CN113706495B

CN113706495B - Machine vision detection system for automatically detecting lithium battery parameters on conveyor belt

Info

Publication number: CN113706495B
Application number: CN202110967897.5A
Authority: CN
Inventors: 卢盛林; 贺珍真; 谷宏亮; 何仕龙
Original assignee: Guangdong OPT Machine Vision Co Ltd
Current assignee: Guangdong OPT Machine Vision Co Ltd
Priority date: 2021-08-23
Filing date: 2021-08-23
Publication date: 2022-11-22
Anticipated expiration: 2041-08-23
Also published as: CN113706495A

Abstract

The invention provides a machine vision detection system for automatically detecting lithium battery parameters on a conveyor belt, which comprises: the image acquisition module is used for continuously acquiring images of the target lithium battery moving at a preset position according to a preset period to obtain a group of target images; the image processing module is used for preprocessing the target image and synthesizing to obtain an image to be detected of the target lithium battery; the image analysis module is used for analyzing the image to be detected to obtain an analysis result; the parameter acquisition module is used for acquiring the to-be-detected parameters of the target lithium battery based on the analysis result; the line scanning camera can be used for photographing and imaging moving objects, the position, the form, the size and other characteristics of the lithium battery can be accurately positioned under the condition that the lithium battery does not stop moving, and the detection efficiency is greatly improved.

Description

Machine vision detection system for automatically detecting lithium battery parameters on conveyor belt

Technical Field

The invention relates to the technical field of moving workpiece detection, in particular to a machine vision detection system for automatically detecting lithium battery parameters on a conveyor belt.

Background

At present, manual detection of lithium battery parameters is mostly adopted in industrial production, but the manual detection of the lithium batteries on a conveyor belt is easy to cause visual fatigue and low in reliability, and the replacement of the lithium batteries needs to be performed manually for a long time to be familiar with products, so that the cost is high, and the production requirements cannot be met;

the existing conveying belt detection device adopts a conveying belt device of a machine vision detection system, but because an image acquisition device is fixed and nonadjustable, acquired images are blurred, lithium battery parameters cannot be accurately detected, or all parameters of lithium batteries cannot be acquired, so that the detection is inaccurate, all parameters of the lithium batteries cannot be completely detected, and the problems of low detection accuracy and low detection efficiency cannot be solved due to the fact that manual adjustment is needed subsequently or the images are detected to obtain the lithium battery parameters.

Machine vision-based inspection methods have been one of the main research directions for automated inspection systems.

Therefore, the invention provides a machine vision detection system for automatically detecting parameters of lithium batteries on a conveyor belt.

Disclosure of Invention

The invention provides a machine vision detection system for automatically detecting parameters of a lithium battery on a conveyor belt, which adopts a line scanning camera to photograph and image a moving object, can accurately position the characteristics of the lithium battery such as position, form, size and the like under the condition that the lithium battery does not stop moving, and greatly improves the detection efficiency. On the basis, an algorithm for measuring relevant parameters of the lithium battery is designed.

The invention provides a machine vision detection system for automatically detecting lithium battery parameters on a conveyor belt, which comprises:

the image acquisition module is used for continuously acquiring images of the target lithium battery moving at a preset position according to a preset period to obtain a group of target images;

the image processing module is used for preprocessing the target image and then synthesizing to obtain an image to be detected of the target lithium battery;

the image analysis module is used for analyzing the image to be detected to obtain an analysis result;

and the parameter acquisition module is used for acquiring the to-be-detected parameters of the target lithium battery based on the analysis result.

Preferably, the machine vision inspection system for automatically detecting parameters of lithium batteries on a conveyor belt comprises:

the first acquisition module is used for controlling each image acquisition device to simultaneously acquire images of the target lithium battery when the target lithium battery reaches a preset position so as to obtain a group of first images;

the calling module is used for calling the equipment identifier of each image acquisition equipment and a preset sample image;

the first acquisition module is used for acquiring a first image parameter of the first image and a second image parameter of the sample image;

a recall correlation module to correlate the sample image with the first image based on the device identification;

the setting module is used for comparing the first image parameter with the second image parameter based on the correlation result to obtain a comparison result, and setting the shooting parameter of each image acquisition device based on the comparison result corresponding to each image acquisition device;

the second acquisition module is used for controlling the image acquisition equipment at the preset position to rotationally acquire the image of the target lithium battery according to the shooting parameters to obtain a group of target images;

wherein the image parameters include: the brightness-to-dark area ratio, a first average brightness corresponding to a bright area and a second average brightness corresponding to a dark area;

wherein the shooting parameters include: exposure range, exposure brightness.

Preferably, the machine vision detection system for automatically detecting parameters of lithium batteries on a conveyor belt, the first obtaining module includes:

the matrix acquisition module is used for acquiring pixel values of all pixel points in the first image, selecting a non-edge pixel point in the first image, acquiring a gray-scale matrix of the first image with the non-edge pixel point as the center, and calculating a first brightness value matrix of the first image with the non-edge pixel point as the center based on the pixel values of all the pixel points and the gray-scale matrix;

the first screening module is used for calculating an edge point judgment value based on the first brightness value matrix and a preset first same-order conversion matrix, determining a corresponding non-edge pixel point as a first edge point when the edge point judgment value meets a first preset range, and screening all first edge pixel points in the first image;

the second screening module is used for acquiring the minimum distance between each first edge pixel point and the rest first edge pixel points, and deleting the corresponding first edge point when the minimum distance is greater than a first preset threshold value;

the contour extraction module is used for connecting the reserved first edge pixel points based on the geometric features of the target lithium battery to obtain the contour of the target lithium battery in the first image;

the area dividing module is used for dividing a first area surrounded in the outline of the target lithium battery into a bright area and a dark area;

the second acquisition module is used for acquiring first average brightness based on brightness values of all first pixel points contained in the bright area, acquiring second average brightness based on brightness values of all second pixel points contained in the dark area, and acquiring a light-to-dark area ratio of the target lithium battery in the first image based on the total number of the first pixel points and the total number of the second pixel points;

the first average brightness, the second average brightness and the light-dark area ratio are first image parameters of the first image.

Preferably, the machine vision detection system for automatically detecting parameters of lithium batteries on a conveyor belt comprises:

the connecting module is used for connecting the reserved first edge pixel points based on the geometric characteristics of the target lithium battery to obtain a first outline of the target lithium battery in the first image;

the first judging module is used for obtaining a first pixel judgment value based on the pixel values of all reserved first edge pixel points, obtaining a second pixel judgment value based on the pixel values of all third pixel points in the area surrounded by the first outline, obtaining a third pixel judgment value based on the pixel values of all fourth pixel points in the area outside the first outline, and taking the first outline as the outline of the target lithium battery in the first image when a first difference value between the first pixel judgment value and the second pixel judgment value and a second difference value between the first pixel judgment value and the third pixel judgment value are all larger than a second preset threshold value;

otherwise, judging that the first outline is unqualified, and re-screening the first edge pixel points based on a second preset range until the obtained first outline is judged to be qualified;

and the correction module is used for filling and smoothing the qualified first contour by adopting a preset correction algorithm to obtain the contour of the target lithium battery in the first image.

the first determining module is used for converting a first area in the first image into a corresponding multi-channel image, performing high-pass filtering based on the gray value and the brightness value of each fifth pixel point in the multi-channel image to obtain a first frequency domain graph and a second frequency domain graph, performing inverse fitting on the first frequency domain graph and the second frequency domain graph to obtain a fitted frequency domain graph, determining an abrupt change area in the frequency domain graph, dividing the frequency domain graph into a low-value frequency domain graph and a high-value frequency domain graph based on the abrupt change area, taking the fifth pixel point corresponding to the high-value frequency domain graph as a sixth pixel point contained in a bright area, and taking the fifth pixel point corresponding to the low-value frequency domain graph as a seventh pixel point contained in a dark area;

a second determining module, configured to obtain a second luminance value matrix corresponding to a fifth pixel point remaining in the first region in the first image except for the sixth pixel point and the seventh pixel point, calculate a first determination value based on the second luminance value matrix and a preset second same-order transformation matrix, when the first determination value is within a third preset range, use a corresponding fifth pixel point as an eighth pixel point included in a bright region, and when the first determination value is within a fourth preset range, use a corresponding fifth pixel point as a ninth pixel point included in a dark region;

a third determining module, configured to perform, based on a gray value, shooting and sorting on remaining fifth pixel points in a first region of the first image except for the sixth pixel point, the seventh pixel point, the eighth pixel point, and the ninth pixel point to obtain a gray value sequence, divide the gray value sequence into a plurality of gray value subsequences according to a preset division rule, calculate a first average gray value of each gray value subsequence, calculate a second average gray value of all pixel points included in a determined bright region and a third average gray value of all pixel points included in a determined dark region, calculate a third difference value between the first average gray value and the second average gray value, and a fourth difference value between the first average gray value and the third average gray value, when the third difference value is smaller than the fourth difference value, use all pixel points corresponding to the gray value subsequences as tenth pixel points included in the bright region, otherwise, use all pixel points corresponding to the gray value subsequences as eleventh pixel points included in the bright region;

and the division submodule is used for dividing the area formed by the sixth pixel point, the eighth pixel point and the tenth pixel point into the bright area and dividing the area formed by the seventh pixel point, the ninth pixel point and the eleventh pixel point into the dark area.

Preferably, the machine vision detection system for automatically detecting the parameters of the lithium battery on the conveyor belt comprises:

the contrast module is used for obtaining a first ratio of the first average brightness to a first preset average brightness, a second ratio of the second average brightness to a second preset average brightness, and a fifth difference value of the light-dark area ratio and the preset light-dark area ratio, which correspond to each image acquisition device, based on the correlation result;

the setting sub-module is used for obtaining exposure brightness setting parameters of each image acquisition device based on a first ratio and a second ratio corresponding to each image acquisition device and a preset exposure brightness corresponding relation, generating a corresponding first setting instruction, obtaining exposure range setting parameters of each image acquisition device based on a fifth difference corresponding to each image acquisition device and the preset exposure range corresponding relation, and generating a corresponding second setting instruction;

the creating module is used for creating a communication link between the parameter calculating module and each image acquisition device;

the first judgment module is used for judging whether the communication link is established successfully or not, if so, judging whether each image acquisition device obtains the communication authorization of the setting submodule or not, and if so, sending a corresponding first setting instruction and a second setting instruction to the image acquisition device;

otherwise, sending out a communication failure prompt message to the monitoring center terminal, and reestablishing a communication link;

the second judgment module is further used for controlling the corresponding image acquisition equipment to send a communication authorization request instruction to the setting submodule when the communication link is not established successfully, receiving a feedback instruction of the setting submodule based on the communication authorization request instruction, obtaining the communication authorization of the setting submodule based on the feedback instruction, and sending a corresponding first setting instruction and a second setting instruction by the image acquisition equipment;

and the third judging module is used for stopping recreating the communication link when the number of times that the image acquisition equipment recreates the communication link exceeds the preset number of times, and sending prompt information that the communication link between the parameter calculating module and the corresponding image acquisition equipment cannot be created to the monitoring center terminal.

the first processing module is used for enhancing the contrast and the definition of the target image;

the image synthesis module is used for extracting first edge features and second edge features of all target images based on a preset algorithm, generating corresponding first edge feature vectors and second edge feature vectors based on the edge features, selecting any target image as a reference image, matching the edge feature vectors of the remaining target images with the first edge feature vectors to obtain corresponding first matching coefficients, connecting the first edges of the target images with the edges of the remaining target images corresponding to the highest first matching coefficients, matching the edge feature vectors of the remaining target images with the second edge feature vectors to obtain corresponding second matching coefficients, connecting the second edges of the target images with the edges of the remaining target images corresponding to the highest second matching coefficients to obtain new reference images, and screening and connecting the new reference images until all the target images are synthesized into an image to be processed;

and the second processing module is used for carrying out detail fusion on the image to be processed based on a preset fusion processing algorithm to obtain the image to be detected of the target lithium battery.

the third processing module is used for performing gray value conversion on the to-be-detected image based on a preset proportional gradient to obtain a preset number of frames of to-be-detected gray image, setting a resolution range based on the area ratio range of the target lithium battery in the target image, generating a resolution gradient based on the resolution range and a preset resolution gradient interval, and generating a to-be-detected gray image set according to the resolution gradient on the to-be-detected gray image;

and the characteristic extraction module is used for extracting lithium battery characteristic data corresponding to each image in the to-be-detected gray image set based on the characteristic classification neural network, converting the lithium battery characteristic data into lithium battery standard characteristic data corresponding to the image format of the to-be-detected image based on the ratio of the resolution ratio and the original resolution ratio of each image in the to-be-detected gray image set, and forming a standard characteristic data set.

the data screening module is used for screening the standard characteristic data set based on a preset characteristic data screening neural network to obtain a parameter to be measured of the target lithium battery;

the detection submodule is used for comparing the parameter to be detected of the target lithium battery with a preset parameter range, judging that the lithium battery is qualified if the parameter to be detected is within the preset parameter range, and sending a judgment result and the parameter to be detected to a monitoring center terminal;

otherwise, the disqualification of the lithium battery is judged, and the judgment result and the parameter to be detected are sent to a monitoring center terminal.

a difference value obtaining module, configured to obtain a maximum difference value between the parameter to be measured and corresponding feature data in the standard feature data set;

and the precision setting module is used for setting the feature extraction precision of the feature classification neural network based on the maximum difference value and a preset precision setting list.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:

FIG. 1 is a machine vision inspection system for automatically inspecting lithium battery parameters on a conveyor belt according to an embodiment of the present invention;

FIG. 2 is a block diagram of an image capture module according to an embodiment of the present invention;

FIG. 3 is a block diagram of a first acquisition module according to an embodiment of the present invention;

FIG. 4 is a block diagram of a contour extraction module according to an embodiment of the present invention;

FIG. 5 is a block diagram of a region partitioning module according to an embodiment of the present invention;

FIG. 6 is a block diagram of a setup module according to an embodiment of the present invention;

FIG. 7 is a diagram of a configuration module according to an embodiment of the present invention;

FIG. 8 is a block diagram of an image parsing module according to an embodiment of the present invention;

FIG. 9 is a diagram of a parameter acquisition module according to an embodiment of the present invention;

fig. 10 is a block diagram of a data screening module according to an embodiment of the present invention.

Detailed Description

The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.

Example 1:

the invention provides a machine vision detection system for automatically detecting lithium battery parameters on a conveyor belt, as shown in figure 1, comprising:

the image acquisition module is used for continuously acquiring images of a target lithium battery moving at a preset position according to a preset period to obtain a group of target images;

and the parameter acquisition module is used for acquiring the parameters to be measured of the target lithium battery based on the analysis result.

In this embodiment, the image acquisition module is implemented by each image acquisition device, and the image acquisition device is a line camera.

In this embodiment, the preset position is a position on the conveyor belt within the shooting range of each image capturing apparatus.

In this embodiment, the target image is an image of the target lithium battery that continuously acquires, in a preset period, the images of the target lithium battery that rotates at a preset position after the shooting parameters are set.

In this embodiment, the target lithium battery is the lithium battery to be tested.

In this embodiment, the image to be measured is used to obtain the parameter to be measured of the target lithium battery.

In this embodiment, the analysis result is a feature data set of the target lithium battery, where the feature data set includes parameters to be measured of the target lithium battery, for example: location, morphology, size of the target lithium battery.

In the embodiment, the machine vision detection system for automatically detecting the parameters of the lithium battery on the conveyor belt is realized by SCI vision software.

The beneficial effects of the above technical scheme are: compared with an artificial size detection method, the industrial size detection method based on machine vision does not produce visual fatigue, is high in accuracy and reliability, does not need to be familiar with products in advance during model changing, does not need labor cost, can meet production requirements, saves labor cost, has strong accuracy, high efficiency and flexibility, can photograph moving objects by adopting image acquisition equipment to form images, can accurately position the position, form, size and other characteristics of the lithium battery under the condition that the lithium battery does not stop moving, and greatly improves detection efficiency.

Example 2:

based on embodiment 1, the image capturing module, as shown in fig. 2, includes:

wherein the image parameters include: a light-to-dark area ratio, a first average brightness corresponding to a bright region, and a second average brightness corresponding to a dark region;

wherein the shooting parameters include: exposure range, exposure brightness.

In this embodiment, the first image is directly captured by each image capturing apparatus.

In this embodiment, the device identifier is a serial number identifier of each image capturing device.

In this embodiment, the sample image is an image obtained when the corresponding image acquisition device acquires the sample lithium battery at the preset position.

In this embodiment, the shooting parameter correspondence relationship is a correspondence relationship between a first ratio of the first average brightness to the first preset average brightness, a second ratio of the second average brightness to the second preset average brightness, and the exposure brightness, which correspond to each image capturing device, and a correspondence relationship between a fifth difference value between a light-dark area ratio and a preset light-dark area ratio, which correspond to each image capturing device, and an exposure range.

In this embodiment, the bright area is an area with high brightness selected as required in the first image, and the dark area is an area with low brightness selected as required in the first image.

The beneficial effects of the above technical scheme are: the method comprises the steps of acquiring a first image through image acquisition on a target lithium battery, comparing the first image parameter of the first image with a second image parameter of a sample image acquired when corresponding image acquisition equipment acquires a sample lithium battery at a preset position to acquire a shooting parameter corresponding to the image acquisition equipment, setting the shooting parameter of the image acquisition equipment according to actual environmental conditions before acquiring an image to be detected of the target lithium battery, automatically adjusting the shooting parameter of each image acquisition equipment according to the environmental conditions, omitting the step of manually debugging the image acquisition equipment, and avoiding the situation that the acquired image to be detected is not clear and the parameter to be detected can not be acquired.

Example 3:

based on embodiment 2, the first obtaining module, as shown in fig. 3, includes:

the matrix acquisition module is used for acquiring pixel values of all pixel points in the first image, selecting a non-edge pixel point in the first image, acquiring a gray scale matrix of the first image with the non-edge pixel point as the center, and calculating a first brightness value matrix of the first image with the non-edge pixel point as the center based on the pixel values of all the pixel points and the gray scale matrix;

the second obtaining module is used for obtaining first average brightness based on brightness values of all first pixel points contained in the bright area, obtaining second average brightness based on brightness values of all second pixel points contained in the dark area, and obtaining a light-dark area ratio of the target lithium battery in the first image based on the total number of the first pixel points and the total number of the second pixel points;

In this embodiment, the non-edge pixels are pixels in the first image except the edge pixels, and the edge pixels are pixels in the outermost periphery of the first image.

In this embodiment, the gray scale matrix is the gray scale of other pixels whose values at the corresponding pixel positions are centered on the non-edge pixel, and the corresponding value at the non-edge pixel is 0.

In this embodiment, the first luminance matrix is the value at the corresponding pixel position, which is the value at the corresponding position in the gray-scale matrix multiplied by the pixel value of the corresponding pixel.

In this embodiment, the first same-order transformation matrix is a preset matrix having the same order as the first luminance value matrix, and the edge point determination value of the corresponding pixel point can be obtained by multiplying the first same-order transformation matrix by the first luminance value matrix of the corresponding pixel point.

In this embodiment, the first preset range is an edge point determination value range corresponding to an edge pixel point of the preset target lithium battery.

In this embodiment, the first preset threshold is the maximum distance between adjacent pixels on the contour line of the target lithium battery, and if the minimum distance between each first edge pixel and the remaining first edge pixels is greater than the maximum distance between adjacent pixels on the contour line of the target lithium battery, the corresponding first edge pixel is not a pixel on the contour line of the target lithium battery, and is therefore deleted.

In this embodiment, the geometric feature of the target lithium battery is the profile feature of the lithium battery stored in advance.

In this embodiment, the first pixel point is a pixel point included in the bright area.

In this embodiment, the second pixel is a pixel included in the dark region.

The beneficial effects of the above technical scheme are: the edge point judgment value of the corresponding pixel point is obtained based on the pixel value, the gray scale and the brightness value of each pixel point on the first image, so that possible pixel points on the contour line of the target lithium battery in the first image are determined, the pixel points on the contour line of the target lithium battery are screened, the pixel values, the gray scales and the brightness values of the pixel points are considered, more accurate screening is achieved, then the minimum distance between each first edge pixel point and the rest first edge pixel points is compared with a first preset threshold value, the first edge pixel points far away from the contour line can be accurately deleted, and the contour of the target lithium battery can be conveniently and accurately constructed.

Example 4:

based on embodiment 3, the contour extraction module, as shown in fig. 4, includes:

the connection module is used for connecting the reserved first edge pixel points based on the geometric characteristics of the target lithium battery to obtain a first outline of the target lithium battery in the first image;

the first judging module is used for obtaining a first pixel judging value based on the pixel values of all reserved first edge pixel points, obtaining a second pixel judging value based on the pixel values of all third pixel points in an area surrounded by the first outline, obtaining a third pixel judging value based on the pixel values of all fourth pixel points in an area outside the first outline, and taking the first outline as the outline of the target lithium battery in the first image when a first difference value between the first pixel judging value and the second pixel judging value and a second difference value between the first pixel judging value and the third pixel judging value are larger than a second preset threshold value;

otherwise, judging that the first contour is unqualified, and re-screening the first edge pixel points based on a second preset range until the obtained first contour is judged to be qualified;

In this embodiment, the first contour is a contour line of the target lithium battery obtained primarily based on the retained geometric feature connection between the first edge pixel point and the target lithium battery.

In this embodiment, the third pixel point is a pixel point in the area surrounded by the first contour.

In this embodiment, the fourth pixel point is a pixel point included in the first contour region.

In this embodiment, the first pixel determination value is a pixel average value of all the first edge pixels.

In this embodiment, the second pixel determination value is the average pixel value of all the third pixels.

In this embodiment, the third pixel determination value is the average pixel value of all the fourth pixels.

In this embodiment, the second preset range is an edge point determination value range corresponding to an edge pixel point of a preset target lithium battery smaller than the first preset range.

The beneficial effects of the above technical scheme are: by comparing the pixel average value of the first edge pixel point on the contour line with the difference value of the pixel average values of all third pixel points in the first contour inclusion region and the pixel average values of all fourth pixel points in the region outside the first contour, whether the first contour is qualified or not is further verified in the aspect of the region pixel value of the first image, and the first contour can be further ensured to be qualified.

Example 5:

based on embodiment 3, as shown in fig. 5, the region dividing module includes:

the first determining module is used for converting a first area in the first image into a corresponding multichannel image, performing high-pass filtering based on the gray value and the brightness value of each fifth pixel point in the multichannel image to obtain a first frequency domain graph and a second frequency domain graph, performing reverse fitting on the first frequency domain graph and the second frequency domain graph to obtain a fitted frequency domain graph, determining an abrupt change area in the frequency domain graph, dividing the frequency domain graph into a low-value frequency domain graph and a high-value frequency domain graph based on the abrupt change area, taking the fifth pixel point corresponding to the high-value frequency domain graph as a sixth pixel point contained in a bright area, and taking the fifth pixel point corresponding to the low-value frequency domain graph as a seventh pixel point contained in a dark area;

a third determining module, configured to perform, based on gray values, shooting and sorting on remaining fifth pixel points in a first region of the first image except for the sixth pixel point, the seventh pixel point, the eighth pixel point, and the ninth pixel point to obtain a gray value sequence, divide the gray value sequence into a plurality of gray value subsequences according to a preset division rule, calculate a first average gray value of each gray value subsequence, calculate a second average gray value of all pixel points included in a determined bright region and a third average gray value of all pixel points included in a determined dark region, calculate a third difference value between the first average gray value and the second average gray value, and a fourth difference value between the first average gray value and the third average gray value, when the third difference value is smaller than the fourth difference value, use all pixel points corresponding to the gray value subsequences as tenth pixel points included in the bright region, otherwise, use all pixel points corresponding to the gray value subsequences as eleventh pixel points included in the bright region;

In this embodiment, converting the first region in the first image into a corresponding multi-channel image is: and converting the first image into a plurality of single-channel images, and fusing the single-channel images into a multi-channel image.

In this embodiment, the fifth pixel points are all the pixel points in the multi-channel image.

In this embodiment, the first frequency domain map is a frequency domain map obtained by performing high-pass filtering on the gray value of the multi-channel image.

In this embodiment, the second frequency domain map is a frequency domain map obtained by high-pass filtering the luminance values of the multi-channel image.

In this embodiment, the fitted frequency domain graph is a frequency domain graph obtained by adding a value corresponding to the fifth pixel point on the frequency domain graph obtained by performing positive-negative conversion on the first frequency domain graph to a value corresponding to the fifth pixel point on the second frequency domain graph.

In this embodiment, the abrupt change region is a region where the abrupt change occurs in the fitted frequency domain map.

In this embodiment, the low-value frequency domain map is a partial frequency domain map with a lower frequency domain value of the region of the fitted frequency domain map except for the abrupt change region.

In this embodiment, the high-value frequency domain map is a partial frequency domain map with higher frequency domain values of the regions except for the abrupt change region in the fitted frequency domain map.

In this embodiment, the second same-order transformation matrix is a matrix of the same order as the second luminance value matrix, and the first determination value corresponding to the fifth pixel point can be obtained by multiplying the second same-order transformation matrix by the second luminance value matrix corresponding to the fifth pixel point.

In this embodiment, the first determination value is a determination value for determining whether the fifth pixel belongs to a bright area or a dark area.

In this embodiment, the third predetermined range is a first determination value range corresponding to the fifth pixel point belonging to the bright area.

In this embodiment, the sixth pixel point is a pixel point included in a bright area determined according to the frequency domain diagram;

and the seventh pixel point is the pixel point contained in the dark area judged according to the frequency domain graph.

In this embodiment, the eighth pixel point is a pixel point included in the bright area determined according to the determination value;

the ninth pixel point is the pixel point contained in the dark area judged according to the judgment value.

In this embodiment, the gray value sequence is obtained by arranging gray values of remaining fifth pixels, except for the sixth pixel, the seventh pixel, the eighth pixel, and the ninth pixel, in the first region of the first image from small to large.

In this embodiment, the gray value sub-sequence is obtained by dividing the gray value sequence according to a preset gray value interval.

In this embodiment, the tenth pixel point is a pixel point included in the bright area determined according to the average gray value difference;

and the eleventh pixel point is a pixel point contained in the dark area determined according to the average gray value difference value.

The beneficial effects of the above technical scheme are: the bright area and the dark area in the target lithium battery area in the first image can be accurately determined by three methods of comparing the frequency domain graph obtained by high-pass filtering the brightness value and the gray value, the calculated judgment value and the gray average value, the situation that the pixel points are screened by using one method and are not considered thoroughly is avoided, the subsequent accurate calculation of the first average brightness, the second average brightness and the light-dark area ratio is facilitated, and further, the more accurate setting of the shooting parameters is facilitated.

Example 6:

based on embodiment 2, the setting module, as shown in fig. 6, includes:

the contrast module is used for obtaining a first ratio of the first average brightness to the first preset average brightness, a second ratio of the second average brightness to the second preset average brightness, and a fifth difference value of the light-dark area ratio to the preset light-dark area ratio, which correspond to each image acquisition device, based on the correlation result;

the first judgment module is used for judging whether the communication link is established successfully or not, if so, judging whether each image acquisition device obtains the communication authorization of the setting sub-module or not, and if so, sending a corresponding first setting instruction and a second setting instruction to the image acquisition device;

the second judging module is further used for controlling the corresponding image acquisition equipment to send a communication authorization request instruction to the setting submodule when the communication link is not successfully established, receiving a feedback instruction of the setting submodule based on the communication authorization request instruction, obtaining the communication authorization of the setting submodule based on the feedback instruction, and sending a corresponding first setting instruction and a corresponding second setting instruction by the image acquisition equipment;

and the third judging module is used for stopping the reestablishment of the communication link when the number of times of reestablishing the communication link by the image acquisition equipment exceeds the preset number of times, and sending prompt information that the communication link between the parameter calculating module and the corresponding image acquisition equipment cannot be established to the monitoring center terminal.

In this embodiment, the preset exposure brightness correspondence relationship is a correspondence relationship between the first ratio, the second ratio, and the exposure brightness of the corresponding image capturing device.

In this embodiment, the first setting instruction is used to instruct the corresponding image capturing device to set the corresponding exposure brightness.

In this embodiment, the exposure range corresponding relationship is a corresponding relationship between the fifth difference and the exposure range of the corresponding image capturing device.

In this embodiment, the second setting instruction is used to instruct the corresponding image capturing device to set the corresponding exposure range.

The beneficial effects of the above technical scheme are: the method comprises the steps of obtaining an image parameter comparison result between a first image and a sample image through calculation, setting shooting parameters of corresponding image acquisition equipment based on the comparison result, realizing the setting of the shooting parameters of each image acquisition equipment according to the field situation, generating corresponding parameter setting instructions based on the set shooting parameters, and detecting a communication link and communication authorization to ensure that the parameter setting instructions can be transmitted safely and completely.

Example 7:

based on embodiment 1, the image processing module, as shown in fig. 7, includes:

In this embodiment, the first edge feature and the second edge feature are image feature parameters corresponding to the first edges and image feature parameters of the second edges of all the target images, and the first edge and the second edge are two edges of the target image that need to be connected with other target images.

In this embodiment, the first edge feature vector is a feature vector parameter corresponding to the first edge feature;

the second edge feature vector is a feature vector parameter corresponding to the second edge feature.

In this embodiment, the matching coefficient calculation process is as follows: and dividing the difference value between a group of feature vectors by the ratio value obtained by the feature vectors, and subtracting the ratio value to obtain the difference value which is the matching coefficient.

In this embodiment, the fusion processing algorithm is a preset image processing algorithm for smoothly filling the image to be processed.

The beneficial effects of the above technical scheme are: extracting edge features from the target image, matching and connecting the edges of all the target images based on the image edge features, and then performing fusion processing to obtain an image to be detected of the target lithium battery, so that an image which embodies the overall appearance of the target lithium battery can be obtained, and a data material is provided for further extracting the image to be detected of the target lithium battery.

Example 8:

based on embodiment 1, as shown in fig. 8, the image analysis module includes:

the third processing module is used for performing gray value conversion on the to-be-detected image based on a preset proportional gradient to obtain to-be-detected gray images of a preset number of frames, setting a resolution range based on the area ratio range of the target lithium battery in the target image, generating a resolution gradient based on the resolution range and a preset resolution gradient interval, and generating a to-be-detected gray image set on the to-be-detected gray image according to the resolution gradient;

and the characteristic extraction module is used for extracting the lithium battery characteristic data corresponding to each image in the to-be-detected gray image set based on the characteristic classification neural network, and converting the lithium battery characteristic data into the lithium battery standard characteristic data corresponding to the image format of the to-be-detected image based on the ratio of the resolution ratio and the original resolution ratio of each image in the to-be-detected gray image set to form a standard characteristic data set.

In this embodiment, the preset proportional gradient is a preset ratio, for example: 1/5, 2/5, 3/5, 4/5 and 5/5.

In this embodiment, the to-be-detected grayscale image is an image obtained by respectively converting the to-be-detected image into an original grayscale value multiplied by a preset proportional gradient.

In this embodiment, a resolution range is set based on an area ratio range of the target lithium battery in the target image, that is, the resolution range corresponding to the target image is determined according to a relationship between the area ratio and the resolution range.

In this embodiment, the characteristic classification neural network is a preset neural network for extracting characteristic data of a lithium battery in the grayscale image to be detected.

In this embodiment, the lithium battery characteristic data includes: position characteristic data, form characteristic data and size characteristic data of the lithium battery.

In this embodiment, the set of to-be-detected grayscale images is an image obtained by setting the resolution of each to-be-detected grayscale image according to the resolution gradient.

In this embodiment, the standard feature data set is a set composed of standard feature data, and the standard brick feature data is feature data corresponding to an image format of a to-be-detected image converted from lithium battery feature data corresponding to each image in the to-be-detected grayscale image set based on a ratio of a resolution to an original resolution of each image in the to-be-detected grayscale image set.

The beneficial effects of the above technical scheme are: the method comprises the steps of carrying out gray value conversion on an image to be detected to obtain a gray image set to be detected, extracting and converting lithium battery characteristic data corresponding to each image in the gray image set to be detected based on a characteristic classification neural network, and achieving more accurate and comprehensive extraction of the characteristic data of a target lithium battery.

Example 9:

based on embodiment 1, as shown in fig. 9, the parameter obtaining module includes:

the data screening module is used for screening the standard characteristic data set based on a preset characteristic data screening neural network to obtain the to-be-tested parameters of the target lithium battery;

otherwise, judging that the lithium battery is unqualified, and sending the judgment result and the parameter to be detected to a monitoring center terminal.

In this embodiment, the parameters to be measured include position coordinate data, shape data, and size data of the target lithium battery.

The beneficial effects of the above technical scheme are: and screening the standard characteristic data set based on a preset characteristic data screening neural network to obtain the parameters to be tested of the target lithium battery, further judging whether the parameters to be tested are qualified, and transmitting the parameters to a monitoring center terminal, so that the full automation of obtaining the target lithium battery is realized, and the functions of remote monitoring and obtaining the parameters to be tested can also be realized.

Example 10:

based on embodiment 9, the data filtering module, as shown in fig. 10, includes:

In this embodiment, setting the feature extraction precision of the feature classification neural network based on the maximum difference and a preset precision setting list includes:

dividing the standard characteristic data set into standard characteristic data subsets corresponding to each parameter to be measured;

based on all the characteristic parameters and the corresponding parameters to be measured in the currently calculated standard characteristic data subset, calculating the maximum difference value between the characteristic parameters and the corresponding parameters to be measured after normalization processing in the currently calculated standard characteristic data subset:

in the formula, i is the ith characteristic parameter in the current calculated standard characteristic data subset, MAX is the maximum difference value between all characteristic parameters after normalization processing in the current calculated standard characteristic data subset and the corresponding parameter to be measured, and x _i For the ith feature parameter before normalization processing in the currently calculated standard feature data subset, (x) _i ) _min For the smallest feature parameter in the currently computed subset of standard feature data, (x) _i ) _max Is the maximum characteristic parameter in the current calculated standard characteristic data subset, and x is the parameter to be measured corresponding to the standard characteristic data subset (the) _max Taking the maximum value; m represents the total number of the characteristic parameters in the standard characteristic data subset;

calculating the feature extraction precision to be set based on the relationship coefficient between the maximum difference between the normalized feature parameter and the corresponding parameter to be measured in each standard feature data subset and the feature extraction precision before setting:

wherein M' is the feature extraction precision to be set, j is the jth standard feature data subset, n is the total number of the standard feature data subsets, and alpha _j Corresponding to the jth standard characteristic data subsetCoefficient of relationship between large difference and feature extraction accuracy, MAX _j Taking the maximum difference value of all characteristic parameters after normalization processing in the jth standard characteristic data subset and the corresponding parameters to be detected, and taking M as the characteristic extraction precision before setting;

for example: the first subset of standard feature data comprises the following feature parameters: 90. 60, 75, the relation coefficient between the maximum difference value and the feature extraction precision which are preset for the parameter to be measured corresponding to the standard feature data subset being 5 is 0.3, and the second standard feature data subset comprises the following feature parameters: 2. 4, 3, the parameter to be measured corresponding to the standard feature data subset is 3, the preset relation coefficient between the maximum difference value and the feature extraction precision is 0.2, and the feature extraction precision M before setting is 0.1;

the first subset of standard feature data contains the feature parameters: the result after 90 normalization was x' ₁ =1; the result after 60 normalization was x' ₁ =0; the result after 75 normalization treatment was x' ₁ =0.5, and the maximum difference between all the normalized feature parameters in the currently calculated standard feature data subset and the corresponding parameters to be measured is 5;

the second subset of standard feature data contains the feature parameters: 2 the result after normalization treatment is x' ₁ =0;4 is x 'after normalization treatment' ₁ =1; the result after 3 normalization treatment was x' ₁ =0.5, and the maximum difference between all the normalized feature parameters in the currently calculated standard feature data subset and the corresponding parameters to be measured is 3;

the feature extraction accuracy to be set is about 0.10393.

In this embodiment, the preset precision setting list is a corresponding relationship between the maximum difference and the extraction precision.

The beneficial effects of the above technical scheme are: and setting the feature extraction precision of the feature classification neural network according to the maximum difference value of the acquired parameters to be detected and the corresponding feature data in the standard feature data set and a preset precision setting list, so that the automatic training and learning of the feature data screening neural network are realized, and the standard feature data set can be screened more accurately in the next screening process.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims

1. A machine vision inspection system for automatically detecting lithium battery parameters on a conveyor belt, comprising:

the image processing module is used for preprocessing the target image and synthesizing to obtain an image to be detected of the target lithium battery;

the parameter acquisition module is used for acquiring the to-be-detected parameters of the target lithium battery based on the analysis result;

the image acquisition module comprises:

wherein the shooting parameters include: exposure range, exposure brightness;

the first obtaining module includes:

the contour extraction module is used for connecting the reserved first edge pixel points based on the geometric characteristics of the target lithium battery to obtain the contour of the target lithium battery in the first image;

2. The system of claim 1, wherein the contour extraction module comprises:

the first judging module is used for obtaining a first pixel judgment value based on the pixel values of all reserved first edge pixel points, obtaining a second pixel judgment value based on the pixel values of all third pixel points in an area surrounded by the first outline, obtaining a third pixel judgment value based on the pixel values of all fourth pixel points in an outer area of the first outline, and taking the first outline as the outline of the target lithium battery in the first image when a first difference value between the first pixel judgment value and the second pixel judgment value and a second difference value between the first pixel judgment value and the third pixel judgment value are all larger than a second preset threshold value;

3. The system of claim 1, wherein the zone partitioning module comprises:

4. The system of claim 1, wherein the setup module comprises:

the creating module is used for creating a communication link between the setting sub-module and each image acquisition device;

the first judging module is used for judging whether the communication link is established successfully or not, if so, judging whether each image acquisition device obtains the communication authorization of the setting submodule or not, and if so, sending a corresponding first setting instruction and a second setting instruction to the image acquisition device;

when the communication link is not successfully established or the image acquisition equipment does not obtain the communication authorization of the setting sub-module, sending a communication failure prompt message to a monitoring center terminal and reestablishing the communication link;

and the third judgment module is used for stopping the reestablishment of the communication link when the number of times of reestablishing the communication link by the image acquisition equipment exceeds the preset number of times, and sending prompt information that the communication link between the setting sub-module and the corresponding image acquisition equipment cannot be established to the monitoring center terminal.

5. The system of claim 1, wherein the image processing module comprises:

6. The system of claim 1, wherein the image analysis module comprises:

7. The system of claim 6, wherein the parameter acquisition module comprises:

the detection submodule is used for comparing the parameter to be detected of the target lithium battery with a preset parameter range, judging that the target lithium battery is qualified if the parameter to be detected is within the preset parameter range, and sending a judgment result and the parameter to be detected to a monitoring center terminal;

8. The system of claim 7, wherein the data screening module comprises: