

The field of view is one of the important parameters of an optical imaging system, which, together with the resolution, determines the amount of information that can be obtained by the optical imaging system. Due to the limitation of the principle of the optical imaging system, the optical imaging system is difficult to reach an imaging range of more than 180 degrees, and in the fields of video monitoring, military industry, biomedicine and the like, the field size of the conventional imaging system cannot meet the application requirement, so that an imaging system with a larger field range is required.

The imaging method with large field range can be divided into two types, namely scanning splicing imaging method and direct shooting imaging method. The scanning splicing imaging method can be divided into single-lens scanning splicing and multi-lens simultaneous shooting splicing, wherein the former needs to acquire images for multiple times, time delay exists, the speed is slow, and the imaging quality is poor due to the existence of jitter in the scanning process; the latter system has high cost, large volume and limitation in application. The direct shooting method adopts the same set of optical imaging lens and detector to image the scene once, has high speed and high efficiency, is a current research hotspot and has wide application prospect.

And in subway tunnel foreign matter detection area, when avoiding detection device to install at the tunnel top, the device accident drops and causes tunnel foreign matter invasion accident, often installs detection device in the tunnel lateral wall, and the device of installing in the lateral wall often can only shoot the foreign matter condition of tunnel one side, can not realize shooting well and control to the foreign matter condition of installation side. In order to meet the requirements of cost saving and risk reduction, direct shooting optical imaging systems with larger view field ranges need to be researched for detecting foreign matters in subways.

Disclosure of Invention

In view of the above requirements, the invention discloses a large-view-field foreign matter detection device and method.

The purpose of the invention is realized by the following technical scheme:

the large-view-field foreign matter detection device comprises a central processing unit, a data storage module, a communication module, a driving module, a power supply module and an imaging module, wherein the central processing unit is connected with the data storage module, the communication module, the driving module, the power supply module and the imaging module respectively.

Furthermore, the large-view-field foreign matter detection device is controlled by the driving module to stay at a specified position during operation, the imaging module is used for shooting images, the shot images and the non-foreign matter images which are stored in the data storage module and shot in advance are compared and analyzed by the central processing unit to judge whether foreign matters exist, and then the judgment result is transmitted to the communication module and transmitted to the control center by the communication module.

Further, the large-visual-field foreign matter detection device runs on a special rail mounted on the side wall, and the driving module controls the device to run and stop on the rail.

Furthermore, the communication module is connected with the central processing unit, receives the processing result information of the central processing unit, sends the information to the control center, and determines further processing measures by the control center.

Further, the imaging module comprises the following components:

a light source for generating light to illuminate a scene;

the reflecting mirror is of a central rotational symmetric structure, is hollow, is internally polished into a mirror surface and is used for reflecting light;

and the camera is used for receiving the light and imaging.

Furthermore, the reflector is composed of a circular table with a center rotationally symmetrical, the upper bottom surface and the lower bottom surface of the circular table are both provided with openings, light rays from the side wall in front of the device directly penetrate through the hollow part of the reflector to be received by the photosensitive surface of the camera, and light rays from the side wall on the same side of the device are received by the photosensitive surface of the camera after being reflected by the reflector.

Furthermore, the lower bottom surface of the reflector is close to the camera and is parallel to the photosensitive surface of the camera, the round table is hollowed, and the inner wall of the round table is polished to be used as the reflector.

Further, the axis of the mirror and the optical axis of the camera are coincident.

Furthermore, the reflecting mirror can be of a truncated cone-shaped structure and a truncated pyramid-shaped structure, the upper bottom surface and the lower bottom surface of the truncated pyramid are both open, the lower bottom surface is close to and parallel to the camera, and the inner side of the lower bottom surface is polished.

Furthermore, the reflecting mirror can be in a truncated cone-shaped structure or a truncated pyramid-shaped structure, and can also be in a splicing structure with a plurality of plane mirrors, any plurality of plane mirrors can be placed around the optical axis according to needs, and the angle of a single plane mirror can be adjusted to meet the needs.

The invention also provides a large-view-field foreign matter detection method, and the detection device comprises the following operation steps:

step 1: and selecting proper device parameters and installation positions according to the actual shooting environment and the shooting requirements. After the installation is finished, shooting the environment to be detected;

step 2: segmenting the shot image to obtain a front scene image and a rear scene image;

and step 3: and respectively processing the front scene image and the rear scene image, and restoring the images into a complete image under a normal visual angle.

Further, the image captured in step 2 is a combined image in which the front scene image and the rear scene image are overlapped, and it is necessary to perform a segmentation process on the combined image.

Further, the segmentation processing of step 2 includes the following operation steps:

step 2.1: when a truncated cone-shaped reflector is selected for shooting, a radial image is obtained, radial cutting is needed, the inner circle after cutting is the radial image of the front scene image, and the outer ring is the radial image of the rear scene image;

step 2.2: when a frustum-shaped reflector or a plane mirror spliced reflector is selected for shooting, a combined image of a front scene image and a rear scene image at different visual angles is obtained, the imaging positions of different positions of a scene on a final image are analyzed according to the imaging principle, and the image is reasonably segmented for further processing.

Further, the step 3 includes the following steps:

step 3.1: when a truncated cone-shaped reflector is selected for shooting, a radial image of a front scene image and a radial image of a rear scene image which are obtained after cutting need to be processed respectively, and conversion from a radial visual angle to a three-dimensional visual angle is achieved;

step 3.2: when a frustum-shaped reflector or a plane mirror splicing reflector is selected for shooting, planar image splicing processing needs to be performed on planar images obtained by cutting under different viewing angles so as to restore the planar images into complete scene images.

Further, the radial image in step 3.1 needs to be processed to realize the conversion from the radial viewing angle to the stereoscopic viewing angle, and the method includes the following operation steps:

step 3.1.1: continuously extracting radial lines of the radial images, dividing the radial lines into three parts which are respectively used for generating a left view, a central view and a right view, and overlapping the extracted images to generate three-view images;

step 3.1.2: blurring the center view picture to compensate for mirror aberration;

step 3.1.3: and taking the central view as a reference, and performing stereo matching by using the left view image and the right view image by using a normalized cross-correlation parallax matching method to obtain a final image.

The invention has the beneficial effects that:

the invention completes the expansion of the view field by single shooting of a single camera by utilizing the reflector;

the device has simple structure and low cost;

the invention can increase the safety of railway lines and reduce the accident rate in the running process of a railway system.

Drawings

FIG. 1 is a schematic view of an overall apparatus of the present invention

FIG. 2 is a schematic view of an imaging module of the device

FIG. 3 is an image relationship diagram of an object plane and an image plane

FIG. 4 is a schematic view of a frustum of a prism

FIG. 5 is a schematic view of a plane splicing type reflector

FIG. 6 is a flow chart of a method of the present invention

Detailed Description

The invention is further elucidated with reference to the drawings and the detailed description. It should be understood that the following detailed description is illustrative of the invention only and is not intended to limit the scope of the invention.

Fig. 1 is a schematic diagram of the overall device of the present invention, and it can be seen from fig. 1 that the foreign object detection device with a large field of view disclosed by the present invention includes a central processing unit 3, a data storage module 5, a communication module 4, a driving module 2, a power supply module 7, and an imaging module composed of a light source 6, a camera 9 and a reflector 8.

The central processing unit 3 is respectively connected with the data storage module 5, the communication module 4, the driving module 2, the power supply module 7 and the imaging module, and controls the modules to be matched with each other and run normally.

The operation process of the foreign matter detection device is as follows: the driving module 2 controls the device to stay at the designated position, the imaging module shoots images, the shot images and the non-foreign object images which are stored in the data storage module 5 and shot in advance are compared, analyzed and judged by the central processing unit 3 to determine whether foreign objects exist, and then the judgment result is transmitted to the communication module 4 and transmitted to the control center by the communication module 4.

The foreign matter detection device runs on a special rail 1 arranged on the side wall, and the driving module 2 controls the foreign matter detection device to move and stay on the rail. When this device was applied to subway tunnel foreign matter and detects, in order to avoid installing when the tunnel top, the device took place the accident and drops the accident, caused the foreign matter invasion risk, so with special rail mounting on the lateral wall of subway tunnel one side.

The communication module 4 is connected with the central processing unit 3, receives the processing result information of the central processing unit 3, and sends the information to the control center, and the control center determines further processing measures.

The imaging module is described with reference to fig. 2.

The reflector 8 is formed by a circular table with a center rotationally symmetrical, the upper bottom surface and the lower bottom surface of the circular table are both opened, the lower bottom surface is close to the camera and is parallel to the photosensitive surface of the camera 9, the interior of the circular table is hollowed, and the inner wall of the circular table is polished to be used as the reflector. In use, the axis of the mirror 8 and the optical axis of the camera 9 are coincident.

When the imaging module is used, light rays from the side wall right in front of the device directly pass through the hollow part of the reflector 8 to be received by the photosensitive surface of the camera 9, and light rays from the side wall on the same side of the device are reflected by the reflector 8 and then received by the photosensitive surface of the camera 9. Therefore, the single camera can simultaneously acquire the scene information before and after the camera through one-time imaging.

FIG. 3 is an imaging relationship diagram of an object plane and an image plane. The angle formed by the reflector 8 and the optical axis is theta. The object plane 10 projects to the image plane 11 at an angle phi. According to the geometric relationship, the relationship between the two is

For achieving better imaging effect, it is desirable that the area of the object plane 10 projected onto the imaging plane 11 is as large as possible, i.e., it is desirable thatAs small as possible, it is desirable that θ, i.e., the angle of the mirror 8 with respect to the optical axis, be as large as possible.

If θ is too large, the area of the upper bottom surface of the opening of the circular truncated cone will be too small, which may affect the imaging of the forward object by the device. Therefore, the radius of the upper bottom surface and the lower bottom surface of the circular truncated cone and the length of the circular truncated cone are reasonably set according to specific use conditions, so that the purpose of smoothly imaging scenes in the front and the rear of the device is met.

Besides reasonably setting parameters, the invention also provides other structures to solve the problem of front and back imaging.

Fig. 4 is a schematic view of a frustum of a prism. The upper and lower bottom surfaces of the frustum-shaped reflector 8 are open, the lower bottom surface is close to and parallel to the camera 9, and the inner side is polished to be used as a reflector. Similar to the truncated cone-shaped reflector, most of the light rays from the front of the device in the field of view directly pass through the upper bottom surface of the opening, pass through the reflector 8 and are directly received by the photosensitive surface of the camera 9, and a small part of the light rays are finally received by the photosensitive surface of the camera 9 after being reflected once or more times by the reflector; all light rays from the rear of the device within the field of view are reflected by the reflector and finally received by the photosensitive surface of the camera 9, so that the front and rear objects of the device can be imaged by a single device through one-time shooting.

FIG. 5 is a schematic view of a planar tiled mirror structure. The reflector 8 of the structure is formed by combining a plurality of plane mirrors which are symmetrically arranged around an optical axis. In the actual use process, the angle between the single plane mirror and the optical axis can be adjusted to meet the requirement. Meanwhile, the plane mirrors are not directly connected but leave a certain gap for shooting a scene in front. If the plane mirrors adopt a structure which is tightly connected with each other, the opening area of the front object only has an upper bottom surface which is surrounded by the plane mirrors. When the inclination angle is larger, the area of the upper bottom surface is limited. This results in an excessively small field of view for imaging the object in front, and with the structure shown in fig. 5, this problem can be effectively avoided, and the field of view for imaging the object in front is enlarged on the premise of ensuring a sufficiently large inclination angle.

Fig. 6 is a flow chart of the method of the present invention, and it can be seen from fig. 6 that the method for detecting foreign matters with a large field of view disclosed by the present invention adopts the apparatus shown in fig. 1, and mainly includes the following steps:

step 1: and selecting proper device parameters and installation positions according to the actual shooting environment and the shooting requirements. After the installation is finished, shooting the environment to be detected;

step 2: segmenting the shot image to obtain a front scene image and a rear scene image;

and step 3: and respectively processing the front scene image and the rear scene image, and restoring the images into a complete image under a normal visual angle.

The invention aims to capture front and rear scenes simultaneously by single shooting of a single device, so that the acquired original image is a combined image formed by overlapping a front scene image and a rear scene image and needs to be subjected to segmentation processing. Different mirrors are used, and the types of the acquired original images are different, so different operation steps are adopted according to the used mirrors.

When the truncated cone-shaped reflector is selected for shooting, a radial image is obtained, and the following operation steps are required:

step 2.1: and performing radial cutting on the acquired radial image, wherein the inner circle after cutting acquires a radial image of the front scene image, and the outer circle acquires a radial image of the rear scene image.

When a frustum-shaped reflector or a plane mirror spliced reflector is selected for shooting, a combined image of front and rear scene plane images at different visual angles is obtained, and the following operation steps are required:

step 2.2: according to the imaging principle, the imaging positions of different positions of a scene on a final image are analyzed, and the image is reasonably segmented for further processing.

After the image segmentation is completed, the obtained images need to be processed respectively according to the situation of using the reflecting mirror, and the images are restored to the images under the normal viewing angle.

When the truncated cone-shaped reflector is selected, the radial images of the front scene and the rear scene obtained through the step 2 need to be subjected to the following operation steps:

step 3.1: the method comprises the following steps of respectively processing a radial image of a front scene image and a radial image of a rear scene image obtained after cutting, and realizing the conversion from a radial visual angle to a three-dimensional visual angle, and specifically comprises the following steps:

step 3.1.1: continuously extracting radial lines of the radial images, dividing the radial lines into three parts which are respectively used for generating a left view, a central view and a right view, and overlapping the extracted images to generate three-view images;

step 3.1.2: blurring the center view picture to compensate for mirror aberration;

step 3.1.3: and taking the central view as a reference, and performing stereo matching by using the left view image and the right view image by using a normalized cross-correlation parallax matching method to obtain a final image.

When a frustum-shaped reflector or a plane mirror spliced reflector is selected, the front and back scene plane images at different visual angles are obtained in the step 2, and the following operation steps are required:

step 3.2: and carrying out plane image splicing treatment on the plane images obtained by cutting under different visual angles so as to restore the plane images into complete scene images.

In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.