CN113092489A - System and method for detecting appearance defects of battery - Google Patents
System and method for detecting appearance defects of battery Download PDFInfo
- Publication number
- CN113092489A CN113092489A CN202110551612.XA CN202110551612A CN113092489A CN 113092489 A CN113092489 A CN 113092489A CN 202110551612 A CN202110551612 A CN 202110551612A CN 113092489 A CN113092489 A CN 113092489A
- Authority
- CN
- China
- Prior art keywords
- image
- appearance
- battery
- side light
- module
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 230000007547 defect Effects 0.000 title claims abstract description 105
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000001514 detection method Methods 0.000 claims abstract description 76
- 238000003384 imaging method Methods 0.000 claims abstract description 62
- 238000005286 illumination Methods 0.000 claims abstract description 20
- 238000007499 fusion processing Methods 0.000 claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 230000002194 synthesizing effect Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000013135 deep learning Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 4
- 239000002537 cosmetic Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 11
- 238000012549 training Methods 0.000 description 5
- 238000002372 labelling Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 239000007788 liquid Substances 0.000 description 3
- 230000037303 wrinkles Effects 0.000 description 3
- 238000005034 decoration Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 239000003292 glue Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011179 visual inspection Methods 0.000 description 2
- 239000000853 adhesive Substances 0.000 description 1
- 230000001070 adhesive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 208000003464 asthenopia Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 235000003642 hunger Nutrition 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000037351 starvation Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8806—Specially adapted optical and illumination features
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8806—Specially adapted optical and illumination features
- G01N2021/8835—Adjustable illumination, e.g. software adjustable screen
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan 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
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/8867—Grading and classifying of flaws using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing
- G01N2021/887—Grading and classifying of flaws using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing the measurements made in two or more directions, angles, positions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan 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
- G01N2021/8883—Scan 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 involving the calculation of gauges, generating models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan 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
- G01N2021/8887—Scan 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
Landscapes
- Chemical & Material Sciences (AREA)
- Biochemistry (AREA)
- Pathology (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Analytical Chemistry (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention provides a system and a method for detecting appearance defects of a battery, which comprises the following steps: the system comprises an imaging module, an image preprocessing module and a defect detection module; the imaging module comprises a bearing table, an imaging camera and a side light source group, the bearing table is used for bearing the detected battery, the imaging camera and the side light source group are positioned above the bearing table, the side light source group comprises 4 side light sources, and the side light sources face the bearing table and are arranged around the bearing table; the imaging module is used for acquiring a plurality of appearance images of the detected battery shot under various illumination environments; the image preprocessing module is used for carrying out fusion processing on the plurality of appearance images to obtain a high-dimensional image; and the defect detection module is used for inputting the high-dimensional image into the appearance defect detection model to obtain the appearance defects of the detected battery. The invention can improve the detection rate of automatic defect detection, reduce the omission factor and improve the detection coverage rate of defect types.
Description
Technical Field
The invention relates to the technical field of battery detection, in particular to a system and a method for detecting appearance defects of a battery.
Background
In the production process of the battery, defects such as surface dirt, scratch, concave-convex points and the like cannot be avoided in the appearance of a small part of the battery, so that the appearance is influenced, and the production and use safety of the battery is greatly influenced. Therefore, the appearance of the battery must be detected after the battery is produced, and the qualified rate of the battery appearance is ensured.
At present, the appearance of the battery is generally detected by adopting manual visual inspection. The manual detection mode is influenced by individual differences of detection personnel, and the consistency of detection standards is difficult to ensure; the device is also easily affected by visual fatigue, and the consistency of detection is difficult to ensure; the defect degree can not be measured digitally by manual visual inspection, and digital management and data analysis are difficult to carry out.
In order to solve the problems, the appearance of the battery is automatically photographed through a camera, and then the appearance defect condition is obtained through analysis according to the photographed image. However, the method cannot simulate a multi-angle and multi-illumination detection mode of human eyes, and causes application problems of low detection rate, frequent missed detection and the like.
Disclosure of Invention
One of the objectives of the present invention is to overcome at least one of the deficiencies of the prior art and to provide a system and a method for detecting appearance defects of a battery.
The technical scheme provided by the invention is as follows:
a system for detecting cosmetic defects in a battery, comprising: the system comprises an imaging module, an image preprocessing module and a defect detection module; the imaging module comprises a bearing table, an imaging camera and a side light source group, wherein the bearing table is used for bearing a detected battery, the imaging camera and the side light source group are positioned above the bearing table, the side light source group comprises 4 side light sources, and the side light sources face the bearing table and are arranged around the bearing table; the imaging module is used for acquiring a plurality of appearance images of the detected battery shot under various illumination environments; the image preprocessing module is used for carrying out fusion processing on the plurality of appearance images to obtain a high-dimensional image; and the defect detection module is used for inputting the high-dimensional image into an appearance defect detection model to obtain the appearance defect of the detected battery.
Further, a right angle is formed between the extending directions of any two adjacent side light sources.
Furthermore, the imaging module is also used for acquiring a low exposure image, a high exposure image and a luminosity image of the detected battery; the image preprocessing module is further configured to perform fusion processing on the low exposure image, the high exposure image and the luminosity image to obtain a high-dimensional image.
Furthermore, the imaging module is also used for obtaining single-angle images of all angles by successively starting one side light source and shooting; and synthesizing the photometric image according to all the obtained single-angle images.
Furthermore, the imaging module is further configured to shoot the detected battery to obtain a low-exposure image when all the side light sources are simultaneously turned on and are in a low-light state.
Furthermore, the imaging module is further configured to shoot the detected battery to obtain a high-exposure image when all the side light sources are simultaneously turned on and are in a high-brightness state.
Further, still include: and the model construction module is used for constructing an appearance defect detection model based on the deep learning network.
The invention also provides a method for detecting the appearance defects of the battery, which is applied to the detection system and comprises the following steps: acquiring a plurality of appearance images of a detected battery shot under various illumination environments; performing fusion processing on the plurality of appearance images to obtain a high-dimensional image; and inputting the high-dimensional image into an appearance defect detection model to obtain the appearance defect of the detected battery.
Further, the acquiring of the appearance image of the detected battery taken in various lighting environments includes: acquiring a low exposure image, a high exposure image and a luminosity image of the detected battery;
the process of fusing the appearance image to obtain a high-dimensional image comprises the following steps: and carrying out fusion processing on the high exposure image, the low exposure image and the luminosity image to obtain a high-dimensional image.
Further, acquiring a photometric image of the detected battery includes: obtaining single-angle images of all angles by sequentially starting one side light source and shooting; and synthesizing the photometric image according to all the obtained single-angle images.
The system and the method for detecting the appearance defects of the battery provided by the invention can at least bring the following beneficial effects:
1. the method and the device have the advantages that the detected battery is shot under the multi-illumination environment and multi-angle detection, various defects are imaged in the picture as obviously as possible, the detection rate of automatic defect detection is improved, the omission ratio is reduced, and the detection coverage rate of the defect types is improved.
2. According to the invention, one side light source is turned on one by one, fixed-point shooting can be carried out by using one camera, multi-angle detection simulating human eyes is obtained, the shooting efficiency is improved, and thus the automatic detection efficiency is further improved.
Drawings
The foregoing features, technical features, advantages and implementations of a system and method for detecting appearance defects of a battery will be further described in the following detailed description of preferred embodiments in a clearly understandable manner with reference to the accompanying drawings.
FIG. 1 is a schematic diagram of a system for detecting defects in the appearance of a battery according to an embodiment of the present invention;
FIG. 2 is a schematic front view of the imaging module of FIG. 1;
FIG. 3 is a flow chart of one embodiment of a method of detecting cosmetic defects in a battery of the present invention;
FIG. 4 is a diagram of a shooting method applied to an example of an implementation scenario of the present invention;
fig. 5 is a perspective view of the imaging module.
The reference numbers illustrate:
100. the image processing device comprises an imaging module, 200, an image preprocessing module, 300, a defect detection module, 110, a bearing table, 120, an imaging camera, 130, a side light source group and 131, a side light source.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
For the sake of simplicity, the drawings only schematically show the parts relevant to the present invention, and they do not represent the actual structure as a product. In addition, in order to make the drawings concise and understandable, components having the same structure or function in some of the drawings are only schematically depicted, or only one of them is labeled. In this document, "one" means not only "only one" but also a case of "more than one".
In one embodiment of the present invention, as shown in fig. 1 and 2, a system for detecting appearance defects of a battery includes:
the system comprises an imaging module 100, an image preprocessing module 200 and a defect detection module 300;
the imaging module 100 includes a carrying platform 110, an imaging camera 120 and a side light source set 130, the carrying platform 110 is used for carrying the battery to be detected, and the imaging camera 120 and the side light source set 130 are located above the carrying platform 110; the side light source set 130 includes 4 side light sources 131, and the 4 side light sources 131 face the carrier table 110 and are disposed around the carrier table 110.
Preferably, a right angle is formed between the extending directions of any two adjacent side light sources. For example, the 4 side light sources 131 are arranged at 90 degrees intervals and are respectively positioned at the front, the back, the left and the right of the bearing table.
The imaging camera 120 is used for photographing and the side light source 131 is used for providing a lighting environment.
The imaging module 100 is configured to obtain appearance images of the detected battery captured in various lighting environments. The lighting environment includes a normal-brightness environment, a low-brightness environment, and a high-brightness environment. At least one appearance image of the detected battery is obtained under different illumination environments.
The imaging effect of the same defect is different under different illumination environments, some defects are obvious in imaging under a high-brightness environment, and some defects are obvious in imaging under a low-brightness environment. Therefore, compared with the shooting in a single illumination environment, the image shot in a multi-illumination environment can obtain more defect imaging information of the detected battery.
In one embodiment, the imaging module 100 is used for acquiring a low exposure image, a high exposure image and a photometric image of a battery to be tested.
The photometric image is an appearance image of the inspected battery photographed under a normal brightness environment according to a first exposure time. The photometric image has a good imaging effect on first defects, and the first defects are defects of changes in the height direction of the surface of the battery and comprise bubbles, pits, foreign matters and wrinkles.
The low exposure image is an appearance image of the battery to be tested, which is shot under a low-brightness environment according to the second exposure time. The second exposure time is less than the first exposure time. Low exposure images have better imaging of the second category of defects, which include print breaks, missing ink, ghosting, and blurring. For example, the first exposure time is 7000us and the second exposure time is 3000 us.
The high exposure image is an appearance image of the battery to be tested, which is shot under a high brightness environment according to the third exposure time. The third exposure time is greater than the first exposure time. The high exposure image has a good imaging effect on the third defects, and the third defects comprise black dirt on the surface of the battery, liquid leakage, fingerprints, residual glue and poor package. For example, the first exposure time is 7000us and the third exposure time is 8000 us.
By acquiring the luminosity image, the low exposure image and the high exposure image, the appearance image of the detected battery under various illumination environments can be obtained.
Further, in order to obtain more defect imaging information, multi-angle detection of human eyes is simulated, and a plurality of single-angle images obtained by shooting the detected battery from different angles under normal brightness are obtained; and obtaining a luminosity image according to the single-angle images, wherein the luminosity image is fused with the detection information of each angle.
Images at different angles can be acquired by disposing imaging cameras at a plurality of angles, respectively, or by disposing one imaging camera and adjusting the position of the imaging camera. The former increases the cost, and the latter increases the complexity of the photographing operation, which is not favorable for improving the photographing efficiency.
For this reason, the following improvements are made:
as shown in fig. 5, fixed-point photographing is performed using one camera without changing the angle of view.
Obtaining single-angle images of all angles by successively starting a side light source and shooting a detected battery; and synthesizing a photometric image according to all the obtained single-angle images.
Specifically, one side light source may be turned on at a time, photographed, turned off, turned on again, and photographed again until all side light sources are turned on. The detected battery can be shot according to the first exposure time under the condition that the side light source is in normal brightness, and a corresponding single-angle image is obtained.
One round can be carried out to obtain four single-angle images; the images obtained at the same angle can be integrated by repeating multiple rounds to obtain single-angle images at 4 angles; and then the single-angle images of the four angles are combined into a photometric image.
Each single angle image is equivalent to an image that simulates the human eye looking from a single angle. By opening a lateral light source and shooting one by one, the battery is detected from four angles of front, back, left and right equivalently respectively, so that the overall appearance of the battery defect can be obtained, and the coverage rate and the detectable rate of the defect detection are improved.
And when all the side light sources are simultaneously started and are in a low-light state, the imaging camera shoots the detected battery to obtain a low-exposure image. And when all the side light sources are simultaneously started and are in a high-brightness state, shooting the detected battery to obtain a high-exposure image.
The image preprocessing module 200 is configured to perform fusion processing on the obtained multiple appearance images to obtain a high-dimensional image. The high-dimensional image fuses the features of the multiple appearance images.
And if the appearance image comprises a luminosity image, a low exposure image and a high exposure image, performing fusion processing on the luminosity image, the low exposure image and the high exposure image to obtain a high-dimensional image.
And a defect detection module 300, configured to input the high-dimensional image into an appearance defect detection model to obtain an appearance defect of the detected battery.
And the model building module 400 is used for building an appearance defect detection model based on the neural network. A construction process of a supervised deep learning network can be adopted to carry out processes such as sample marking, sample training, parameter tuning and the like, so as to obtain an appearance defect detection model meeting the preset requirements.
In the embodiment, the detected battery is shot under the multi-illumination environment and multi-angle detection, so that various defects are imaged in the picture as obviously as possible, the detection rate of automatic defect detection is improved, the omission factor is reduced, and the detection coverage rate of defect types is improved; by gradually starting one lateral light source, a camera can be used for fixed-point shooting, multi-angle detection of human eyes is simulated, shooting efficiency is improved, and therefore automatic detection efficiency is further improved.
In an embodiment of the present invention, as shown in fig. 3, a method for detecting an appearance defect of a battery based on the foregoing system for detecting an appearance defect of a battery includes:
step S100 acquires appearance images of the battery under inspection taken under various lighting environments.
The lighting environment includes a normal-brightness environment, a low-brightness environment, and a high-brightness environment. At least one appearance image of the detected battery is obtained under different illumination environments.
The imaging effect of the same defect is different under different illumination environments, some defects are obvious in imaging under a high-brightness environment, and some defects are obvious in imaging under a low-brightness environment. Therefore, compared with the shooting in a single illumination environment, the image shot in a multi-illumination environment can obtain more defect imaging information of the detected battery.
The step S100 includes:
step S110 acquires a low exposure image, a high exposure image, and a photometric image of the battery under test.
The photometric image is an appearance image of the inspected battery photographed under a normal brightness environment according to a first exposure time. The photometric image has a good imaging effect on first defects, and the first defects are defects of changes in the height direction of the surface of the battery and comprise bubbles, pits, foreign matters and wrinkles.
The low exposure image is an appearance image of the battery to be tested, which is shot under a low-brightness environment according to the second exposure time. The second exposure time is less than the first exposure time. Low exposure images have better imaging of the second category of defects, which include print breaks, missing ink, ghosting, and blurring. For example, the first exposure time is 7000us and the second exposure time is 3000 us.
The high exposure image is an appearance image of the battery to be tested, which is shot under a high brightness environment according to the third exposure time. The third exposure time is greater than the first exposure time. The high exposure image has a good imaging effect on the third defects, and the third defects comprise black dirt on the surface of the battery, liquid leakage, fingerprints, residual glue and poor package. For example, the first exposure time is 7000us and the third exposure time is 8000 us.
By acquiring the luminosity image, the low exposure image and the high exposure image, the appearance image of the detected battery under various illumination environments can be obtained.
Further, in order to obtain more defect imaging information, multi-angle detection of human eyes is simulated, and a plurality of single-angle images obtained by shooting the detected battery from different angles under normal brightness are obtained; and obtaining a luminosity image according to the single-angle images, wherein the luminosity image is fused with the detection information of each angle.
Images at different angles can be acquired by disposing imaging cameras at a plurality of angles, respectively, or by disposing one imaging camera and adjusting the position of the imaging camera. The former increases the cost, and the latter increases the complexity of the photographing operation, which is not favorable for improving the photographing efficiency.
For this reason, the following improvements are made:
as shown in fig. 5, fixed-point photographing is performed using one camera without changing the angle of view.
And sequentially starting a side light source for the side light source group, and shooting the detected battery according to the first exposure time under normal brightness to obtain a single-angle image. If the first round is carried out, four single-angle images are obtained; the images obtained at the same angle can be integrated by repeating multiple rounds to obtain single-angle images at 4 angles; and then the single-angle images of the four angles are combined into a photometric image.
Each single angle image is equivalent to an image that simulates the human eye looking from a single angle. By opening a lateral light source and shooting one by one, the battery is detected from four angles of front, back, left and right equivalently respectively, so that the overall appearance of the battery defect can be obtained, and the coverage rate and the detectable rate of the defect detection are improved.
And when all the side light sources are simultaneously started and are in a low-light state, the imaging camera shoots the detected battery to obtain a low-exposure image. And when all the side light sources are simultaneously started and are in a high-brightness state, shooting the detected battery to obtain a high-exposure image.
Step S200 performs fusion processing on the obtained multiple appearance images to obtain a high-dimensional image.
And if the appearance image comprises a luminosity image, a low exposure image and a high exposure image, performing fusion processing on the luminosity image, the low exposure image and the high exposure image to obtain a high-dimensional image, wherein the high-dimensional image is fused with the characteristics of the three images.
Step S300, inputting the high-dimensional image into an appearance defect detection model to obtain the appearance defects of the detected battery.
Step S400 constructs an appearance defect detection model based on the neural network.
A construction process of a supervised deep learning network can be adopted to carry out processes such as sample marking, sample training, parameter tuning and the like, so as to obtain an appearance defect detection model meeting the preset requirements.
In the embodiment, the detected battery is shot under the multi-illumination environment and multi-angle detection, so that various defects are imaged in the picture as obviously as possible, the detection rate of automatic defect detection is improved, the omission factor is reduced, and the detection coverage rate of defect types is improved; by gradually starting one lateral light source, a camera can be used for fixed-point shooting, multi-angle detection of human eyes is simulated, shooting efficiency is improved, and therefore automatic detection efficiency is further improved.
The invention also provides a concrete implementation scene example, and the system and the method for detecting the appearance defects of the battery provided by the invention are applied to the appearance detection of the soft package battery, and a main flow chart is as follows, and comprises the following steps:
step S1: the front side and the back side of the battery are photographed by adopting three photographing methods respectively to obtain a front side low exposure picture, a front side high exposure picture, a front side angle picture, a back side low exposure picture, a back side high exposure picture and a back side angle picture.
The arrangement of the side light source, imaging camera and the carrier table is shown in fig. 5.
As shown in fig. 4, three shooting methods are specifically as follows:
the first method comprises the following steps: and taking a low exposure picture (four light sources are turned on and four light sources are low and bright).
And (3) starting the light source 1-4, taking the 1 st picture by using the exposure value of the camera of 3000us, and obtaining a low exposure picture.
And the second method comprises the following steps: and taking a high exposure picture one by one (four light sources are turned on and four light sources are highlighted).
And (4) starting the light source 1-4, taking the 2 nd picture with the camera exposure value of 8000us, and obtaining a high exposure picture.
And the third is that: and (3) shooting a plurality of pictures at different angles, and synthesizing to obtain one luminosity picture (at least 4 pictures are shot, the light sources are spaced at 90 degrees and are sequentially lightened for shooting).
Only the light source 1 is turned on, the exposure value of the camera is 7000us, and the 3 rd picture is taken; only the light source 2 is turned on, the exposure value of the camera is 7000us, and the 4 th picture is taken; only the light source 3 is turned on, the camera exposure value is 7000us, and the 5 th picture is taken; only the light source 4 is turned on, the exposure value of the camera is 7000us, and the 6 th picture is taken; and combining the 3 rd to 6 th pictures into a luminosity picture.
The luminosity picture has a better imaging effect on the first type of defects, the low-exposure picture has a better imaging effect on the second type of defects, and the high-exposure picture has a better imaging effect on the third type of defects. The first type of defects are variation defects in the height direction of the surface of the battery, including bubbles, pits, foreign matter, and wrinkles. The second category of defects includes print breaks, ink starvation, ghosting, and blurring. The third type of defects comprises black dirt on the surface of the battery, liquid leakage, fingerprints, adhesive residues and poor package.
The camera is used for fixed-point shooting, the visual angle does not need to be changed, the efficiency of a shooting link can be improved, and automatic shooting is facilitated.
Note that the above shooting sequence is only an example, and the shooting sequence can be adjusted as necessary.
Step S2: and (5) preprocessing the picture.
And synthesizing the front high-exposure picture, the front low-exposure picture and the front photometric picture into a front high-dimensional picture.
And synthesizing the back side high exposure picture, the back side low exposure picture and the back side luminosity picture into a back side high-dimensional picture.
Each high-exposure picture, each low-exposure picture and each luminosity picture are single-channel N-bit images (namely each pixel is represented by N bits, for example, N is 8), the three pictures are fused, each pixel corresponds to one another, and a three-channel N-bit (3 x N bit depth in total) image is constructed to obtain a high-dimensional image. Thus, the synthesized high-dimensional picture integrates the characteristics of a high-exposure picture, a low-exposure picture and a luminosity picture.
Step S3: and (4) neural network processing.
And constructing a deep learning network of multi-channel input.
Obtaining a large number of unfused high-exposure pictures, low-exposure pictures and luminosity pictures as training samples in advance, respectively labeling the training samples, wherein the high-exposure pictures are labeled with a third type of defects, the low-exposure pictures are labeled with a second type of defects, the luminosity pictures are labeled with a first type of defects, and three labeling results of the three pictures are fused to generate a total labeling file, wherein the total labeling file comprises position coordinate information and category information of the first-third types of defects. According to the training process of the supervised deep learning network, a backbone network structure of the learning network is modified into a multi-channel input network structure, the fused high-dimensional pictures and the fused marking files are input into the modified network, the network is trained, parameters are adjusted and optimized, and the appearance defect detection model meeting the preset requirements is obtained.
In the actual appearance detection, according to the shooting and preprocessing links, high-dimensional images of the front surface and the back surface are input for the appearance defect detection model. The model carries out reasoning according to the parameters of the model and the input high-dimensional image to finish the position and type detection of the defects of the front side and the back side.
In the embodiment, different from the traditional photographing detection mode, various defects are imaged in the picture as obviously as possible by utilizing the fusion of three photographing methods; the method has good detection effect on the first to third defects, and the detection coverage rate of the defect types is high; the existing optical detection method has the characteristics of high detection stability, good repeatability and the like, and obtains a good detection effect in practical application.
It should be noted that the above embodiments can be freely combined as necessary. The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A system for detecting cosmetic defects in a battery, comprising:
the system comprises an imaging module, an image preprocessing module and a defect detection module;
the imaging module comprises a bearing table, an imaging camera and a side light source group, wherein the bearing table is used for bearing a detected battery, the imaging camera and the side light source group are positioned above the bearing table, the side light source group comprises 4 side light sources, and the side light sources face the bearing table and are arranged around the bearing table;
the imaging module is used for acquiring a plurality of appearance images of the detected battery shot under various illumination environments;
the image preprocessing module is used for carrying out fusion processing on the plurality of appearance images to obtain a high-dimensional image;
and the defect detection module is used for inputting the high-dimensional image into an appearance defect detection model to obtain the appearance defect of the detected battery.
2. The detection system of claim 1, wherein:
the extending directions of any two adjacent side light sources form a right angle.
3. The detection system of claim 1, wherein:
the imaging module is also used for acquiring a low exposure image, a high exposure image and a luminosity image of the detected battery;
the image preprocessing module is further configured to perform fusion processing on the low exposure image, the high exposure image and the luminosity image to obtain a high-dimensional image.
4. The detection system of claim 3, wherein:
the imaging module is also used for obtaining single-angle images of all angles by successively starting one side light source and shooting; and synthesizing the photometric image according to all the obtained single-angle images.
5. The detection system of claim 3, wherein:
the imaging module is also used for shooting the detected battery to obtain a low-exposure image when all the side light sources are simultaneously turned on and are in a low-light state.
6. The detection system of claim 3, wherein:
the imaging module is also used for shooting the detected battery to obtain a high-exposure image when all the side light sources are simultaneously started and are in a high-brightness state.
7. The detection system of claim 1, further comprising:
and the model construction module is used for constructing an appearance defect detection model based on the deep learning network.
8. A method for detecting defects in the appearance of a battery, the method being applied to the detection system according to any one of claims 1 to 7, the method comprising:
acquiring a plurality of appearance images of a detected battery shot under various illumination environments;
performing fusion processing on the plurality of appearance images to obtain a high-dimensional image;
and inputting the high-dimensional image into an appearance defect detection model to obtain the appearance defect of the detected battery.
9. The method for detecting the appearance defect of the battery according to claim 8, wherein:
the acquiring of the appearance image of the detected battery shot under various illumination environments comprises:
acquiring a low exposure image, a high exposure image and a luminosity image of the detected battery;
the process of fusing the appearance image to obtain a high-dimensional image comprises the following steps:
and carrying out fusion processing on the high exposure image, the low exposure image and the luminosity image to obtain a high-dimensional image.
10. The method for detecting battery appearance defects according to claim 9, wherein acquiring a photometric image of the detected battery comprises:
obtaining single-angle images of all angles by sequentially starting one side light source and shooting;
and synthesizing the photometric image according to all the obtained single-angle images.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110551612.XA CN113092489A (en) | 2021-05-20 | 2021-05-20 | System and method for detecting appearance defects of battery |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110551612.XA CN113092489A (en) | 2021-05-20 | 2021-05-20 | System and method for detecting appearance defects of battery |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113092489A true CN113092489A (en) | 2021-07-09 |
Family
ID=76665940
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110551612.XA Pending CN113092489A (en) | 2021-05-20 | 2021-05-20 | System and method for detecting appearance defects of battery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113092489A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115436288A (en) * | 2022-09-20 | 2022-12-06 | 山东大学 | Lens defect detection device and imaging method |
| CN116051542A (en) * | 2023-03-06 | 2023-05-02 | 深圳市深视智能科技有限公司 | Defect detection method and defect detection device |
| CN116908212A (en) * | 2023-09-12 | 2023-10-20 | 厦门微亚智能科技股份有限公司 | Battery cell blue film appearance defect detection method and system based on feature extraction |
| CN117705814A (en) * | 2024-02-05 | 2024-03-15 | 宁德时代新能源科技股份有限公司 | Battery shell cover welding detection system and method |
| CN117825395A (en) * | 2024-03-06 | 2024-04-05 | 宁德时代新能源科技股份有限公司 | Bare cell defect detection system and method |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020001405A1 (en) * | 2000-06-30 | 2002-01-03 | Nidek Co., Ltd. | Defect inspection method and defect inspection apparatus |
| US20110176024A1 (en) * | 2010-01-15 | 2011-07-21 | Samsung Electronics Co., Ltd. | Image Fusion Apparatus and Method |
| CN106375675A (en) * | 2016-08-30 | 2017-02-01 | 中国科学院长春光学精密机械与物理研究所 | Aerial camera multi-exposure image fusion method |
| CN107607546A (en) * | 2017-09-19 | 2018-01-19 | 佛山缔乐视觉科技有限公司 | Leather defect inspection method, system and device based on photometric stereo vision |
| CN110286134A (en) * | 2019-07-26 | 2019-09-27 | 上海御微半导体技术有限公司 | A kind of defect detecting device and its method |
| CN110335221A (en) * | 2019-03-21 | 2019-10-15 | 西安电子科技大学 | A kind of more exposure image fusion methods based on unsupervised learning |
| CN110473178A (en) * | 2019-07-30 | 2019-11-19 | 上海深视信息科技有限公司 | A kind of open defect detection method and system based on multiple light courcess fusion |
| CN110473179A (en) * | 2019-07-30 | 2019-11-19 | 上海深视信息科技有限公司 | A kind of film surface defects detection method, system and equipment based on deep learning |
| KR20200007292A (en) * | 2018-07-12 | 2020-01-22 | (주) 인텍플러스 | Apparatus for inspecting exterior of battery |
| CN111896550A (en) * | 2020-03-31 | 2020-11-06 | 广西师范大学 | Surface defect detection device and method |
| CN112240887A (en) * | 2020-12-14 | 2021-01-19 | 惠州高视科技有限公司 | Battery appearance defect detection system and method |
| CN112258434A (en) * | 2020-09-30 | 2021-01-22 | 广东电网有限责任公司广州供电局 | Detail-preserving multi-exposure image fusion algorithm in static scene |
| CN112345539A (en) * | 2020-11-05 | 2021-02-09 | 菲特（天津）检测技术有限公司 | Aluminum die casting surface defect detection method based on deep learning |
| CN112630230A (en) * | 2020-10-20 | 2021-04-09 | 天津大学 | Online surface defect detection method based on photometric stereo method |
| CN112821868A (en) * | 2020-12-31 | 2021-05-18 | 浙江爱旭太阳能科技有限公司 | Control method of PL detection system |
-
2021
- 2021-05-20 CN CN202110551612.XA patent/CN113092489A/en active Pending
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020001405A1 (en) * | 2000-06-30 | 2002-01-03 | Nidek Co., Ltd. | Defect inspection method and defect inspection apparatus |
| US20110176024A1 (en) * | 2010-01-15 | 2011-07-21 | Samsung Electronics Co., Ltd. | Image Fusion Apparatus and Method |
| CN106375675A (en) * | 2016-08-30 | 2017-02-01 | 中国科学院长春光学精密机械与物理研究所 | Aerial camera multi-exposure image fusion method |
| CN107607546A (en) * | 2017-09-19 | 2018-01-19 | 佛山缔乐视觉科技有限公司 | Leather defect inspection method, system and device based on photometric stereo vision |
| KR20200007292A (en) * | 2018-07-12 | 2020-01-22 | (주) 인텍플러스 | Apparatus for inspecting exterior of battery |
| CN110335221A (en) * | 2019-03-21 | 2019-10-15 | 西安电子科技大学 | A kind of more exposure image fusion methods based on unsupervised learning |
| CN110286134A (en) * | 2019-07-26 | 2019-09-27 | 上海御微半导体技术有限公司 | A kind of defect detecting device and its method |
| CN110473178A (en) * | 2019-07-30 | 2019-11-19 | 上海深视信息科技有限公司 | A kind of open defect detection method and system based on multiple light courcess fusion |
| CN110473179A (en) * | 2019-07-30 | 2019-11-19 | 上海深视信息科技有限公司 | A kind of film surface defects detection method, system and equipment based on deep learning |
| CN111896550A (en) * | 2020-03-31 | 2020-11-06 | 广西师范大学 | Surface defect detection device and method |
| CN112258434A (en) * | 2020-09-30 | 2021-01-22 | 广东电网有限责任公司广州供电局 | Detail-preserving multi-exposure image fusion algorithm in static scene |
| CN112630230A (en) * | 2020-10-20 | 2021-04-09 | 天津大学 | Online surface defect detection method based on photometric stereo method |
| CN112345539A (en) * | 2020-11-05 | 2021-02-09 | 菲特（天津）检测技术有限公司 | Aluminum die casting surface defect detection method based on deep learning |
| CN112240887A (en) * | 2020-12-14 | 2021-01-19 | 惠州高视科技有限公司 | Battery appearance defect detection system and method |
| CN112821868A (en) * | 2020-12-31 | 2021-05-18 | 浙江爱旭太阳能科技有限公司 | Control method of PL detection system |
Non-Patent Citations (1)
| Title |
|---|
| 张军等: "不同曝光值图像的直接融合方法", 软件学报, vol. 22, no. 04, 15 April 2011 (2011-04-15), pages 813 - 825 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115436288A (en) * | 2022-09-20 | 2022-12-06 | 山东大学 | Lens defect detection device and imaging method |
| CN116051542A (en) * | 2023-03-06 | 2023-05-02 | 深圳市深视智能科技有限公司 | Defect detection method and defect detection device |
| CN116908212A (en) * | 2023-09-12 | 2023-10-20 | 厦门微亚智能科技股份有限公司 | Battery cell blue film appearance defect detection method and system based on feature extraction |
| CN116908212B (en) * | 2023-09-12 | 2023-12-08 | 厦门微亚智能科技股份有限公司 | Battery cell blue film appearance defect detection method and system based on feature extraction |
| CN117705814A (en) * | 2024-02-05 | 2024-03-15 | 宁德时代新能源科技股份有限公司 | Battery shell cover welding detection system and method |
| CN117825395A (en) * | 2024-03-06 | 2024-04-05 | 宁德时代新能源科技股份有限公司 | Bare cell defect detection system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113092489A (en) | System and method for detecting appearance defects of battery | |
| CN108683907A (en) | Optics module picture element flaw detection method, device and equipment | |
| JPH09275272A (en) | Soldering check device and method taking advantage of relative neuron network | |
| CN109741307A (en) | Veiling glare detection method, veiling glare detection device and the veiling glare detection system of camera module | |
| US20220222855A1 (en) | System and method for determining whether a camera component is damaged | |
| CN115032196B (en) | Full-scribing high-flux color pathological imaging analysis instrument and method | |
| CN112070747A (en) | LED lamp bead defect detection method and device | |
| CN116128881B (en) | Lens detection system and method for defect product synchronous marking | |
| CN109714591A (en) | Based on the picture quality subjective evaluation method and system to assessment label | |
| CN108989789B (en) | Real-time evaluation method for imaging quality of camera | |
| CN211179500U (en) | Multi-light source optical detection system | |
| JP2006139777A (en) | Method and device for detecting flat panel display device by visual model | |
| CN109741285A (en) | A kind of construction method and system of underwater picture data set | |
| CN112595496A (en) | Method, device and equipment for detecting defects of near-eye display equipment and storage medium | |
| CN113189109A (en) | Flaw judgment system and flaw judgment method based on artificial intelligence | |
| CN108389182B (en) | Image quality detection method and device based on deep neural network | |
| CN109886936A (en) | A kind of low contrast defect inspection method and device | |
| CN109145861A (en) | Emotion identification device and method, head-mounted display apparatus, storage medium | |
| CN114913086B (en) | Face image quality enhancement method based on generation countermeasure network | |
| CN109447965A (en) | A kind of detection method and device showing screen | |
| CN114219758A (en) | Defect detection method, system, electronic device and computer readable storage medium | |
| CN108596901A (en) | A kind of implementation method of intelligence test and appraisal real shooting photo imaging image quality | |
| CN109060831A (en) | A kind of automatic dirty detection method based on bottom plate fitting | |
| US20240096059A1 (en) | Method for classifying images and method for optically examining an object | |
| CN115225876B (en) | Seamless fusion splicing system for multiple laser projectors |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |