WO2021169335A1 - 一种锂电池极耳激光焊点视觉在线检测方法 - Google Patents

一种锂电池极耳激光焊点视觉在线检测方法 Download PDF

Info

Publication number
WO2021169335A1
WO2021169335A1 PCT/CN2020/122843 CN2020122843W WO2021169335A1 WO 2021169335 A1 WO2021169335 A1 WO 2021169335A1 CN 2020122843 W CN2020122843 W CN 2020122843W WO 2021169335 A1 WO2021169335 A1 WO 2021169335A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
pcm
tab
ear
battery
Prior art date
Application number
PCT/CN2020/122843
Other languages
English (en)
French (fr)
Inventor
陈忠
郝煜亚
谢声扬
陈伯乐
郭光明
张宪民
Original Assignee
华南理工大学
广东东博自动化设备有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华南理工大学, 广东东博自动化设备有限公司 filed Critical 华南理工大学
Publication of WO2021169335A1 publication Critical patent/WO2021169335A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84Systems specially adapted for particular applications
    • G01N21/88Investigating the presence of flaws or contamination
    • G01N21/8851Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01BMEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00Measuring arrangements characterised by the use of optical techniques
    • G01B11/002Measuring arrangements characterised by the use of optical techniques for measuring two or more coordinates
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84Systems specially adapted for particular applications
    • G01N21/88Investigating the presence of flaws or contamination
    • G01N21/95Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/0002Inspection of images, e.g. flaw detection
    • G06T7/0004Industrial image inspection
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/20Image preprocessing
    • G06V10/25Determination of region of interest [ROI] or a volume of interest [VOI]
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84Systems specially adapted for particular applications
    • G01N21/88Investigating the presence of flaws or contamination
    • G01N21/8851Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8887Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10004Still image; Photographic image
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30108Industrial image inspection
    • G06T2207/30152Solder

Definitions

  • the invention relates to the field of laser solder joint detection in a lithium battery automatic packaging process, and in particular to a visual online detection method for laser solder joints of lithium battery tabs.
  • the defect detection of the solder joints of the pole ears of lithium battery cells and the PCM plate pole ears adopts traditional visual inspection algorithms.
  • the main algorithms are: shape template matching, spot analysis, contour matching, etc.; preprocessing algorithms usually include: grayscale linear transformation, Corrosion, expansion, etc.
  • the traditional visual inspection method cannot effectively solve the inspection problem caused by the folds of the lithium battery tabs
  • manual visual inspection of the solder joints of the lithium battery is still widely used.
  • the manual visual inspection method has disadvantages such as low efficiency and unstable effect. Therefore, it is very important to propose an effective visual solder joint inspection method for the deterioration of image quality caused by the wrinkle and bending of the lithium battery tabs.
  • the purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for visual online detection of the laser solder joints of the tabs of the lithium battery.
  • the invention overcomes the problem that the traditional detection method is difficult to detect the wrinkle lug solder joints, and at the same time meets the requirements of detection accuracy and online detection.
  • the invention selects a specific detection algorithm according to the contrast of the tab image, and reduces the detection ROI area according to the positioning relationship.
  • a visual on-line detection method for laser solder joints of lithium battery tabs including the steps:
  • PCM circuit protection module
  • solder joint detection in a square ROI. If a solder joint is detected, the estimated solder joint coordinates will be compensated to obtain the actual solder joint coordinates in the image of the electrode ear of the battery.
  • two sets of dome light sources and the black and white industrial camera module are used to perform a global image of the cell tab and the PCM plate tab of the lithium battery. collection.
  • the light source adopts blue light illumination; two industrial cameras are installed opposite to each other, and the lithium battery is located between the two industrial cameras during image collection; the two cameras separately collect the global image of the PCM plate tab and the global image of the battery cell tab; the tab The global image covers the left and right two pole ears; the two industrial camera modules (camera and lens) are exactly the same, and the overlapping area of the field of view needs to completely cover the left and right pole ears; the imaging effect that should be achieved by combining the camera exposure time and the light intensity of the light source is Make the gray value of the normal non-wrinkle tab in the image 255.
  • the cell tab global image is divided into a cell left tab local image and a cell right tab local image;
  • PCM The global image of the pole ear is divided into a partial image of the left pole of the PCM board and a partial image of the right pole of the PCM board.
  • Each partial image of the tab is required to completely contain the contour of the tab, and the partial image of the tab should cover all the positions of the tab of the lithium battery in the global image that may appear due to the lithium battery clamping error.
  • the step of judging the contrast of the obtained partial image of the tabs of the lithium battery includes:
  • the gray level of the background area far from the light source irradiated by the lithium battery is in the low gray value range, and the range is about 0 to 20;
  • the gray scale of the overall lug shadow caused by the shadow or the bending and tilt of the lug is in the middle gray value range, and the range is about 30 to 80; there is no wrinkle in the local image of the lug, and the grayscale of the non-welded area of the normal lug is in a high grayscale.
  • Value range the range is approximately 200 to 255;
  • is greater than or equal to the preset constant const, the partial image of the polar ear is judged as normal contrast, otherwise it is low contrast; the constant const is calculated by comparing the sum of the number of pixels in the high gray value interval of a normal polar ear image with no wrinkles and medium gray The quotient of the sum of the number of pixels in the degree value interval is used as a reference, and the average value can be continuously accumulated in real-time operation to reduce the error.
  • the tab ROI extracted from the partial image of the tab needs to meet: the left edge of the ROI coincides with the left edge of the contour of the tab , The right edge of the ROI coincides with the right edge of the ear contour, and the lower edge of the ROI coincides with the lower edge of the ear contour.
  • the step of extracting the lug ROI from the local image of the lug by adopting a corresponding method according to the determined contrast result includes:
  • the OTSU algorithm is used for thresholding, and the method of finding the smallest circumscribed rectangle is adopted in the binary image.
  • the rectangle with the largest area among all circumscribed rectangles is the pole ear ROI;
  • the cumulative gray pixel number approximation method is used for thresholding, and the method of finding the smallest circumscribed rectangle is adopted in the binary image.
  • the rectangle with the largest area among all circumscribed rectangles is the lug ROI;
  • the resulting tab ROI is the tab image.
  • the thresholding is performed using the cumulative gray-scale pixel approximation method, specifically as follows: under normal circumstances, the polar ear has a 255 gray value in the image, and the sum of its pixels floats around an average value T, and the local image pixels of the polar ear The length is l, and the number of pixels corresponding to each gray value in the low-contrast polar ear partial image is N i , where i represents the gray value, N represents the number of pixels, and S represents the pixels that are accumulated from 255 gray levels and decremented by 1 Sum, stop the accumulation when S is greater than T for the first time and get the gray value n.
  • the thresholding operation is performed with t as the parameter.
  • the parameter T is obtained by pre-calculating the number of 255 gray pixels of the normal image of the non-wrinkled battery cell tab and the normal image of the non-wrinkled PCM tab, and can be continuously accumulated and averaged to update T during real-time operation.
  • the partial image of the PCM lug with normal contrast is extracted to obtain the PCM lug image with the normal contrast and the low-contrast PCM lug image.
  • a low-contrast PCM pole ear image is obtained.
  • spot analysis or Hough circle transformation method is used to detect solder joints.
  • the solder joint detection step includes:
  • the Log-Gabor even filter is used for filtering processing, and the Hough circle transform is used to detect the circle after processing;
  • the step of accumulating the circle results detected by the Hough circle transformation after each Log-Gabor even filtering and performing the filtering operation includes:
  • All the circles retained in the previous step are arranged from small to large according to the center y coordinate. If the distance between the y coordinates of two adjacent circles is less than ⁇ y , then the y coordinates of the two circles are considered to be the same;
  • All the circles retained in the previous step are arranged according to the center x coordinate from small to large. If the distance between the x coordinates of two adjacent circles is less than ⁇ x , then the x coordinates of the two circles are considered to be the same;
  • the step of mapping and predicting the welding point coordinates in the cell tab image by detecting the welding point coordinates in the PCM tab image and the calibration information obtained in advance, compare the right tab image of the PCM and the battery core.
  • the welding point coordinates of the left pole ear image are transformed, and the transformation formula is as follows:
  • ⁇ BL ⁇ is the image coordinate system of the left ear of the battery
  • ⁇ FR ⁇ is the image coordinate system of the right ear of the PCM
  • the origin of the coordinate system is the upper left corner of the image
  • the right is the positive x direction
  • the downward is the positive y direction.
  • P ij represents the coordinate vector of the welding point, where i represents the row, j represents the column, the PCM right pole ear image length w 1 , height h 1 , the battery cell left pole ear image length w 1 ', height h 1 ', a, a' , B, b′ are the pre-calibration information read from the txt file, a is the pixel length of the P-calibration distance from the left edge of the image in the PCM right polar ear image, and b is the bottom of the P-calibration distance image in the PCM right polar ear image
  • the pixel length of the edge, a′ is the pixel length of the left pole ear image of the battery corresponding to the solder point P calibration ′ from the right edge of the image, b′ is the pixel length of the left pole ear image of the battery corresponding to the solder point P calibration ′ from the bottom edge of the image .
  • ⁇ BR ⁇ is the image coordinate system of the right ear of the battery
  • ⁇ FL ⁇ is the image coordinate system of the left ear of the PCM
  • the origin of the coordinate system is the upper left corner of the image
  • the right is the positive x direction
  • the downward is the positive y direction.
  • P mn represents the coordinate vector of the solder joints, where m represents the row, n represents the column, PCM left pole ear image length w 2 , height h 2 , battery cell right pole ear image length w 2 ', height h 2 ', c, c′ , D, d′ are the pre-calibration information read from the txt file, c is the pixel length of the solder joint P calibration distance from the right edge of the image in the PCM left polar ear image, and d is the P calibration distance image bottom in the PCM left polar ear image The pixel length of the edge, c′ is the pixel length of the right pole ear image of the battery corresponding to the solder point P calibration ′ from the left edge of the image, and d′ is the pixel length of the right pole ear image of the battery corresponding to the solder point P calibration ′ from the bottom edge of the image .
  • the pre-calibration information is obtained in advance through an additional program, specifically:
  • the right pole ear image of the PCM and the left pole ear image of the battery are obtained;
  • the acquisition of pre-calibration information is obtained in advance through an additional program, specifically:
  • the right pole ear image of the PCM and the left pole ear image of the battery are obtained;
  • the pre-calibration information read from the txt file can only use a, b, a', b'or only use c, d, c', d'; when only a, b, a', b'are used, use replace use replace When using only c, d, c′, d′, use replace use replace And in real-time operation, it is possible to repeatedly record a, b, a', b', c, d, c', d'for accumulating average values, thereby reducing errors.
  • the method of spot analysis is used to search for the solder joints on the square ROI. If the spot analysis fails to find the solder joints Point, the histogram equalization process is performed on the image, and the Hough circle detection is used to search for solder joints. According to the search results in the square ROI, the formula for compensating the coordinates of the pole ear solder joints of the mapped battery is as follows (the origin of the image coordinate system is in the upper left corner):
  • J is the actual welding point coordinates in the pole ear image coordinate system of the battery
  • (x temp , y temp ) is the coordinates of the welding point in the square ROI image coordinate system ⁇ Square ⁇
  • l is the length of the square ROI. If detected For multiple circles, only the one closest to the center of the rectangle is used.
  • the ROI side length l is a preset parameter, which is related to the installation position of the image acquisition mechanism and the model selection, l is about 2 times the diameter of the solder joint and the pixel length
  • the present invention has the following beneficial effects:
  • the present invention improves the problem of misjudgment of unqualified products caused by traditional algorithms in the detection of wrinkled ear images.
  • the present invention meets the real-time requirements of actual automated production.
  • Figure 1 is a flow chart of a visual online inspection method for laser solder joints of lithium battery tabs
  • Fig. 2 is a flow chart of judging the contrast of the partial pole ear image in an embodiment of the present invention
  • Fig. 3 is a flowchart of low-contrast PCM tab image processing in an embodiment of the present invention.
  • Fig. 4 is a schematic diagram of obtaining calibration information in the step of mapping the solder joint coordinates of the PCM tab image from the solder joint coordinates of the PCM tab image in the embodiment of the present invention
  • Fig. 5 is a schematic structural diagram of a visual online inspection system for laser solder joints of lithium battery tabs in an embodiment of the present invention.
  • a visual online inspection system for laser solder joints of lithium battery tabs is provided, and the structure of the system is shown in FIG. 5.
  • the lithium battery 113 is transported to the visual inspection station by the feeding mechanism 112, and the lithium battery is fixed and held by the clamp 114.
  • the visual inspection station has dome light sources 106 and 108, industrial cameras 107 and 110, and light source control After the lithium battery 113 is placed, the visual inspection method of the present invention starts to work.
  • Figure 1 shows a flow chart of a visual online inspection method for laser solder joints of lithium battery tabs, which includes the following steps:
  • PCM circuit protection module
  • two sets of dome light sources 106, 108 and industrial cameras 107, 110 are used for acquisition.
  • the global image of the PCM tab and the global image of the battery cell tab with a size of 2592*1944 are obtained.
  • the tabs without wrinkles are highlighted in the image.
  • the tabs with wrinkles have most of the shadows.
  • the battery cell pole ear global image is divided into a cell left pole ear local image and a cell right pole ear local image;
  • the PCM pole ear global image is divided into a PCM board left pole ear local image and a PCM board right pole ear local image .
  • set the partial image of the left pole ear as a starting point (0,450), a rectangular area with a length of 800, and a width of 600 in the global image
  • set the partial image of the right pole ear as a starting point (1700,600), a length of 800, and a width of 600. 600 rectangular area.
  • the resulting PCM left pole ear partial image, PCM right pole ear partial image, battery core left pole ear partial image, and battery core right pole ear partial image are all in the partial image.
  • Figure 2 shows the flow chart of the contrast of the partial image of the tabs of the lithium battery obtained by judging, including the steps:
  • (3-1) Divide the gray histogram statistical results of the partial image of the extreme ears into 3 intervals: low gray value interval, medium gray value interval, and high gray value interval;
  • the gray level of the background area far from the light source irradiated by the lithium battery is in the low gray value range, and the range is about 0 to 20;
  • the gray scale of the overall lug shadow caused by the shadow or the bending and tilt of the lug is in the middle gray value range, and the range is about 30 to 80; there is no wrinkle in the local image of the lug, and the grayscale of the non-welded area of the normal lug is in a high grayscale.
  • Value range the range is approximately 200 to 255;
  • is greater than or equal to the preset constant const, the partial image of the polar ear is judged as normal contrast, otherwise it is low contrast; the constant const is calculated by comparing the sum of the number of pixels in the high gray value interval of a normal polar ear image with no wrinkles and medium gray The quotient of the sum of the number of pixels in the degree value interval is used as a reference, and the average value can be continuously accumulated in real-time operation to reduce the error.
  • the constant const is 3, and the grayscale histogram of the partial image of the left polar ear of the PCM is counted.
  • the sum of the number of pixels in the highlight gray value interval is 10 times the sum of the number of pixels in the middle gray value interval, and it is judged as normal Contrast polar ear image.
  • the sum of the number of pixels in the PCM right polar ear partial image is 1.5 times the sum of the number of pixels in the middle gray value range. It is judged as low contrast; similarly, the left electrode of the battery is judged as low. Contrast, the right pole ear of the battery is judged as normal contrast.
  • the ROI of the pole ear extracted from the local image of the pole meets: the left edge of the ROI coincides with the left edge of the outline of the pole, the right edge of the ROI coincides with the right edge of the outline of the pole, and the lower edge of the ROI coincides with the lower edge of the outline of the pole.
  • different threshold parameters are used to perform threshold image segmentation to extract the polar ear ROI; the PCM left polar ear partial image is of normal contrast, so the OTSU threshold is used for binarization, and the largest connected area in the binarized image is extracted
  • the smallest circumscribed rectangle of the domain obtains the PCM left polar ear image; the PCM right polar ear image is of low contrast, and the number of pixels is accumulated from the gray value of 255. When the accumulated sum reaches 85000, the accumulation stops. At this time, the gray The degree is 35, so the binarization operation is performed on the partial image of the PCM right pole ear with a threshold of 35.
  • the circumscribed rectangle with the largest area is extracted from the binary image to obtain the PCM right pole ear image; similarly, the left pole ear of the battery is obtained.
  • Image image of the right pole ear of the battery cell.
  • ⁇ 0 is the center wavelength of 45
  • ( ⁇ x , ⁇ y ) T
  • ⁇ 0 is the parameter that determines the bandwidth.
  • ⁇ BL ⁇ is the image coordinate system of the left ear of the battery
  • ⁇ FR ⁇ is the image coordinate system of the right ear of the PCM
  • the origin of the coordinate system is the upper left corner of the image
  • the right is the positive x direction
  • the downward is the positive y direction.
  • P ij represents the coordinate vector of the welding point, where i represents the row, j represents the column, the PCM right pole ear image length w 1 , height h 1 , the battery cell left pole ear image length w 1 ', height h 1 ', a, a' , B, b′ are the pre-calibration information read from the txt file, a is the pixel length of the P-calibration distance from the left edge of the image in the PCM right polar ear image, and b is the bottom of the P-calibration distance image in the PCM right polar ear image
  • the pixel length of the edge, a′ is the pixel length of the left pole ear image of the battery corresponding to the solder point P calibration ′ from the right edge of the image, b′ is the pixel length of the left pole ear image of the battery corresponding to the solder point P calibration ′ from the bottom edge of the image .
  • ⁇ BR ⁇ is the image coordinate system of the right ear of the battery
  • ⁇ FL ⁇ is the image coordinate system of the left ear of the PCM
  • the origin of the coordinate system is the upper left corner of the image
  • the right is the positive x direction
  • the downward is the positive y direction.
  • P mn represents the coordinate vector of the solder joints, where m represents the row, n represents the column, PCM left pole ear image length w 2 , height h 2 , battery cell right pole ear image length w 2 ', height h 2 ', c, c′ , D, d′ are the pre-calibration information read from the txt file, c is the pixel length of the solder joint P calibration distance from the right edge of the image in the PCM left polar ear image, and d is the P calibration distance image bottom in the PCM left polar ear image The pixel length of the edge, c′ is the pixel length of the right pole ear image of the battery corresponding to the solder point P calibration ′ from the left edge of the image, and d′ is the pixel length of the right pole ear image of the battery corresponding to the solder point P calibration ′ from the bottom edge of the image .
  • the pre-calibration information is obtained in advance through an additional program, specifically:
  • the side length of the square ROI is 20 pixels.
  • the spot analysis method is used to search for the solder joints in the square ROI. If the spot analysis fails to search for the solder joints, the histogram equalization process is performed on the image, and then the Hough circle is used to detect the solder joints. point.
  • the formula for compensating the coordinates of the pole ear solder joints of the mapped battery is as follows (the origin of the image coordinate system is in the upper left corner):
  • J is the actual solder joint coordinates in the pole ear image coordinate system of the battery
  • (x temp , y temp ) is the solder joint coordinates in the square ROI image coordinate system ⁇ Square ⁇
  • l is 20. If multiple circles are detected Just use the one closest to the center of the rectangle.
  • solder joint inspection of all PCM tab images and battery cell tab images has been completed, and the solder joints of the lithium battery 113 are marked as good or defective according to the test results. If it is a good product, the unloading mechanism 109 transports the lithium battery 113 to the good product unloading conveyor belt 103;

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Analytical Chemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biochemistry (AREA)
  • Theoretical Computer Science (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Quality & Reliability (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Processing (AREA)
  • Connection Of Batteries Or Terminals (AREA)

Abstract

一种锂电池(113)极耳激光焊点视觉在线检测方法,包括:采集PCM极耳全局图像与电芯极耳全局图像;对全局图像进行划分,得到锂电池极耳局部图像;对极耳局部图像的对比度进行判断;依据判断的对比度结果采取相应方法从极耳局部图像得到极耳图像;依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测;通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标;在电芯极耳图像中,以得到的每一个映射预估焊点坐标为中心生成一个正方形ROI;在正方形ROI中进行焊点检测,得到实际的电芯极耳图像中的焊点坐标。该方法克服了传统视觉算法难以检测褶皱极耳焊点的问题,实现了检测准确性并满足在线检测的要求。

Description

一种锂电池极耳激光焊点视觉在线检测方法 技术领域
本发明涉及锂电池自动化封装工艺中的激光焊点检测领域,尤其涉及一种锂电池极耳激光焊点视觉在线检测方法。
背景技术
随着机器视觉技术的不断发展,视觉缺陷检测在工业领域应用越来越广泛。目前,在手机等数码产品的锂电池自动化Pack封装工艺中普遍采用了机器视觉方法对锂电池电芯极耳与PCM(电路保护模块)板极耳激光焊接的焊点进行缺陷在线检测。然而,由于电芯极耳材质非常软且纤薄,所以极易褶皱弯曲。而极耳的褶皱极易产生低对比度和局部阴影成像特征,极耳弯曲倾斜又容易引起极耳图像产生整体阴影特征。因此电芯极耳的这种图像特征严重影响传统视觉算法的焊点识别,造成识别不出焊点或识别焊点数目遗漏或增加的问题,从而误判良品为不合格。针对这一问题,检测设备厂商采用在视觉检测工位前加入一个碾压除皱工艺的方法,但这种方法存在明显缺陷:降低了生产节拍,提高了生产成本,使用一段时间后除皱效果会显著降低。
目前,锂电池电芯极耳与PCM板极耳焊点缺陷检测采用传统的视觉检测算法,主要算法有:形状模板匹配、斑点分析、轮廓匹配等;预处理算法通常有:灰度线性变换、腐蚀、膨胀等。而一旦图像有局部阴影或整体图像出现低对比度特征,常常会造成这类匹配检测算法和斑点分析等焊点检测方法失效。完善光源照明方案仍然难以普遍性地改善或解决极耳褶皱和弯曲倾斜引起的图像质量劣化问题。所以传统视觉检测算法难以实现极耳褶皱和弯曲倾斜情况下的视觉焊点识别。
由于传统的视觉检测方法无法有效解决锂电池极耳褶皱引起的检测问题,所以锂电池焊点人工目检法仍然被大量使用。但人工目检法存在效率低和效果不稳定等缺点。因此,针对锂电池极耳褶皱和弯曲倾斜带来图像质量劣化,提出有效的视觉焊点检测方法是十分重要的。
发明内容
本发明的目的在于克服现有技术的不足,提供一种锂电池极耳激光焊点视觉在线检测方法。本发明克服了传统检测方法难以检测褶皱极耳焊点的问题,同时满足了检测准确性和在线检测的要求。本发明依据极耳图像对比度选择具体检测算法,并根据定位关系缩小检测ROI区域。
本发明的目的能够通过以下技术方案实现:
一种锂电池极耳激光焊点视觉在线检测方法,包括步骤:
采集PCM(电路保护模块)极耳全局图像与电芯极耳全局图像;
对PCM板极耳全局图像和电芯极耳全局图像进行划分,得到锂电池极耳局部图像;
对得到的锂电池极耳局部图像的对比度进行判断;
依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI,得到PCM极耳图像和电芯极耳图像;
依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测;
通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标;
在电芯极耳图像中,以得到的每一个映射预估焊点坐标为中心生成一个正方形ROI;
在正方形ROI中进行焊点检测,如果检测到焊点就对预估的焊点坐标 进行补偿,得到实际的电芯极耳图像中的焊点坐标。
具体地,所述采集PCM极耳全局图像与电芯极耳全局图像的步骤中,采用2组圆顶光源和黑白工业相机模组对锂电池的电芯极耳与PCM板极耳进行全局图像采集。
所述光源采用蓝光照明;两个工业相机对置安装,图像采集时锂电池位于两个工业相机中间;两个相机分别采集PCM板极耳全局图像和电芯极耳全局图像;所述极耳全局图像涵盖左、右2个极耳;两个工业相机模组(相机与镜头)完全相同,视场重合区域需要完全覆盖左右极耳;综合相机曝光时间和光源光强应该达到的成像效果是使得正常无褶皱极耳在图像中的灰度值为255。
具体地,所述对PCM板极耳全局图像和电芯极耳全局图像进行划分的步骤中,电芯极耳全局图像划分为电芯左极耳局部图像和电芯右极耳局部图像;PCM极耳全局图像划分为PCM板左极耳局部图像和PCM板右极耳局部图像。每个极耳局部图像均要求完整地包含该极耳轮廓,且极耳局部图像应该涵盖全局图像中锂电池极耳因为锂电池装夹误差导致可能出现的所有位置。
具体地,所述对得到的锂电池极耳局部图像的对比度进行判断的步骤中,包括:
将极耳局部图像的灰度直方图统计结果划分为3区间:低灰度值区间、中等灰度值区间、高灰度值区间;
极耳局部图像中在锂电池周围远离光源照射的背景区域的灰度处于低灰度值区间,范围大约为0到20;极耳局部图像中极耳上的焊点、褶皱极耳引起的局部阴影或者极耳弯曲倾斜造成的整体极耳阴影的灰度处于中等 灰度值区间,范围大约为30到80;极耳局部图像中无褶皱正常极耳的非焊点区域灰度处于高灰度值区间,范围大约为200到255;
通过高灰度值区间像素和S h与中等灰度值区间像素和S m的比值δ,判断极耳局部图像的对比度:
如果δ大于等于预设的常数const,则将极耳局部图像判断为正常对比度,否则为低对比度;其中常数const通过对一个无褶皱正常极耳图像的高灰度值区间像素数总和与中等灰度值区间像素数总和的商做参照,并且可以在实时运行中不断累计求平均值来减小误差。
具体地,所述依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI的步骤中,从极耳局部图像提取的极耳ROI需要满足:ROI左边缘与极耳轮廓左边缘重合,ROI右边缘与极耳轮廓右边缘重合,ROI下边缘与极耳轮廓下边缘重合。
具体地,所述依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI的步骤中,包括:
对于正常对比度的极耳局部图像,采用OTSU算法进行阈值化,并在二值图像中采用寻找最小外接矩形法,所有外接矩形中面积最大的矩形就是极耳ROI;
对于低对比度的极耳局部图像,采用累加灰度像素数逼近法进行阈值化,并在二值图像中采用寻找最小外接矩形法,所有外接矩形中面积最大的矩形就是极耳ROI;
得到的极耳ROI即为极耳图像。
更进一步地,所述采用累加灰度像素逼近法进行阈值化,具体为:正常情况下极耳在图像中是255灰度值,其像素总和在一个平均值T附近浮 动,极耳局部图像像素长度为l,低对比度极耳局部图像中每个灰度值对应的像素数量为N i,其中i表示灰度值,N表示像素数量,S表示从255灰度开始按照递减1进行累加的像素和,当累加到S第一次大于T时停止累加并得到灰度值n,灰度值为n时,如果S减T的差大于2l那么255减去n就是阈值t,如果S减T的差小于等于2l那么256减n就是阈值t,以t为参数进行阈值化操作。其中参数T通过预先对无褶皱电芯极耳正常图像和无褶皱PCM极耳正常图像进行255灰度像素数量统计来获取,并且可以在实时运行中进行不断累加求平均值更新T。
阈值t的求解公式为:
Figure PCTCN2020122843-appb-000001
Figure PCTCN2020122843-appb-000002
具体地,所述依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测的步骤中,正常对比度的PCM极耳局部图像提取后获得正常对比度的PCM极耳图像,低对比度的PCM极耳局部图像提取后获得低对比度的PCM极耳图像。对于正常对比度的PCM极耳图像,采用斑点分析或霍夫圆变换的方法进行焊点检测。
具体地,所述依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测的步骤中,对于低对比度的PCM极耳图像,焊点检测步骤包括:
对PCM极耳图像进行直方图均衡化处理;
采用Log-Gabor even滤波器进行滤波处理,处理后进行霍夫圆变换检 测圆;
改变Log-Gabor even滤波器的参数σ 0,重复滤波处理和检测圆若干次;
将每一次Log-Gabor even滤波后进行霍夫圆变换检测到的圆结果,累计在一起进行筛选操作,最后保留真正的焊点。
更进一步地,所述将每一次Log-Gabor even滤波后进行霍夫圆变换检测到的圆结果,累计在一起进行筛选操作的步骤中,包括:
从第一个圆开始循环遍历,合并圆心坐标距离小于ε的圆,新圆的圆心坐标为所合并圆的圆心坐标的平均值,并标记合并圆的数量;
对所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x坐标距离小于ε x,那么认为这两个圆x坐标是同一的;
统计同一x坐标下所有合并圆的数目和,如果和大于等于C x,保留该x坐标下的所有圆,反之则删去;
对上一步保留的所有圆按照圆心y坐标从小到大进行排列,如果相邻两个圆y坐标距离小于ε y,那么认为这两个圆y坐标是同一的;
统计同一y坐标下所有合并圆的数目和,如果和大于等于C y,保留该y坐标下的所有圆,反之则删去;
对上一步保留的所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x坐标距离小于ε x,那么认为这两个圆x坐标是同一的;
统计同一x坐标下所有合并圆的数目和,如果这个和大于等于C x′那么就保留该x坐标下的所有圆,反之则删去;
其中C x′>C x
具体地,所述通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标的步骤中,对PCM右极耳图 像和电芯左极耳图像的焊点坐标进行变换,变换公式如下:
Figure PCTCN2020122843-appb-000003
其中,{BL}为电芯左极耳图像坐标系,{FR}为PCM右极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P ij表示焊点坐标向量,其中i表示行、j表示列,PCM右极耳图像长w 1,高h 1,电芯左极耳图像长w 1’,高h 1’,a、a′、b、b′是从txt文档中读取的预先标定信息,a为PCM右极耳图像中焊点P 标定距离图像左边缘的像素长度,b是PCM右极耳图像中P 标定距离图像底边缘的像素长度,a′为电芯左极耳图像对应焊点P 标定′距离图像右边缘的像素长度,b′为电芯左极耳图像对应焊点P 标定′距离图像底边缘的像素长度。
对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换,变换公式如下:
Figure PCTCN2020122843-appb-000004
其中,{BR}为电芯右极耳图像坐标系,{FL}为PCM左极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P mn表示焊点坐标向量,其中m表示行,n表示列,PCM左极耳图像长w 2、高h 2、电芯右极耳图像长w 2’、高h 2’,c、c′、d、d′是从txt文档中读取的预先标定信息,c为PCM左极耳图像中焊点P 标定距离图像右边缘的像素长度,d是PCM左极耳图像中P 标定距离图像底边缘的像素长度,c′为电芯右极耳 图像对应焊点P 标定′距离图像左边缘的像素长度,d′为电芯右极耳图像对应焊点P 标定′距离图像底边缘的像素长度。
更进一步地,在对PCM右极耳图像和电芯左极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,具体为:
输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
通过对比度判断及极耳ROI提取,得到PCM右极耳图像和电芯左极耳图像;
在PCM右极耳图像中指定一个焊点的位置( FRx, FRy),求解a,b,计算公式为:
Figure PCTCN2020122843-appb-000005
在电芯左极耳图像中指定与PCM右极耳对应焊点的位置( BLx, BLy),求解a′,b′,计算公式为:
Figure PCTCN2020122843-appb-000006
将a、b、a′、b′的值保存在xml或txt文件中。
更进一步地,在对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,具体为:
输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
通过对比度判断及极耳ROI提取,得到PCM右极耳图像和电芯左极耳图像;
在PCM左极耳图像中指定一个焊点的位置( FLx, FLy),求解c,d,计算 公式为:
Figure PCTCN2020122843-appb-000007
在电芯右极耳图像中指定与PCM左极耳对应焊点的位置( BRx, BRy),求解c′,d′,计算公式为:
Figure PCTCN2020122843-appb-000008
将c、d、c′、d′的值保存在xml或txt文件中。
所述从txt文件中读取的预先标定信息可以只使用a、b、a′、b′或只使用c、d、c′、d′;只使用a、b、a′、b′时,用
Figure PCTCN2020122843-appb-000009
代替
Figure PCTCN2020122843-appb-000010
Figure PCTCN2020122843-appb-000011
代替
Figure PCTCN2020122843-appb-000012
只使用c、d、c′、d′时,用
Figure PCTCN2020122843-appb-000013
代替
Figure PCTCN2020122843-appb-000014
Figure PCTCN2020122843-appb-000015
代替
Figure PCTCN2020122843-appb-000016
并且在实时运行中可以通过重复记录a、b、a′、b′、c、d、c′、d′进行累加求平均值,从而减小误差。
具体地,所述在正方形ROI中进行焊点检测,得到实际的电芯极耳图像中的焊点坐标的步骤中,采用斑点分析的方法对正方形ROI搜索焊点,如果斑点分析搜索不到焊点,则对图像进行直方图均衡化处理,再使用霍夫圆检测搜索焊点。通过正方形ROI内搜索结果对映射的电芯极耳焊点坐标进行补偿公式如下(图像坐标系原点在左上角):
Figure PCTCN2020122843-appb-000017
其中J为实际的电芯极耳图像坐标系下焊点坐标,(x temp,y temp)为焊点在正方形ROI图像坐标系{Square}中的坐标,l为正方形ROI的长度,如果检测到多个圆就只用距离矩形中心最近的一个。其中ROI边长l为预设参 数,与图像采集机构安装位置和选型型号有关,l大概2倍焊点圆直径像素长度
本发明相较于现有技术,具有以下的有益效果:
(1)本发明在不增加硬件成本的基础上,改善了传统算法对褶皱极耳图像检测造成的良品误判不合格问题。
(2)本发明满足实际自动化生产的实时性要求。
附图说明
图1为一种锂电池极耳激光焊点视觉在线检测方法的流程图;
图2是本发明实施例中局部极耳图像对比度判断的流程图;
图3是本发明实施例中低对比度PCM极耳图像处理的流程图;
图4是本发明实施例中从PCM极耳图像的焊点坐标映射电芯极耳图像焊点坐标步骤中获取标定信息的示意图;
图5是本发明实施例中锂电池极耳激光焊点视觉在线检测系统的结构示意图。
具体实施方式
下面结合实施例及附图对本发明作进一步详细的描述,但本发明的实施方式不限于此。
实施例
在本实施例中,提供一个锂电池极耳激光焊点视觉在线检测系统,所述系统的结构如图5所示。在所述系统中,锂电池113被上料机构112搬运到视觉检测工位上,由夹具114固定保持锂电池,视觉检测工位有圆顶光源106和108,工业相机107和110,光源控制器104,图像采集卡105,运算处理与控制单元102,显示设备101;锂电池113放置好后,本发明的 视觉检测方法开始工作。
如图1所示为一种锂电池极耳激光焊点视觉在线检测方法的流程图,包括步骤:
(1)采集PCM(电路保护模块)极耳全局图像与电芯极耳全局图像;
在本实施例中,采用2组圆顶光源106、108和工业相机107、110进行采集。得到了尺寸为2592*1944的PCM极耳全局图像和电芯极耳全局图像,每个全局图像中有左右2个极耳被拍摄到,在图像中没有褶皱的极耳是高亮的,有存在褶皱的极耳有大部分阴影。
(2)划分全局图像,得到4个锂电池极耳局部图像;
所述电芯极耳全局图像划分为电芯左极耳局部图像和电芯右极耳局部图像;所述PCM极耳全局图像划分为PCM板左极耳局部图像和PCM板右极耳局部图像。
在本实施例中,在全局图像中设置左极耳的局部图像为起点(0,450)、长800、宽600的矩形区域,右极耳局部图像为起点(1700,600)、长800,、宽600的矩形区域。从而得到的PCM左极耳局部图像、PCM右极耳局部图像、电芯左极耳局部图像和电芯右极耳局部图像的每个极耳轮廓都完全处于局部图像内。
(3)判断得到的锂电池极耳局部图像的对比度;
如图2所示为判断得到的锂电池极耳局部图像的对比度的流程图,包括步骤:
(3-1)将极耳局部图像的灰度直方图统计结果划分为3区间:低灰度值区间、中等灰度值区间、高灰度值区间;
极耳局部图像中在锂电池周围远离光源照射的背景区域的灰度处于低 灰度值区间,范围大约为0到20;极耳局部图像中极耳上的焊点、褶皱极耳引起的局部阴影或者极耳弯曲倾斜造成的整体极耳阴影的灰度处于中等灰度值区间,范围大约为30到80;极耳局部图像中无褶皱正常极耳的非焊点区域灰度处于高灰度值区间,范围大约为200到255;
(3-2)通过高灰度值区间像素和S h与中等灰度值区间像素和S m的比值δ,判断极耳局部图像的对比度:
如果δ大于等于预设的常数const,则将极耳局部图像判断为正常对比度,否则为低对比度;其中常数const通过对一个无褶皱正常极耳图像的高灰度值区间像素数总和与中等灰度值区间像素数总和的商做参照,并且可以在实时运行中不断累计求平均值来减小误差。
在本实施例中,取常数const为3,统计PCM左极耳局部图像的灰度直方图,高亮灰度值区间像素数目和是中等灰度值区间像素数目和的10倍,判断为正常对比度的极耳图像,PCM右极耳局部图像中高亮灰度值区间像素数目和是中等灰度值区间像素数目和的1.5倍,判断为低对比度;类似地,电芯左极耳判断为低对比度,电芯右极耳判断为正常对比度。
(4)依据判断的对比度结果采用相应方法从极耳局部图像中提取极耳ROI,得到极耳图像;
从极耳局部图像提取的极耳ROI满足:ROI左边缘与极耳轮廓左边缘重合,ROI右边缘与极耳轮廓右边缘重合,ROI下边缘与极耳轮廓下边缘重合。
在本实施例中,通过对比度判断结果采用不同阈值参数进行阈值图像分割提取极耳ROI;PCM左极耳局部图像是正常对比度所以采用OTSU阈值进行二值化,提取二值化图像中面积最大连通域的最小外接矩形得到了 PCM左极耳图像;PCM右极耳图像为低对比度,采用从255灰度值开始累加像素数量求和,当累加的和达到85000时,停止累加,此时的灰度为35,所以对PCM右极耳局部图像进行阈值为35的二值化操作,同样在二值图像中提取面积最大的外接矩形得到PCM右极耳图像;类似地,得到电芯左极耳图像,电芯右极耳图像。
(5)依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测;
在本实施例中,根据对比度不同采取不同的方法进行PCM极耳图像的焊点检测:对PCM左极耳图像采用斑点分析的方法,检测到了全部6个焊点;对于低对比度的PCM右极耳图像,焊点检测步骤如图3所示,包括:
(5-1)对PCM极耳图像进行直方图均衡化处理;
(5-2)采用Log-Gabor even滤波器进行滤波处理,处理后进行霍夫圆变换检测圆;
(5-3)改变Log-Gabor even滤波器的参数σ 0,重复步骤(5-2)10次,σ 0的参数从0.01到0.10,;
(5-4)将每一次Log-Gabor even滤波后进行霍夫圆变换检测到的圆结果,累计在一起进行3次基于不同方向重复次数的筛选操作:依次是x方向、y方向、x方向,每次筛选的结果作为下一次筛选的输入,最后保留真正的焊点。
Log-Gabor even滤波器表达式如下:
Figure PCTCN2020122843-appb-000018
其中ω 0是中心波长取值45,ω=(ω xy) T,σ 0是决定带宽的参数,每次 滤波后进行霍夫圆变换检测圆,将所有圆结果累积在一起,按照以下步骤进行筛选:
(5-4-1)从第一个圆开始循环遍历,合并圆心坐标距离小于3的圆,新圆的圆心坐标为所合并圆的圆心坐标的平均值,并标记合并圆的数量;
(5-4-2)首先对所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x像素坐标距离小于6,那么认为这两个圆x坐标是同一的;
(5-4-3)统计同一x坐标下所有圆的数目和,如果这个数目大于等于14那么就保留该x坐标下的所有圆,反之则删去;
(5-4-4)对上一步保留的所有圆按照圆心y坐标从小到大进行排列,如果相邻两个圆y坐标距离小于5,那么认为这两个圆y坐标是同一的;
(5-4-5)统计同一y坐标下所有圆的数目和,如果这个数目大于等于25那么就保留该y坐标下的所有圆,反之则删去;
(5-4-6)对上一步保留的所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x坐标距离小于5,那么认为这两个圆x坐标是同一的;
(5-4-7)统计同一x坐标下所有圆的数目和,如果这个数目大于等于16那么就保留该x坐标下的所有圆,反之则删去。
完成筛选后得到PCM右极耳图像中全部6个焊点。
(6)通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标;
对PCM右极耳图像和电芯左极耳图像的焊点坐标进行变换,变换公式如下:
Figure PCTCN2020122843-appb-000019
其中,{BL}为电芯左极耳图像坐标系,{FR}为PCM右极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P ij表示焊点坐标向量,其中i表示行、j表示列,PCM右极耳图像长w 1,高h 1,电芯左极耳图像长w 1’,高h 1’,a、a′、b、b′是从txt文档中读取的预先标定信息,a为PCM右极耳图像中焊点P 标定距离图像左边缘的像素长度,b是PCM右极耳图像中P 标定距离图像底边缘的像素长度,a′为电芯左极耳图像对应焊点P 标定′距离图像右边缘的像素长度,b′为电芯左极耳图像对应焊点P 标定′距离图像底边缘的像素长度。
对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换,变换公式如下:
Figure PCTCN2020122843-appb-000020
其中,{BR}为电芯右极耳图像坐标系,{FL}为PCM左极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P mn表示焊点坐标向量,其中m表示行,n表示列,PCM左极耳图像长w 2、高h 2、电芯右极耳图像长w 2’、高h 2’,c、c′、d、d′是从txt文档中读取的预先标定信息,c为PCM左极耳图像中焊点P 标定距离图像右边缘的像素长度,d是PCM左极耳图像中P 标定距离图像底边缘的像素长度,c′为电芯右极耳图像对应焊点P 标定′距离图像左边缘的像素长度,d′为电芯右极耳图像对应 焊点P 标定′距离图像底边缘的像素长度。
在对PCM右极耳图像和电芯左极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,获取示意图如图4所示,具体为:
(6-1-1)输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
(6-1-2)通过步骤(2)-(5)得到PCM右极耳图像和电芯左极耳图像;
(6-1-3)在PCM右极耳图像中指定一个焊点的位置( FRx, FRy),求解a,b,计算公式为:
Figure PCTCN2020122843-appb-000021
(6-1-4)在电芯左极耳图像中指定与PCM右极耳图像对应焊点的位置( BLx, BLy),求解a′,b′,计算公式为:
Figure PCTCN2020122843-appb-000022
(6-1-5)将a、b、a′、b′的值保存在xml或txt文件中。
在对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,具体为:
(6-2-1)输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
(6-2-2)通过步骤(2)-(5)得到PCM右极耳图像和电芯左极耳图像;
(6-2-3)在PCM左极耳图像中指定一个焊点的位置( FLx, FLy),求解 c,d,计算公式为:
Figure PCTCN2020122843-appb-000023
(6-2-4)在电芯右极耳图像中指定与PCM左极耳图像对应焊点的位置( BRx, BRy),求解c′,d′,计算公式为:
Figure PCTCN2020122843-appb-000024
(6-2-5)将c、d、c′、d′的值保存在xml或txt文件中。
现在得到了电芯左极耳图像中被映射过来的6个焊点的坐标,和电芯右极耳图像中被映射过来的6个焊点的坐标。
(7)以得到的每一个映射预估焊点坐标为中心生成一个正方形ROI;
在本实施例中,正方形ROI边长为20个像素。
(8)在正方形ROI中进行焊点检测,如果检测到焊点就对预估的焊点坐标进行补偿,得到实际的电芯极耳图像中的焊点坐标。
具体地,所述步骤(8)中,采用斑点分析的方法对正方形ROI搜索焊点,如果斑点分析搜索不到焊点,则对图像进行直方图均衡化处理,再使用霍夫圆检测搜索焊点。通过正方形ROI内搜索结果对映射的电芯极耳焊点坐标进行补偿公式如下(图像坐标系原点在左上角):
Figure PCTCN2020122843-appb-000025
其中J为实际的电芯极耳图像坐标系下焊点坐标,(x temp,y temp)为焊点在正方形ROI图像坐标系{Square}中的坐标,l为20,如果检测到多个圆就只用距离矩形中心最近的一个。
完成了对所有PCM极耳图像与电芯极耳图像的焊点检测,根据检测结果标记锂电池113焊点焊接是良品还是不良品。如果是良品,由下料机构109搬运锂电池113到良品下料输送带103;如果是不良品,由下料机构109搬运锂电池113到不良品下料输送带111。
上述实施例为本发明较佳的实施方式,但本发明的实施方式并不受上述实施例的限制,其他的任何未背离本发明的精神实质与原理下所作的改变、修饰、替代、组合、简化,均应为等效的置换方式,都包含在本发明的保护范围之内。

Claims (10)

  1. 一种锂电池极耳激光焊点视觉在线检测方法,其特征在于,包括步骤:
    采集PCM极耳全局图像与电芯极耳全局图像;
    对PCM板极耳全局图像和电芯极耳全局图像进行划分,得到锂电池极耳局部图像;
    对得到的锂电池极耳局部图像的对比度进行判断;
    依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI,得到PCM极耳图像和电芯极耳图像;
    依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测;
    通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标;
    在电芯极耳图像中,以得到的每一个映射预估焊点坐标为中心生成一个正方形ROI;
    在正方形ROI中进行焊点检测,如果检测到焊点就对预估的焊点坐标进行补偿,得到实际的电芯极耳图像中的焊点坐标。
  2. 根据权利要求1所述的方法,其特征在于,所述采集PCM极耳全局图像与电芯极耳全局图像的步骤中,采用2组圆顶光源和黑白工业相机模组对锂电池的电芯极耳与PCM板极耳进行全局图像采集;
    所述光源采用蓝光照明;两个工业相机对置安装,图像采集时锂电池位于两个工业相机中间;两个相机分别采集PCM极耳全局图像和电芯极耳全局图像;所述极耳全局图像涵盖左、右2个极耳;两个工业相机模组完全相同,视场重合区域需要完全覆盖左右极耳;综合相机曝光时间和光源光强应该达到的成像效果是使得正常无褶皱极耳在图像中的灰度值为255。
  3. 根据权利要求1所述的方法,其特征在于,所述对PCM板极耳全局图像和电芯极耳全局图像进行划分的步骤中,电芯极耳全局图像划分为电芯左极耳局部图像和电芯右极耳局部图像;PCM极耳全局图像划分为 PCM板左极耳局部图像和PCM板右极耳局部图像;每个极耳局部图像均要求完整地包含该极耳轮廓,且极耳局部图像应该涵盖全局图像中锂电池极耳因为锂电池装夹误差导致可能出现的所有位置。
  4. 根据权利要求1所述的方法,其特征在于,所述对得到的锂电池极耳局部图像的对比度进行判断的步骤中,包括:
    将极耳局部图像的灰度直方图统计结果划分为3区间:低灰度值区间、中等灰度值区间、高灰度值区间;
    通过高灰度值区间像素和S h与中等灰度值区间像素和S m的比值δ,判断极耳局部图像的对比度:
    如果δ大于等于预设的常数const,则将极耳局部图像判断为正常对比度,否则为低对比度;其中常数const通过对一个无褶皱正常极耳图像的高灰度值区间像素数总和与中等灰度值区间像素数总和的商做参照,并且可以在实时运行中不断累计求平均值来减小误差。
  5. 根据权利要求1所述的方法,其特征在于,所述依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI的步骤中,从极耳局部图像提取的极耳ROI需要满足:ROI左边缘与极耳轮廓左边缘重合,ROI右边缘与极耳轮廓右边缘重合,ROI下边缘与极耳轮廓下边缘重合。
  6. 根据权利要求1所述的方法,其特征在于,所述依据判断的对比度结果采取相应方法从极耳局部图像中提取极耳ROI的步骤中,包括:
    对于正常对比度的极耳局部图像,采用OTSU算法进行阈值化,并在二值图像中采用寻找最小外接矩形法,所有外接矩形中面积最大的矩形就是极耳ROI;
    对于低对比度的极耳局部图像,采用累加灰度像素数逼近法进行阈值化,并在二值图像中采用寻找最小外接矩形法,所有外接矩形中面积最大的矩形就是极耳ROI;
    得到的极耳ROI即为极耳图像;
    采用累加灰度像素逼近法进行阈值化,阈值t求解公式为:
    Figure PCTCN2020122843-appb-100001
    Figure PCTCN2020122843-appb-100002
    其中,正常情况下极耳在图像中是255灰度值,其像素总和在一个平均值T附近浮动,l为极耳局部图像像素长度,N i为低对比度极耳局部图像中每个灰度值对应的像素数量,i表示灰度值;S表示从255灰度开始按照递减1进行累加的像素和,n表示当累加到S第一次大于T时得到的灰度值。
  7. 根据权利要求1所述的方法,其特征在于,所述依据判断的对比度采取相应方法进行PCM极耳图像的焊点检测的步骤中,对于低对比度的PCM极耳图像,焊点检测步骤包括:
    对PCM极耳图像进行直方图均衡化处理;
    采用Log-Gabor even滤波器进行滤波处理,处理后进行霍夫圆变换检测圆;
    改变Log-Gabor even滤波器的参数σ 0,重复滤波处理和检测圆若干次;
    将每一次Log-Gabor even滤波后进行霍夫圆变换检测到的圆结果,累计在一起进行筛选操作,最后保留真正的焊点。
  8. 根据权利要求7所述的方法,其特征在于,所述将每一次Log-Gabor even滤波后进行霍夫圆变换检测到的圆结果,累计在一起进行筛选操作的步骤中,包括:
    从第一个圆开始循环遍历,合并圆心坐标距离小于ε的圆,新圆的圆心坐标为所合并圆的圆心坐标的平均值,并标记合并圆的数量;
    对所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x坐标 距离小于ε x,那么认为这两个圆x坐标是同一的;
    统计同一x坐标下所有合并圆的数目和,如果和大于等于C x,保留该x坐标下的所有圆,反之则删去;
    对上一步保留的所有圆按照圆心y坐标从小到大进行排列,如果相邻两个圆y坐标距离小于ε y,那么认为这两个圆y坐标是同一的;
    统计同一y坐标下所有合并圆的数目和,如果和大于等于C y,保留该y坐标下的所有圆,反之则删去;
    对上一步保留的所有圆按照圆心x坐标从小到大进行排列,如果相邻两个圆x坐标距离小于ε x,那么认为这两个圆x坐标是同一的;
    统计同一x坐标下所有合并圆的数目和,如果这个和大于等于C x′那么就保留该x坐标下的所有圆,反之则删去;
    其中C x′>C x
  9. 根据权利要求1所述的方法,其特征在于,所述通过检测的PCM极耳图像中的焊点坐标和预先获取的标定信息,映射预估电芯极耳图像中的焊点坐标的步骤中,对PCM右极耳图像和电芯左极耳图像的焊点坐标进行变换,变换公式如下:
    Figure PCTCN2020122843-appb-100003
    其中,{BL}为电芯左极耳图像坐标系,{FR}为PCM右极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P ij表示焊点坐标向量,其中i表示行、j表示列,PCM右极耳图像长w 1,高h 1,电芯左极耳图像长w 1’,高h 1’,a、a′、b、b′是从txt文档中读取的预先标定信息,a为PCM右极耳图像中焊点P 标定距离图像左边缘的像素长度,b是PCM右极耳图像中P 标定距离图像底边缘的像素长度,a′为电芯左极耳图像对应焊点P 标定′距离图像右边缘的像素长度,b′为电芯左极耳图像对应 焊点P 标定′距离图像底边缘的像素长度;
    对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换,变换公式如下:
    Figure PCTCN2020122843-appb-100004
    其中,{BR}为电芯右极耳图像坐标系,{FL}为PCM左极耳图像坐标系,坐标系原点均为图像左上角,向右为x正方向,向下为y正方向,P mn表示焊点坐标向量,其中m表示行,n表示列,PCM左极耳图像长w 2、高h 2、电芯右极耳图像长w 2’、高h 2’,c、c′、d、d′是从txt文档中读取的预先标定信息,c为PCM左极耳图像中焊点P 标定距离图像右边缘的像素长度,d是PCM左极耳图像中P 标定距离图像底边缘的像素长度,c′为电芯右极耳图像对应焊点P 标定′距离图像左边缘的像素长度,d′为电芯右极耳图像对应焊点P 标定′距离图像底边缘的像素长度;
    在对PCM右极耳图像和电芯左极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,具体为:
    输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
    通过对比度判断及ROI提取,得到PCM右极耳图像和电芯左极耳图像;
    在PCM右极耳图像中指定一个焊点的位置( FRx, FRy),求解a,b,计算公式为:
    Figure PCTCN2020122843-appb-100005
    在电芯左极耳图像中指定与PCM右极耳图像对应焊点的位置( BLx, BLy),求解a′,b′,计算公式为:
    Figure PCTCN2020122843-appb-100006
    将a、b、a′、b′的值保存在xml或txt文件中;
    在对PCM左极耳图像和电芯右极耳图像的焊点坐标进行变换的过程中,预先标定信息的获取是通过额外的程序事先得到的,具体为:
    输入正常无褶皱极耳的PCM板极耳全局图像和电芯极耳全局图像;
    通过对比度判断及ROI提取,得到PCM右极耳图像和电芯左极耳图像;
    在PCM左极耳图像中指定一个焊点的位置( FLx, FLy),求解c,d,计算公式为:
    Figure PCTCN2020122843-appb-100007
    在电芯右极耳图像中指定与PCM左极耳图像对应焊点的位置( BRx, BRy),求解c′,d′,计算公式为:
    Figure PCTCN2020122843-appb-100008
    将c、d、c′、d′的值保存在xml或txt文件中。
  10. 根据权利要求1所述的方法,其特征在于,所述在正方形ROI中进行焊点检测,得到实际的电芯极耳图像中的焊点坐标的步骤中,采用斑点分析的方法对正方形ROI搜索焊点,如果斑点分析搜索不到焊点,则对图像进行直方图均衡化处理,再使用霍夫圆变换搜索焊点;通过正方形ROI内搜索结果对映射的电芯极耳焊点坐标进行补偿公式如下:
    Figure PCTCN2020122843-appb-100009
    其中J为实际的电芯极耳图像坐标系下焊点坐标,(x temp,y temp)为焊点在矩形ROI图像坐标系{Square}中的坐标,l为矩形ROI的长度,如果检测 到多个圆就只用距离矩形中心最近的一个。
PCT/CN2020/122843 2020-02-25 2020-10-22 一种锂电池极耳激光焊点视觉在线检测方法 WO2021169335A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010114940.9 2020-02-25
CN202010114940.9A CN111307818B (zh) 2020-02-25 2020-02-25 一种锂电池极耳激光焊点视觉在线检测方法

Publications (1)

Publication Number Publication Date
WO2021169335A1 true WO2021169335A1 (zh) 2021-09-02

Family

ID=71156802

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/122843 WO2021169335A1 (zh) 2020-02-25 2020-10-22 一种锂电池极耳激光焊点视觉在线检测方法

Country Status (2)

Country Link
CN (1) CN111307818B (zh)
WO (1) WO2021169335A1 (zh)

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112363069A (zh) * 2020-09-18 2021-02-12 万向一二三股份公司 一种锂离子电池极耳断裂检测方法
CN113744244A (zh) * 2021-09-03 2021-12-03 广东奥普特科技股份有限公司 测量锂电池极片边缘到极耳边缘距离的在线视觉检测系统
CN114037657A (zh) * 2021-10-12 2022-02-11 上海电机学院 一种结合区域生长与环形校正的锂电池极耳缺陷检测方法
CN114549531A (zh) * 2022-04-26 2022-05-27 广州市易鸿智能装备有限公司 一种锂电池卷绕OverHang对中纠偏控制系统及方法
CN114594114A (zh) * 2022-03-09 2022-06-07 广东兆众自动化设备有限公司 一种锂电芯全自动在线无损检测方法
CN115082504A (zh) * 2022-08-23 2022-09-20 菏泽学院 一种太阳能光伏板光斑识别方法
CN115100600A (zh) * 2022-06-30 2022-09-23 苏州市新方纬电子有限公司 一种电池包的产线智能检测方法及系统
CN115178909A (zh) * 2022-07-11 2022-10-14 苏州汇川控制技术有限公司 焊接质量的检测方法、装置、终端设备与介质
CN115360319A (zh) * 2022-09-20 2022-11-18 楚能新能源股份有限公司 一种多极耳电芯极片信息追溯方法
CN115661040A (zh) * 2022-09-28 2023-01-31 广州市易鸿智能装备有限公司 一种锂电池电芯涂胶的检测方法及装置、电子设备
CN115829913A (zh) * 2022-08-10 2023-03-21 宁德时代新能源科技股份有限公司 裸电芯外观检测方法、装置、计算机设备和存储介质
CN116008294A (zh) * 2022-12-13 2023-04-25 无锡微准科技有限公司 一种基于机器视觉的键帽表面颗粒缺陷检测方法
CN116596932A (zh) * 2023-07-18 2023-08-15 北京阿丘机器人科技有限公司 电池顶盖极柱外观检测方法、装置、设备及存储介质
CN117078580A (zh) * 2022-05-10 2023-11-17 广州镭晨智能装备科技有限公司 焊点检测方法、装置及计算机可读存储介质
CN117109447A (zh) * 2023-10-24 2023-11-24 钛玛科(北京)工业科技有限公司 自适应极耳宽度检测方法、装置及设备
CN117309889A (zh) * 2023-11-27 2023-12-29 宁德时代新能源科技股份有限公司 极片边缘检测装置及方法
WO2024032002A1 (zh) * 2022-08-10 2024-02-15 宁德时代新能源科技股份有限公司 电芯检测方法、装置、系统、计算机设备和存储介质
CN117570852A (zh) * 2024-01-15 2024-02-20 钛玛科(北京)工业科技有限公司 极耳顶点坐标检测方法、装置及设备
CN117593309A (zh) * 2024-01-19 2024-02-23 常熟理工学院 基于立体重构的芯片焊点检测方法、系统及存储介质
CN117723491A (zh) * 2024-02-07 2024-03-19 宁德时代新能源科技股份有限公司 电芯防爆阀的检测系统及检测方法
CN117718593A (zh) * 2024-02-01 2024-03-19 宁德时代新能源科技股份有限公司 极柱的焊接方法及焊接系统
WO2024066267A1 (zh) * 2022-09-27 2024-04-04 宁德时代新能源科技股份有限公司 检测极耳数量的方法和装置

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111307818B (zh) * 2020-02-25 2021-07-20 华南理工大学 一种锂电池极耳激光焊点视觉在线检测方法
CN111976151A (zh) * 2020-08-06 2020-11-24 漳浦金盛智能科技有限公司 一种平面口罩机
CN112184816B (zh) * 2020-09-28 2024-03-22 珠海博明视觉科技有限公司 一种基于电池极耳的锂电池柔性定位方法
CN113552133B (zh) * 2021-06-30 2022-11-15 广东利元亨智能装备股份有限公司 一种电芯裙边开裂检测方法及视觉检测装置
CN113469991B (zh) * 2021-07-15 2022-03-18 广东奥普特科技股份有限公司 一种锂电池极耳激光焊点视觉在线检测方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05109858A (ja) * 1991-10-15 1993-04-30 Nec Corp Tabはんだ付け検査装置
CN103487440A (zh) * 2013-08-28 2014-01-01 东莞市三瑞自动化科技有限公司 一种基于机器视觉的电池极耳检测方法及其检测系统
CN103846538A (zh) * 2012-11-29 2014-06-11 上海航天设备制造总厂 图像识别装置及利用所述装置的电池阵焊接系统
KR20150033268A (ko) * 2013-09-24 2015-04-01 주식회사 엘지화학 이차전지의 초음파 용접상태에 관한 비전측정 최적화방법 및 그 장치
CN108355981A (zh) * 2018-01-08 2018-08-03 西安交通大学 一种基于机器视觉的电池连接器质量检测方法
CN109142368A (zh) * 2018-07-23 2019-01-04 广州超音速自动化科技股份有限公司 锂电池极片打皱检测方法及极耳焊接检测系统
CN111307818A (zh) * 2020-02-25 2020-06-19 华南理工大学 一种锂电池极耳激光焊点视觉在线检测方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104439727B (zh) * 2014-11-20 2016-08-17 刘厚德 用于电池激光焊接的夹具
CN106825958B (zh) * 2017-03-16 2018-09-28 深圳市光大激光科技股份有限公司 电芯自动焊接检测装置和方法
CN108280837B (zh) * 2018-01-25 2020-06-26 电子科技大学 基于变换的x射线图像中bga焊球轮廓提取方法
CN109187546A (zh) * 2018-07-23 2019-01-11 广州超音速自动化科技股份有限公司 锂电池极耳糅合检测方法及极耳焊接检测系统
CN109187547A (zh) * 2018-07-23 2019-01-11 广州超音速自动化科技股份有限公司 锂电池极耳焊点焊破检测方法及极耳焊接检测系统
CN110135521A (zh) * 2019-05-28 2019-08-16 陕西何止网络科技有限公司 基于卷积神经网络的极片极耳缺陷检测模型、检测方法及系统
CN110910363A (zh) * 2019-11-15 2020-03-24 上海交通大学 基于机器视觉和深度学习的虚焊检测方法、系统及介质

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH05109858A (ja) * 1991-10-15 1993-04-30 Nec Corp Tabはんだ付け検査装置
CN103846538A (zh) * 2012-11-29 2014-06-11 上海航天设备制造总厂 图像识别装置及利用所述装置的电池阵焊接系统
CN103487440A (zh) * 2013-08-28 2014-01-01 东莞市三瑞自动化科技有限公司 一种基于机器视觉的电池极耳检测方法及其检测系统
KR20150033268A (ko) * 2013-09-24 2015-04-01 주식회사 엘지화학 이차전지의 초음파 용접상태에 관한 비전측정 최적화방법 및 그 장치
CN108355981A (zh) * 2018-01-08 2018-08-03 西安交通大学 一种基于机器视觉的电池连接器质量检测方法
CN109142368A (zh) * 2018-07-23 2019-01-04 广州超音速自动化科技股份有限公司 锂电池极片打皱检测方法及极耳焊接检测系统
CN111307818A (zh) * 2020-02-25 2020-06-19 华南理工大学 一种锂电池极耳激光焊点视觉在线检测方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
WANG SHIWEN: "Automated Optical Inspection of Tabs Welding Defects in Lithium Batteries", CHINESE MASTER'S THESES FULL-TEXT DATABASE, TIANJIN POLYTECHNIC UNIVERSITY, CN, 15 January 2019 (2019-01-15), CN, XP055841854, ISSN: 1674-0246 *

Cited By (37)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112363069B (zh) * 2020-09-18 2023-07-07 万向一二三股份公司 一种锂离子电池极耳断裂检测方法
CN112363069A (zh) * 2020-09-18 2021-02-12 万向一二三股份公司 一种锂离子电池极耳断裂检测方法
CN113744244A (zh) * 2021-09-03 2021-12-03 广东奥普特科技股份有限公司 测量锂电池极片边缘到极耳边缘距离的在线视觉检测系统
CN114037657A (zh) * 2021-10-12 2022-02-11 上海电机学院 一种结合区域生长与环形校正的锂电池极耳缺陷检测方法
CN114594114A (zh) * 2022-03-09 2022-06-07 广东兆众自动化设备有限公司 一种锂电芯全自动在线无损检测方法
CN114549531A (zh) * 2022-04-26 2022-05-27 广州市易鸿智能装备有限公司 一种锂电池卷绕OverHang对中纠偏控制系统及方法
CN114549531B (zh) * 2022-04-26 2022-07-15 广州市易鸿智能装备有限公司 一种锂电池卷绕OverHang对中纠偏控制系统及方法
CN117078580A (zh) * 2022-05-10 2023-11-17 广州镭晨智能装备科技有限公司 焊点检测方法、装置及计算机可读存储介质
CN115100600B (zh) * 2022-06-30 2024-05-31 苏州市新方纬电子有限公司 一种电池包的产线智能检测方法及系统
CN115100600A (zh) * 2022-06-30 2022-09-23 苏州市新方纬电子有限公司 一种电池包的产线智能检测方法及系统
CN115178909A (zh) * 2022-07-11 2022-10-14 苏州汇川控制技术有限公司 焊接质量的检测方法、装置、终端设备与介质
CN115178909B (zh) * 2022-07-11 2023-10-27 苏州汇川控制技术有限公司 焊接质量的检测方法、装置、终端设备与介质
WO2024032002A1 (zh) * 2022-08-10 2024-02-15 宁德时代新能源科技股份有限公司 电芯检测方法、装置、系统、计算机设备和存储介质
CN115829913A (zh) * 2022-08-10 2023-03-21 宁德时代新能源科技股份有限公司 裸电芯外观检测方法、装置、计算机设备和存储介质
WO2024032000A1 (zh) * 2022-08-10 2024-02-15 宁德时代新能源科技股份有限公司 裸电芯外观检测方法、装置、计算机设备和存储介质
CN115829913B (zh) * 2022-08-10 2024-05-17 宁德时代新能源科技股份有限公司 裸电芯外观检测方法、装置、计算机设备和存储介质
CN115082504B (zh) * 2022-08-23 2022-11-11 菏泽学院 一种太阳能光伏板光斑识别方法
CN115082504A (zh) * 2022-08-23 2022-09-20 菏泽学院 一种太阳能光伏板光斑识别方法
CN115360319B (zh) * 2022-09-20 2023-04-14 楚能新能源股份有限公司 一种多极耳电芯极片信息追溯方法
CN115360319A (zh) * 2022-09-20 2022-11-18 楚能新能源股份有限公司 一种多极耳电芯极片信息追溯方法
WO2024066267A1 (zh) * 2022-09-27 2024-04-04 宁德时代新能源科技股份有限公司 检测极耳数量的方法和装置
CN115661040A (zh) * 2022-09-28 2023-01-31 广州市易鸿智能装备有限公司 一种锂电池电芯涂胶的检测方法及装置、电子设备
CN116008294B (zh) * 2022-12-13 2024-03-08 无锡微准科技有限公司 一种基于机器视觉的键帽表面颗粒缺陷检测方法
CN116008294A (zh) * 2022-12-13 2023-04-25 无锡微准科技有限公司 一种基于机器视觉的键帽表面颗粒缺陷检测方法
CN116596932B (zh) * 2023-07-18 2024-02-09 北京阿丘机器人科技有限公司 电池顶盖极柱外观检测方法、装置、设备及存储介质
CN116596932A (zh) * 2023-07-18 2023-08-15 北京阿丘机器人科技有限公司 电池顶盖极柱外观检测方法、装置、设备及存储介质
CN117109447B (zh) * 2023-10-24 2024-01-05 钛玛科(北京)工业科技有限公司 自适应极耳宽度检测方法、装置及设备
CN117109447A (zh) * 2023-10-24 2023-11-24 钛玛科(北京)工业科技有限公司 自适应极耳宽度检测方法、装置及设备
CN117309889A (zh) * 2023-11-27 2023-12-29 宁德时代新能源科技股份有限公司 极片边缘检测装置及方法
CN117309889B (zh) * 2023-11-27 2024-05-14 宁德时代新能源科技股份有限公司 极片边缘检测装置及方法
CN117570852A (zh) * 2024-01-15 2024-02-20 钛玛科(北京)工业科技有限公司 极耳顶点坐标检测方法、装置及设备
CN117570852B (zh) * 2024-01-15 2024-03-26 钛玛科(北京)工业科技有限公司 极耳顶点坐标检测方法、装置及设备
CN117593309A (zh) * 2024-01-19 2024-02-23 常熟理工学院 基于立体重构的芯片焊点检测方法、系统及存储介质
CN117593309B (zh) * 2024-01-19 2024-03-19 常熟理工学院 基于立体重构的芯片焊点检测方法、系统及存储介质
CN117718593A (zh) * 2024-02-01 2024-03-19 宁德时代新能源科技股份有限公司 极柱的焊接方法及焊接系统
CN117718593B (zh) * 2024-02-01 2024-06-11 宁德时代新能源科技股份有限公司 极柱的焊接方法及焊接系统
CN117723491A (zh) * 2024-02-07 2024-03-19 宁德时代新能源科技股份有限公司 电芯防爆阀的检测系统及检测方法

Also Published As

Publication number Publication date
CN111307818B (zh) 2021-07-20
CN111307818A (zh) 2020-06-19

Similar Documents

Publication Publication Date Title
WO2021169335A1 (zh) 一种锂电池极耳激光焊点视觉在线检测方法
CN111462110A (zh) 焊缝质量检测方法、装置、系统及电子设备
WO2021168733A1 (zh) 缺陷图像的缺陷检测方法、装置及计算机可读存储介质
CN113393426B (zh) 一种轧钢板表面缺陷检测方法
CN113834816A (zh) 基于机器视觉的光伏电池缺陷在线检测方法及系统
US20110199491A1 (en) Calibration index determination device, calibration device, calibration performance evaluation device, system, method, and program
CN116468687A (zh) 划伤缺陷检测方法、装置及存储介质、电子设备
CN115272168A (zh) 一种极片质量检测方法、系统、设备及存储介质
CN110426395B (zh) 一种太阳能el电池硅片表面检测方法及装置
CN115752969A (zh) 一种铝箔封口密封性检测方法及系统和设备
CN115587993A (zh) 一种基于ai与机器视觉算法的口罩焊点缺陷检测方法
CN117058411B (zh) 电池棱边外形瑕疵识别方法、装置、介质和设备
CN115213563B (zh) 激光智能焊接方法及系统
CN115984360A (zh) 一种基于图像处理的计算干滩长度的方法及系统
CN112161996A (zh) 一种基于图像识别的光伏电池板焊带烧毁检测方法及系统
CN114494142A (zh) 一种基于深度学习的移动终端中框缺陷检测方法及装置
CN117690846B (zh) 太阳能硅片视觉检测方法、装置、设备及存储介质
CN112132807B (zh) 一种基于颜色相似度分割的焊缝区域提取方法和装置
CN115345877B (zh) 一种用于太阳能电池板虚焊异常的检测方法
CN117474902B (zh) 一种桶面料包的缺包检测方法、系统、设备和存储介质
CN116448666B (zh) 一种基于机器视觉的锂电池盒的检测装置及检测方法
CN112364783B (zh) 一种零件检测方法、装置及计算机可读存储介质
WO2024000373A1 (zh) 检测焊印的方法、装置和电子设备
CN114549397A (zh) 一种基于图像处理技术的焊印识别检测长宽的方法
CN116309383A (zh) 基于机器视觉的载带检测方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20921960

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20921960

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 13/02/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20921960

Country of ref document: EP

Kind code of ref document: A1