阅读说明：本技术 一种圆钢修磨质量检测系统中光照强度的控制方法 (Control method for illumination intensity in round steel grinding quality detection system ) 是由 邹堃 胡继康 申屠理锋 李劲 奚嘉奇 于 2018-06-22 设计创作，主要内容包括：本发明揭示了一种圆钢修磨质量检测系统中光照强度的控制方法,通过照明光源照射被测圆钢；通过工业相机捕捉圆钢表面图像并输入数据处理模块；通过数据处理模块把图像转换为视频信号发送到显示模块,同时根据钢坯信息计算光照强度,并将控制信号发送到可调光源驱动模块；通过可调驱动模块调节光源亮度,以达到数据处理模块设定的光照强度。本发明采用在线计算来进行实时调节,用实例数据来进行自学习的,可以自适应调节修磨工序中的光照强度,有效减少圆钢修磨质量检测系统的误判。(The invention discloses a control method of illumination intensity in a round steel grinding quality detection system, which irradiates a detected round steel through an illumination light source; capturing a round steel surface image through an industrial camera and inputting the round steel surface image into a data processing module; converting the image into a video signal through a data processing module, sending the video signal to a display module, calculating the illumination intensity according to the billet information, and sending a control signal to an adjustable light source driving module; the brightness of the light source is adjusted through the adjustable driving module so as to achieve the illumination intensity set by the data processing module. The invention adopts online calculation to carry out real-time adjustment, carries out self-learning by using example data, can adaptively adjust the illumination intensity in the grinding process, and effectively reduces the misjudgment of the round steel grinding quality detection system.)

1. A control method for illumination intensity in a round steel grinding quality detection system is characterized by comprising the following steps: the method comprises the following steps:

A. irradiating the round steel to be detected by an illumination light source;

B. capturing a round steel surface image through an industrial camera and inputting the round steel surface image into a data processing module;

C. converting the image into a video signal through a data processing module, sending the video signal to a display module, calculating the illumination intensity according to the billet information, and sending a control signal to an adjustable light source driving module;

D. the brightness of the light source is adjusted through the adjustable driving module so as to achieve the illumination intensity set by the data processing module.

2. The method for controlling the illumination intensity in the round steel grinding quality detection system according to claim 1, wherein the method comprises the following steps: in step C, the calculation method includes:

C1. obtaining type C of steel₀，And obtaining the corresponding driving current value I₀；

C2. Obtaining the diameter D of the billet₀，D₀∈[D₁，D_N]Calculating the current coefficient alpha₀；

C3. Calculating the current drive current value I₁。

3. The method for controlling the illumination intensity in the round steel grinding quality detection system according to claim 2, wherein the method comprises the following steps: in step C2, the current coefficient α is set₀The calculation formula of (2) is as follows:

4. the method for controlling the illumination intensity in the round steel grinding quality detection system according to claim 3, wherein the method comprises the following steps: in step C3, the driving current value I₁The calculation formula of (2) is as follows:

I₁＝I₀×α₀×k_i

in the formula, k_iIs the self-learning coefficient after the current ith self-learning, and the initial value k thereof₀Is 1.

5. The method for controlling the illumination intensity in the round steel grinding quality detection system according to claim 4, wherein the method comprises the following steps: the self-learning method comprises the following steps:

for the working conditions satisfying the conditions, after the coping operation is started, the defect brightness P 'is calculated in the image captured by the industrial camera'_QAnd background luminance P'_GDifference therebetween, then the i +1 th learning is currently performedLater self-learning coefficient k_i+1Comprises the following steps:

where Δ k is the learning step, P₀，P₁To identify a threshold.

Technical Field

The invention relates to a round steel grinding quality detection technology, in particular to a control method of illumination intensity in a round steel grinding quality detection system.

Background

In the round steel grinding process, an operator needs to grind the billet according to the defects marked in the previous flaw detection process on the surface of the round steel. And then, manually or machine vision is used for confirming that the defects are eliminated, and then the coping operation is finished. If the defect is not eliminated, the thinning is carried out again until the defect is eliminated. Because the defects of the steel billet are small and difficult to distinguish, the coping operation usually needs good lighting conditions. Currently, shop lights of sufficient brightness are typically installed at the sharpening station to provide illumination. However, the steel slabs have a large difference in the light reflectance of the surface due to the difference in the processing method. For example, the surface of the shot-blasted billet is almost completely black; the turned steel billet has the silver bright color of metal. The fixed illumination intensity is difficult to be applied to each steel billet, and the surface of the steel billet is dazzled due to over-strong illumination; and the illumination is too weak, so that the surface defects are not easy to distinguish. This problem is more obvious when using machine vision round steel coping quality detecting system, can increase systematic erroneous judgement. With the popularization of LED lamps convenient to adjust illumination and the application of machine vision systems, light systems capable of being adjusted in a self-adaptive mode become more and more significant.

Currently, some methods and apparatuses for automatically adjusting the illumination intensity have appeared, such as a flash intensity control method and an image pickup device (CN 104243842A), and a method for controlling the light intensity by detecting the size of a target image has been proposed. The essence of the method is that the illumination intensity is adjusted by judging the distance of an object, and the position of a workpiece is fixed in round steel grinding, so the method is not suitable for being adopted. As another example, a light source adaptive control method (CN 106506980 a) based on determining the pixel intensity of an image proposes a method of adjusting the illumination intensity by using LabVIEW to obtain the mean value of the brightness of an object image. The LabVIEW system is costly and can simply use the processor in the machine vision system to calculate pixel brightness from the rows. In addition, in round bar thinning, the contrast between defects and billet surface brightness is of greater concern. Rather than the overall image average brightness. Therefore, this method is not suitable for use.

Disclosure of Invention

The invention aims to provide a method for controlling the illumination intensity in a round steel grinding quality detection system, which can reduce the misjudgment of the round steel grinding quality detection system by adaptively adjusting the illumination intensity in a grinding process.

A control method for illumination intensity in a round steel grinding quality detection system comprises the following steps:

A. irradiating the round steel to be detected by an illumination light source;

B. capturing a round steel surface image through an industrial camera and inputting the round steel surface image into a data processing module;

C. converting the image into a video signal through a data processing module, sending the video signal to a display module, calculating the illumination intensity according to the billet information, and sending a control signal to an adjustable light source driving module;

D. the brightness of the light source is adjusted through the adjustable driving module so as to achieve the illumination intensity set by the data processing module.

In step C, the calculation method includes:

C1. obtaining the type of steel gradeAnd obtaining the corresponding driving current value I₀；

C2. Obtaining the diameter D of the billet₀，Calculating the current coefficient alpha₀；

C3. Calculating the current drive current value I₁。

In step C2, the current coefficient α is set₀The calculation formula of (2) is as follows:

in step C3, the driving current value I₁The calculation formula of (2) is as follows:

I₁＝I₀×α₀×k_i

in the formula, k_iIs the self-learning coefficient after the current ith self-learning, and the initial value k thereof₀Is 1.

The self-learning method comprises the following steps:

for the working conditions satisfying the conditions, after the coping operation is started, the defect brightness P 'is calculated in the image captured by the industrial camera'_QAnd background luminance P'_GDifference between them, then the current self-learning coefficient k after the i +1 th learning_i+1Comprises the following steps:

where Δ k is the learning step, P₀，P₁To identify a threshold.

By adopting the technical scheme of the invention, online calculation is adopted for real-time adjustment, and example data is used for self-learning, so that the illumination intensity in the grinding process can be adjusted in a self-adaptive manner, and the misjudgment of the round steel grinding quality detection system is effectively reduced.

Drawings

In the present invention, like reference numerals refer to like features throughout, wherein:

FIG. 1 is a schematic diagram of a control method of the present invention.

Detailed Description

The technical scheme of the invention is further explained by combining the drawings and the embodiment.

The control method for the illumination intensity in the round steel grinding quality detection system adopts a control system shown in figure 1, and comprises a light source 1, an industrial camera 3, a data processing module 4, a display module 5 and an adjustable light source driving module 6, and the main control steps are as follows: firstly, a factory lighting source 1 emits light with certain intensity to irradiate the round steel 2 to be measured; the surface image of the round steel 2 is captured by an industrial camera 3; the data processing module 4 converts the image into a video signal and sends the video signal to the display module 5, and meanwhile, the illumination intensity is calculated according to the information such as the type and the diameter of the billet steel and the like, and the illumination intensity is sent to the adjustable light source driving module 6; the adjustable driving module 6 adjusts the brightness of the light source to reach the illumination intensity set by the data processing module 4. Wherein:

the light source 1 is an adjustable lamp, and an adjustable LED factory illuminating lamp is usually selected. The adjustable drive module 6 is typically a current adjustable switching power supply. The industrial camera 3 may be a CCD or CMOS industrial camera with a relatively strong interference rejection capability. The data processing module 4 can be a processing unit such as an industrial personal computer, a PLC, a DSP and the like. The display module 5 may select a suitable display.

When the light source 1 irradiates the round steel 2 to be measured, the industrial camera 3 shoots an image of the surface of the round steel and transmits the image to the data processing module 4, the data processing module 4 adjusts the illumination intensity, and the specific adjusting method can comprise three parts of calibration, calculation and self-learning.

Firstly, the calibration packet part comprises the following steps:

1. calibration

The calibration is carried out in an off-line state, the off-line detection precision is high, the method is the best mode for determining the set value of the illumination intensity, and the characteristics of the calibration system in the off-line mode can provide stable and appropriate initial parameters for subsequent calculation.

1.1 calibrating billet variety parameters

The surface condition of the billet steel adopting different treatment methods is greatly different, and the invention needs to be at a certain diameter D_BThe influence of different kinds of steel billets on the imaging is calibrated down, D_BTypical billet diameters. First, the surface treatment of the steel slabs needs to be classified into M types, and the value of M is generally determined according to the number of the surface treatment methods of the steel slabs. For example, if the billet incoming material is divided into turning, shot blasting and unprocessed, then M may be taken to be 3. For each class, the diameter is uniformly taken as D_BThe current I of the driving module is adjusted, so that the defect brightness P is obtained in the image acquired by the industrial camera_QAnd the background brightness P_GDifference between the two values is at identification threshold P₀，P₁In between, i.e

P₀≤|P_Q-P_G|≤P₁

Defective brightness P_QThe brightness of the defect area can be obtainedThe value may be the luminance value of a certain feature point. Background brightness P_GThe brightness mean value of the background area can be taken, and the brightness mean value of some characteristic points can also be taken. Threshold value P₀The value is taken as the threshold value P which is suitable for the human eye to distinguish the defects more obviously₁The value is preferably taken so that the image is not dazzling and does not have glare.

For M steel billets, the billet type number is C₁，C₂，…，C_(M-1)，C_MTaking the diameter D from each kind of billet_BThe proper illumination driving current value I is obtained by measurement_C1，I_C2，…，I_C(M-1)，I_CMTherefore, the defects and the background on each steel billet can be distinguished more obviously without dazzling.

1.2 calibrating diameter parameters of steel billets

For a round steel billet, the surface curvature of the round steel billet affects the projection image of the defect on an industrial camera, and the larger the curvature of the round steel, the smaller the diameter, and the more difficult the defect is to be distinguished on the image, so that for a billet with a smaller diameter, the illumination needs to be properly enhanced. Selecting N calibration points within the value range of the diameter D of the round steel, wherein the diameter of each calibration point is D₁，D₂，…，D_(N-1)，D_NThe diameter has a larger value range, so that the value of N can be more; otherwise, it may be less, but should generally not be less than 3. D₁Can be taken as the lower diameter limit, D_NCan be taken as the upper diameter limit, and D should be included in the value_BI.e. the diameter should be D₁，…，D_B，…，D_N。

Then in a certain class of steel C_BIn, take the diameter as D₁，…，D_B，…，D_NThe proper illumination driving current value I of the steel billet is obtained by measurement_D1，…，I_B，…，I_DNThe same procedure as in the A1 procedure allows the defects on each billet to be clearly distinguished from the background. Steel grade C_BCan be taken as a common typical steel grade and is C₁，C₂，…，C_(M-1)，C_MWithin the range. The illumination can be driven by current for convenient calculationValue I_D1，…，I_B，…，I_DNAre all divided by I_BObtaining the current coefficient alpha_D1，…，α_B，…，α_DNWherein the current coefficient α_B＝1。

Two, calculating

After the online operation, in the data processing module 4, the calculation process is as follows:

2.1 reading Steel type C₀，Finding out the corresponding drive current value I obtained in the step of 1.1 calibration₀。

2.2 reading billet diameter D₀，D₀∈[D₁，D_N]Calculating the current coefficient alpha₀The following were used:

2.3 calculating the Current drive Current value I_I，I₁＝I₀×α₀×k_i

Wherein: k is a radical of_iIs the self-learning coefficient after the current ith self-learning, and the initial value k thereof₀Is 1.

Third, self-learning

In actual use, factors such as aging of a light source, change of ambient light, and dust deposition on a camera lens cause matching deviation between a driving current value calculated by using offline calibration data and actually required illumination intensity, and the deviation generally increases with the use time and finally even accumulates to an unacceptable degree for a user. It is very necessary to correct the deviation by on-line self-learning. In the present invention, the device requires constant self-learning to correct the parameters during use. Because the grinding incoming material types are more and the working conditions are different, the self-learning needs to meet certain conditions.

In order to approach the working condition of off-line calibration as much as possible, steel grade C can be selected_BDiameter D_BDuring steel billet grindingA self-learning process is initiated. If the steel grade and the diameter are difficult to find in the calibration for a long time due to production scheduling, product specification change and the like, the diameter value range can be properly widened, but the steel grade should not be changed.

For the working condition meeting the condition, after the coping operation is started, the defect brightness P 'is calculated in the image collected by the industrial camera'_QAnd background luminance P'_GDifference between them, then the current self-learning coefficient k after the i +1 th learning_i+1Comprises the following steps:

wherein, the value of the Δ k is a learning step length, the value of the Δ k can be adjusted according to the self-learning frequency, if the use condition allows, the self-learning is more frequent, and the Δ k can be smaller; otherwise, the larger the size of the sample. Δ k may be a fixed step size or a dynamic step size. If the designer is more transparent about each illumination influence factor of the application occasion and the application working condition is more familiar, the fixed step length can be used. Conversely, dynamic step sizes may be used. But in general, this will not be the case due to changes in system parameters.