CN116843680B - IGBT power module radiator surface defect identification method - Google Patents
IGBT power module radiator surface defect identification method Download PDFInfo
- Publication number
- CN116843680B CN116843680B CN202311097425.4A CN202311097425A CN116843680B CN 116843680 B CN116843680 B CN 116843680B CN 202311097425 A CN202311097425 A CN 202311097425A CN 116843680 B CN116843680 B CN 116843680B
- Authority
- CN
- China
- Prior art keywords
- distribution coefficient
- defect
- coefficient
- image
- radiator
- 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.)
- Active
Links
- 230000007547 defect Effects 0.000 title claims abstract description 202
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000009826 distribution Methods 0.000 claims abstract description 179
- 238000011156 evaluation Methods 0.000 claims abstract description 68
- 230000007797 corrosion Effects 0.000 claims abstract description 48
- 238000005260 corrosion Methods 0.000 claims abstract description 48
- 230000002159 abnormal effect Effects 0.000 claims abstract description 29
- 238000012545 processing Methods 0.000 claims abstract description 27
- 238000001931 thermography Methods 0.000 claims abstract description 7
- 230000000694 effects Effects 0.000 claims description 16
- 238000001914 filtration Methods 0.000 claims description 9
- 230000008439 repair process Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 238000013461 design Methods 0.000 claims description 5
- 230000006872 improvement Effects 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 238000002790 cross-validation Methods 0.000 claims description 4
- 239000000463 material Substances 0.000 claims description 4
- 238000003708 edge detection Methods 0.000 claims description 3
- 238000003331 infrared imaging Methods 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 abstract description 7
- 230000005856 abnormality Effects 0.000 abstract description 6
- 238000004227 thermal cracking Methods 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 description 10
- 238000001514 detection method Methods 0.000 description 6
- 238000012423 maintenance Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000011179 visual inspection Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000003706 image smoothing Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Computational Linguistics (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a method for identifying surface defects of a radiator of an IGBT power module, which relates to the technical field of radiator surface defect image identification, and adopts a high-resolution industrial camera and an infrared thermal imaging camera to acquire images of the surface of the radiator in real time, acquire a plurality of images A and B, extract bubble characteristics, crack characteristics, corrosion characteristics and dirt characteristics in the plurality of images A after image processing through a characteristic extraction algorithm, extract characteristics of the plurality of images B, extract local temperature abnormal characteristics and surface thermal crack distribution characteristics, convert the local temperature abnormal characteristics and the surface thermal crack distribution characteristics into numerical indexes, realize comprehensive defect assessment, and introduce a comprehensive assessment coefficient PG (PG 1 ，PG 2 ) And comprehensively evaluating the characteristics of bubbles, cracks, corrosion, dirt, local temperature abnormality and surface thermal cracking defects, comparing the characteristics with the set standard defect characteristics to obtain an evaluation result, and accurately judging the state and deviation degree of the surface defects of the radiator.
Description
Technical Field
The invention relates to the technical field of radiator surface defect image recognition, in particular to a method for recognizing surface defects of an IGBT power module radiator.
Background
The IGBT power module radiator is an important component used in power electronic equipment, and is mainly used for radiating heat and reducing temperature so as to ensure the normal operation of the power module. These heat sinks are typically exposed to high temperature and pressure environments and are prone to defects such as bubbles, cracks, corrosion, dirt, etc. during long-term operation, resulting in reduced heat dissipation and even affecting the stability and life of the device.
The existence of surface defects of the radiator of the IGBT power module can seriously affect the working performance and reliability of equipment. Thus, early identification and accurate assessment of the extent of defects is critical to maintaining and continuing use of the power module. The conventional defect recognition method may require a lot of time and labor and has a limited accuracy, so that it is necessary to develop a more efficient and accurate defect recognition method.
Current defect identification methods are mostly based on manual visual inspection or use of some specific sensors to obtain surface information. These methods are not only time-consuming and laborious, but also susceptible to human subjective factors, leading to unstable recognition results. Furthermore, existing methods typically focus on only a single defect feature, making it difficult to fully evaluate the defect state of the heat spreader surface.
Disclosure of Invention
(one) solving the technical problems
Aiming at the defects of the prior art, the invention provides a method for identifying the surface defects of an IGBT power module radiator, which aims to solve the problems of low defect identification efficiency, low accuracy, difficult overall assessment of defect states and the like in the prior art. The method is based on a high-resolution industrial camera and an infrared thermal imaging camera, combines an image processing technology and a feature extraction algorithm, and comprehensively identifies and evaluates defects such as bubbles, cracks, corrosion, dirt, local temperature abnormality, surface thermal cracks and the like on the surface of a radiator through comprehensive evaluation coefficients PG (PG 1, PG 2).
(II) technical scheme
In order to achieve the above purpose, the invention is realized by the following technical scheme: the method for identifying the surface defects of the radiator of the IGBT power module comprises the following steps,
shooting and acquiring a plurality of frames of images on the surface of the radiator by adopting a high-resolution industrial camera, acquiring a plurality of frames of images A, and acquiring infrared imaging on the surface of the radiator in real time by adopting an infrared thermal imaging camera to acquire a plurality of frames of images B;
performing image processing on the multi-frame image A and the multi-frame image B, wherein the image processing comprises noise removal, contrast increase, graying processing and filtering processing; and adjusting and calculating a focal size proportion value f (x, y), and dynamically extracting image features according to the focal size proportion value f (x, y);
establishing an identification model, extracting bubble characteristics, crack characteristics, corrosion characteristics and dirt characteristics in a multi-frame image A after image processing, and analyzing and calculating the characteristics to obtain a bubble distribution coefficient Qp and a first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs and fouling distribution coefficient Wg; extracting features of the multi-frame image B after image processing, extracting local temperature abnormal features and surface thermal crack distribution features, and obtaining an abnormal feature distribution coefficient Rd and a second crack distribution coefficient Lw 2 The method comprises the steps of carrying out a first treatment on the surface of the The first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Matching, if the numbers are the same, the identification effect is correct, and if the deviation exceeds a preset deviation value, the identification is needed again;
labeling the corresponding features and classifications of each image according to the existing images and the extracted features, and storing the labeled features and classifications as a total data set SJ;
dividing the marked total data set SJ into a training set and a testing set, training by adopting a cross-validation method, and optimizing model parameters to minimize prediction errors;
the bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Correlating the corrosion distribution coefficient Fs and the dirt distribution coefficient Wg to obtain a first defect evolution evaluation coefficient PG 1 The abnormal characteristic distribution coefficient Rd and the second crack distribution coefficient Lw are calculated 2 Associated, obtain the second defect evolution assessment coefficient PG 2 Estimating the first defect evolution estimation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 Fitting and calculating to obtain comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) The formula of (2) is:
wherein m is expressed as a first defect evolution estimation coefficient PG1 and a second defect evolution estimation coefficient PG 2 W represents a preset error threshold value, and n represents an error threshold value factor;
will be healdSum of estimated coefficients PG (PG) 1 ，PG 2 ) And comparing the defect characteristic with the standard defect characteristic to obtain a corresponding defect compensation scheme.
Preferably, the focal point size ratio f (x, y) is calculated by a two-dimensional gaussian function as:
wherein x and y are represented as the image abscissa and ordinate, and α is the gaussian kernel; f is the focal size proportion of (x, y), a mask is used for scanning each pixel of the image in specific operation, and the value of a pixel point at the center of the mask is the weighted average gray value of pixels in the neighborhood determined by the mask; a Z multiplied by Z mask is adopted, Z is more than or equal to 4, and the obtained filtering effect provides an image with higher image quality for edge detection.
Preferably, the bubble distribution coefficient Qp, the first crack distribution coefficient Lw 1 The corrosion distribution coefficient Fs and the fouling distribution coefficient Wg are calculated by the following formulas:
wherein: np represents the number of detected bubbles in the image, N represents the total number of areas in the image or the total number of pixels of the heat sink surface; the bubble distribution coefficient Qp represents the duty ratio of bubbles on the surface of the radiator; nw1 represents the number of first-kind cracks detected in the imageQuantity, first crack distribution coefficient Lw 1 Representing the duty cycle of a first type of crack on the surface of the radiator; ns represents the number of corrosion areas detected in the image, and the corrosion distribution coefficient Fs represents the duty ratio of the corrosion areas on the radiator surface; ng represents the number of dirt areas detected in the image, and the dirt distribution coefficient Wg represents the duty cycle of the dirt areas on the radiator surface.
Preferably, image B is set to I (x, y), where (x, y) represents the pixel coordinates in the image; in order to extract the local temperature anomaly characteristic, the anomaly characteristic distribution coefficient Rd is obtained by calculation using the following formula:
wherein Σ represents summing all pixels of the image B, μ represents the average temperature of the image B, σ represents the standard deviation of the image B; the abnormal characteristic distribution coefficient Rd represents the degree of deviation of the temperature of each pixel point in the image from the average temperature.
Preferably, the image B is set to be I (x, y), i.e. a single-channel gray-scale image, and the second crack distribution coefficient Lw is obtained by calculating according to the following formula in order to extract the surface thermal crack distribution characteristics 2 ：
Where, Σ represents summing all pixels of image B,a laplace operator representing the image B; second crack distribution coefficient Lw 2 Representing the quadratic square of the gray value rate of change of each pixel in the image.
Preferably, the first crack distribution coefficient Lw is calculated by an absolute error matching method 1 And a second crack distribution coefficient Lw 2 Absolute error, if the error is smaller than the preset deviation value, the identification effect is considered to be correct;
or, calculating the first crack distribution coefficient Lw by adopting a relative error matching method 1 And a second crack distribution coefficient Lw 2 Relative error, i.e. the difference between the values divided by the first crackDistribution coefficient Lw of lines 1 And a second crack distribution coefficient Lw 2 If the relative error is smaller than the preset deviation value, the identification effect is considered to be correct;
or, calculating the first crack distribution coefficient Lw by adopting a root mean square error matching method 1 And a second crack distribution coefficient Lw 2 The root mean square error, namely the average value of the sum of squares of the differences on each pixel point, is considered to be correct if the root mean square error is smaller than a preset deviation value.
Preferably, the first defect evolution estimation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 The method is obtained by calculation through the following association formula:
wherein G, H, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw, respectively 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 The weight coefficient value of (2) is adjusted and set by a user; beta is expressed as a modified natural number.
Preferably, a set of standard defect features is set according to expertise and experimental data, wherein the features represent the features which the radiator should have in an ideal state;
the comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comparing the defect state with the standard defect characteristics to obtain an evaluation result, wherein the evaluation result is used for evaluating the defect state and the deviation degree of the surface of the radiator.
According to the evaluation result, a targeted defect compensation scheme is formulated; the defect remedy includes repair or cleaning measures for bubbles, cracks, corrosion and dirt defects, and also includes improvements suggestions for radiator materials, structures or designs.
Preferably, standard defect features are set, including standard Qp:0.2 (20%), standard Lw 1 :0.15 (15%), standard Fs:0.1 (10%), standard Wg:0.1 (10%), standard Rd:0.12 (12%), standard Lw2:0.08 (8%); G. h, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 G=0.4, h=0.4, c=0.3, d=0.2, e=0.1, f=0.5;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Qp: identifying a poor bubble defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 1 : identifying a first crack defect condition;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Fs: identifying a poor corrosion defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Wg: identifying a poor fouling defect condition;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Rd: identifying abnormal characteristics and poor defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 2 The second crack defect state is identified as poor.
(III) beneficial effects
The invention provides a method for identifying surface defects of a radiator of an IGBT power module. The beneficial effects are as follows:
(1) According to the method for identifying the surface defects of the radiator of the IGBT power module, the high-resolution industrial camera and the infrared thermal imaging camera are adopted to collect images of the surface of the radiator in real time, and a plurality of images A and a plurality of images B are obtained, so that more comprehensive and detailed surface information is provided; the image processing technology is used for denoising, contrast increasing, graying and filtering the multi-frame image A and the multi-frame image B, so that the images are clearer, and the feature extraction and analysis are facilitated.
(2) According to the method for identifying the surface defects of the radiator of the IGBT power module, the bubble characteristics, the crack characteristics, the corrosion characteristics and the dirt characteristics in the multi-frame image A after image processing are extracted through the characteristic extraction algorithm, the characteristics of the multi-frame image B are extracted, the local temperature abnormal characteristics and the surface thermal crack distribution characteristics are extracted, and therefore the defect condition of the surface of the radiator is comprehensively evaluated.
(3) The IGBT power module radiator surface defect identification method introduces a comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comprehensively evaluating the defect characteristics of bubbles, cracks, corrosion, dirt, local temperature abnormality, surface thermal cracks and the like, comparing the defect characteristics with the set standard defect characteristics to obtain an evaluation result, and accurately judging the state and deviation degree of the surface defects of the radiator.
(4) The invention provides a method for identifying surface defects of an IGBT power module radiator, which aims to solve the problems of low defect identification efficiency, low accuracy, difficult overall assessment of defect states and the like in the prior art. The method is based on a high-resolution industrial camera and an infrared thermal imaging camera, combines an image processing technology and a feature extraction algorithm, and utilizes a comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comprehensively identifying and evaluating defects such as bubbles, cracks, corrosion, dirt, local temperature abnormality, surface thermal cracks and the like on the surface of the radiator.
Drawings
FIG. 1 is a schematic flow chart of a method for identifying surface defects of a radiator of an IGBT power module;
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The IGBT power module radiator is an important component used in power electronic equipment, and is mainly used for radiating heat and reducing temperature so as to ensure the normal operation of the power module. These heat sinks are typically exposed to high temperature and pressure environments and are prone to defects such as bubbles, cracks, corrosion, dirt, etc. during long-term operation, resulting in reduced heat dissipation and even affecting the stability and life of the device.
The existence of surface defects of the radiator of the IGBT power module can seriously affect the working performance and reliability of equipment. Thus, early identification and accurate assessment of the extent of defects is critical to maintaining and continuing use of the power module. The conventional defect recognition method may require a lot of time and labor and has a limited accuracy, so that it is necessary to develop a more efficient and accurate defect recognition method.
Current defect identification methods are mostly based on manual visual inspection or use of some specific sensors to obtain surface information. These methods are not only time-consuming and laborious, but also susceptible to human subjective factors, leading to unstable recognition results. Furthermore, existing methods typically focus on only a single defect feature, making it difficult to fully evaluate the defect state of the heat spreader surface.
Example 1
The invention provides a method for identifying surface defects of a radiator of an IGBT power module, referring to FIG. 1, comprising the following steps,
shooting and acquiring a plurality of frames of images on the surface of the radiator by adopting a high-resolution industrial camera, acquiring a plurality of frames of images A, and acquiring infrared imaging on the surface of the radiator in real time by adopting an infrared thermal imaging camera to acquire a plurality of frames of images B;
performing image processing on the multi-frame image A and the multi-frame image B, wherein the image processing comprises noise removal, contrast increase, graying processing and filtering processing; and adjusting and calculating a focal size proportion value f (x, y), and dynamically extracting image features according to the focal size proportion value f (x, y); through removing noise, the image can be cleaner, interference is reduced, and the method is helpful for more accurately extracting the characteristics of the surface of the radiator; the contrast of the image is adjusted to enhance details and features in the image, so that differences of different parts in the image are more obvious; converting a color image to a grayscale image can simplify the complexity of image processing and help highlight texture and shape features of the image; filtering may smooth the image and remove high frequency noise from the image. The proper filtering algorithm can help to retain useful information in the image and eliminate unnecessary details in the image, so that the stability of defect detection is improved;
establishing an identification model, extracting bubble characteristics, crack characteristics, corrosion characteristics and dirt characteristics in a multi-frame image A after image processing, and analyzing and calculating the characteristics to obtain a bubble distribution coefficient Qp and a first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs and fouling distribution coefficient Wg; extracting features of the multi-frame image B after image processing, extracting local temperature abnormal features and surface thermal crack distribution features, and obtaining an abnormal feature distribution coefficient Rd and a second crack distribution coefficient Lw 2 The method comprises the steps of carrying out a first treatment on the surface of the The first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Matching, if the numbers are the same, the identification effect is correct, and if the deviation exceeds a preset deviation value, the identification is needed again;
by establishing a comprehensive identification model and simultaneously extracting a plurality of characteristics such as bubbles, cracks, corrosion, dirt and the like, the system can comprehensively consider the characteristics of different defect types, and the bubble distribution coefficient Qp and the first crack distribution coefficient Lw are calculated 1 The defect information can be converted into a numerical index by the corrosion distribution coefficient Fs and the scale distribution coefficient Wg. The indexes can quantitatively describe the defect condition of the surface of the radiator, and are convenient for comparison and assessment of the severity of the defect; and extracting local temperature abnormal characteristics and surface thermal crack distribution characteristics from the multi-frame image B after image processing, thereby being beneficial to detecting thermal problems and cracks on the surface of the radiator. These abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 Further information is provided, so that the performance and health of the radiator can be more comprehensively evaluated; the first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Matching is carried out, a preset deviation value is set, and the identification effect can be verified. If the deviation exceeds a preset value, the identification can be triggered in time again, so that the reliability and the accuracy of defect detection are improved;
labeling the corresponding features and classifications of each image according to the existing images and the extracted features, and storing the labeled features and classifications as a total data set SJ;
dividing the marked total data set SJ into a training set and a testing set, training by adopting a cross-validation method, and optimizing model parameters to minimize prediction errors; the generalization capability of the model can be better evaluated through cross verification, and the problem of over fitting is avoided, so that the robustness and the accuracy of the defect identification model are improved.
The bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Correlating the corrosion distribution coefficient Fs and the dirt distribution coefficient Wg to obtain a first defect evolution evaluation coefficient PG 1 The abnormal characteristic distribution coefficient Rd and the second crack distribution coefficient Lw are calculated 2 Associated, obtain the second defect evolution assessment coefficient PG 2 Estimating the first defect evolution estimation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 Fitting and calculating to obtain comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) The formula of (2) is:
wherein m is expressed as a first defect evolution estimation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 W represents a preset error threshold value, and n represents an error threshold value factor; comprehensively considering the influence of a plurality of characteristics on the defects to obtain a more comprehensive defect evaluation result;
the comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comparing the defect characteristic with the standard defect characteristic to obtain a corresponding defect compensation scheme. And a corresponding defect compensation scheme is formulated, so that the defects are repaired in time, and the normal operation of the radiator is ensured.
In this embodiment, by establishing an identification model, extracting image features, analyzing and calculating the features, defects on the surface of the IGBT power module radiator can be effectively identified, and converted into a numerical index, so as to implement comprehensive defect evaluation. This will be beneficial to improving the efficiency and reliability of defect detection, ensuring the stability of the surface quality and performance of the heat sink; the defect identification and evaluation system is used for comprehensively evaluating the influence of different characteristics on defects by constructing a data set, training an identification model by adopting a cross-validation method, providing a comprehensive, accurate and reliable defect evaluation result and providing a beneficial scheme for defect compensation. Such a system helps to improve product quality and safety in industrial production while reducing failure rate and maintenance costs.
Example 2, this example is an explanation made in example 1, specifically, the formula for calculating the focal point size ratio value f (x, y) by a two-dimensional gaussian function is:
wherein x and y are represented as the image abscissa and ordinate, and α is the gaussian kernel; f is the focal size proportion of (x, y), a mask is used for scanning each pixel of the image in specific operation, and the value of a pixel point at the center of the mask is the weighted average gray value of pixels in the neighborhood determined by the mask; a Z multiplied by Z mask is adopted, Z is more than or equal to 4, and the obtained filtering effect provides an image with higher image quality for edge detection. By calculating the focal point size proportion value by using a two-dimensional Gaussian function, various beneficial effects such as image smoothing, noise elimination, edge preservation, image quality improvement and the like can be realized, and the method has important significance for applications such as defect detection, feature extraction and the like. In the industrial defect identification, the method can improve the accuracy and the stability of the radiator surface defect identification.
Because the Gaussian filter can reduce noise and discontinuity in the image, the visual quality of the image can be improved, the image is more attractive and natural, the focal size proportion value f (x, y) is calculated by using a two-dimensional Gaussian function, and the focal size can be dynamically adjusted according to the specific condition of the image so as to adapt to the characteristics of different image areas. This helps to optimise the image processing effect, especially in the case of defects or abnormal features in the image.
Example 3, this example is in practiceThe explanation is made in example 1, specifically, the bubble distribution coefficient Qp, the first crack distribution coefficient Lw 1 The corrosion distribution coefficient Fs and the fouling distribution coefficient Wg are calculated by the following formulas:
wherein: np represents the number of detected bubbles in the image, N represents the total number of areas in the image or the total number of pixels of the heat sink surface; the bubble distribution coefficient Qp represents the duty ratio of bubbles on the surface of the radiator; nw1 represents the number of the first type of cracks detected in the image, and the first crack distribution coefficient Lw 1 Representing the duty cycle of a first type of crack on the surface of the radiator; ns represents the number of corrosion areas detected in the image, and the corrosion distribution coefficient Fs represents the duty ratio of the corrosion areas on the radiator surface; ng represents the number of dirt areas detected in the image, and the dirt distribution coefficient Wg represents the duty cycle of the dirt areas on the radiator surface. The number of defects is converted into a specific numerical index. This helps to more accurately describe the distribution of different types of defects on the surface of the heat sink, and thus better quantitatively assess the defects.
In the present embodiment, the bubble distribution coefficient Qp, the first crack distribution coefficient Lw 1 The corrosion distribution coefficient Fs and the fouling distribution coefficient Wg represent the duty cycle of the different defect types across the surface of the radiator. This helps to analyze the distribution trend of different types of defects, determining whichThe type of defect is more severe and thus is preferred. Meanwhile, the change of the defect proportion can be tracked, and potential problems can be found in time; by calculating the distribution coefficients of different types of defects on the surface of the radiator, the fault early warning function can be realized. Once the defect proportion of a certain type is obviously increased, the defect type can mean that the problem of the defect type is serious, and measures are taken in time for maintenance or replacement; according to the calculated defect distribution coefficient, a defect compensation scheme can be specifically formulated. For example, if the bubble distribution coefficient Qp is higher, it is considered to optimize the welding process or to use a better heat sink design, thereby reducing the formation of bubbles;
by calculating the bubble distribution coefficient Qp, the first crack distribution coefficient Lw 1 The corrosion distribution coefficient Fs and the dirt distribution coefficient Wg can realize quantitative evaluation and analysis of defects, and provide important information beneficial to fault early warning and defect compensation scheme formulation. This will help to improve the performance and reliability of the heat sink and ensure its stable operation in industrial applications.
Embodiment 4, which is an explanation made in embodiment 1, specifically, an image B is set to I (x, y), where (x, y) represents pixel coordinates in the image; in order to extract the local temperature anomaly characteristic, the anomaly characteristic distribution coefficient Rd is obtained by calculation using the following formula:
wherein Σ represents summing all pixels of the image B, μ represents the average temperature of the image B, σ represents the standard deviation of the image B; the abnormal characteristic distribution coefficient Rd represents the degree of deviation of the temperature of each pixel point in the image from the average temperature.
In this embodiment, a local temperature abnormality region on the surface of the heat sink can be quickly identified. These abnormal areas may be caused by faults, poor connections, or other problems, and timely identifying detection and handling of these anomalies helps to prevent potential faults. By using the average temperature μ and standard deviation σ of image B, rd normalizes the temperature values in image B so that images with different scales and ranges can be compared and analyzed. This helps to eliminate the difficulties caused by the different temperature value ranges; rd calculates the degree of temperature deviation for each pixel point, which can help locate the local temperature anomaly region on the surface of the heat sink.
The abnormal characteristic distribution coefficient Rd can be used for realizing the rapid detection and positioning of local temperature abnormality, and extracting the local temperature characteristic of an image, so that beneficial information is provided for the analysis and fault diagnosis of the radiator surface temperature distribution. This will help to ensure the performance and reliability of the heat sink while improving safety and efficiency in the industrial process.
Embodiment 5, which is an explanation of embodiment 1, specifically, the image B is set to be I (x, y), that is, a single-channel gray scale image, and the second crack distribution coefficient Lw is obtained by calculating the following formula for extracting the surface thermal crack distribution characteristics 2 ：
Where, Σ represents summing all pixels of image B,a laplace operator representing the image B; second crack distribution coefficient Lw 2 Representing the quadratic square of the gray value rate of change of each pixel in the image.
In the present embodiment, the second crack distribution coefficient Lw is calculated 2 The hot crack areas on the surface of the heat sink can be quickly identified. Thermal cracking is typically manifested as rapid changes in gray values, through a second crack distribution coefficient Lw 2 The calculation of (a) can accurately locate these areas; thermal cracking is generally manifested as a sharp change in gray value, a second crack distribution coefficient Lw 2 These features can be highlighted to better understand the crack distribution of the heat spreader surface. Due to the second crack distribution coefficient Lw 2 For the change rate of gray valueThe quadratic square can enhance the region with larger gray value variation in the image, so that the hot cracking is more obvious. This helps to improve the contrast and sharpness of the image, so that the presence of thermal cracks is more easily observed; by calculating a second crack distribution coefficient Lw 2 The gray value change rate of each pixel point is obtained, so that the heat crack area on the surface of the radiator can be positioned. Therefore, the fault position can be more accurately determined, the quick repair of the hot cracks is facilitated, and the further expansion of the faults is avoided.
Example 6, which is an explanation of example 1, specifically, the absolute error matching method is used to calculate the first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Absolute error, if the error is smaller than the preset deviation value, the identification effect is considered to be correct; the method is simple and visual, and is suitable for scenes sensitive to the absolute value of the error.
Calculating a first crack distribution coefficient Lw by adopting a relative error matching method 1 And a second crack distribution coefficient Lw 2 Relative error, i.e. the difference in values divided by the first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 If the relative error is smaller than the preset deviation value, the identification effect is considered to be correct; this approach is applicable to scenes that are relatively sensitive to numerical size.
Calculating a first crack distribution coefficient Lw by adopting a root mean square error matching method 1 And a second crack distribution coefficient Lw 2 The root mean square error, namely the average value of the sum of squares of the differences on each pixel point, is considered to be correct if the root mean square error is smaller than a preset deviation value. The method comprehensively considers the errors on all the pixel points and is suitable for scenes sensitive to the whole errors.
In this embodiment, a plurality of error matching methods are used to evaluate the first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Errors among the defects are beneficial to improving the accuracy and the robustness of defect identification. By setting a preset deviation value, whether the identification effect is correct or not can be rapidly verified, and parameters or an optimization algorithm can be timely adjusted so as to further improve the reliability of defect identification. This will be of benefit to guaranteeing accuracy and stability of radiator surface defect discernment, ensures product quality and security.
Embodiment 7, which is the explanation of embodiment 1, specifically, the first defect-evolution-assessment coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 The method is obtained by calculation through the following association formula:
wherein G, H, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw, respectively 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 The weight coefficient value of (2) is adjusted and set by a user; beta is expressed as a modified natural number.
In the present embodiment, the first defect evolution evaluation coefficient PG is calculated by recognition 1 And a second defect evolution evaluation coefficient PG 2 And a comprehensive evaluation result of defect evolution can be obtained. The method is helpful for analyzing the evolution trend of the defects, finding whether the trend of gradual deterioration exists or not, and timely taking measures for maintenance and repair. Calculating a first defect evolution assessment coefficient PG by using a correlation formula 1 And a second defect evolution evaluation coefficient PG 2 Comprehensive defect evaluation, weight adjustment and evolution trend analysis can be realized, and the method is helpful for comprehensively and flexibly evaluating the evolution condition of the surface defects of the radiator. The method is beneficial to improving the accuracy of defect identification and evaluation, guaranteeing the performance and reliability of the radiator and ensuring the stable operation of the product in industrial application.
Example 8, which is an explanation of example 1, specifically, a set of standard defect features, which represent features that the radiator should have in an ideal state, are set according to expertise and experimental data; these features may include bubble free, crack free, corrosion free, fouling free, uniform temperature distribution, etc. The ideal state is clarified, and a reference standard is provided for defect evaluation of the surface of the radiator. The standard defect feature provides a reference standard that helps to evaluate whether the defective condition of the heat sink surface meets expectations. According to the comparison of the actual evaluation result and the standard defect characteristics, the health condition of the radiator can be judged, and guidance is provided for subsequent repair and maintenance.
The comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comparing the defect state with the standard defect characteristics to obtain an evaluation result, wherein the evaluation result is used for evaluating the defect state and the deviation degree of the surface of the radiator. By comparing with the ideal state, whether the radiator has defects such as bubbles, cracks, corrosion, dirt and the like can be determined, and the severity of the defects can be evaluated.
In this embodiment, a set of standard defect features is set as a reference standard for evaluating the defect state and deviation of the surface of the heat sink. By and synthesis of the evaluation coefficient PG (PG 1 ，PG 2 ) The comparison can be carried out, whether the radiator has defects or not can be accurately judged, and the severity of the radiator is evaluated, so that corresponding maintenance and repair measures are adopted, the performance and reliability of the radiator are ensured, and the stable operation of the radiator in industrial application is ensured.
According to the evaluation result, a targeted defect compensation scheme is formulated; the defect remedy includes repair or cleaning measures for bubbles, cracks, corrosion and dirt defects, and also includes improvements suggestions for radiator materials, structures or designs.
Embodiment 9, which is an explanation made in embodiment 7, specifically, standard defect features are set, and the standard defect features include standard Qp:0.2 (20%), standard Lw 1 :0.15 (15%), standard Fs:0.1 (10%), standard Wg:0.1 (10%), standard Rd:0.12 (12%), standard Lw 2 :0.08 (8%); G. h, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and secondCrack distribution coefficient Lw 2 G=0.4, h=0.4, c=0.3, d=0.2, e=0.1, f=0.5;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Qp: identifying a poor bubble defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 1 : identifying a first crack defect condition;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Fs: identifying a poor corrosion defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Wg: identifying a poor fouling defect condition;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Rd: identifying abnormal characteristics and poor defect state;
if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 2 The second crack defect state is identified as poor.
In a specific example, in an image,
G. h, C, D, E, F are weighting coefficients for adjusting the importance of different features in the overall evaluation.
Standard defect features may be set based on expert knowledge or experimental data, for example, assuming that our standard defect features are as follows:
standard Qp:0.2 (20%) Standard Lw 1 :0.15 (15%) standard Fs:0.1 (10%) standard Wg:0.1 (10%) standard Rd:0.12 (12%) Standard Lw 2 ：0.08(8％)
According to actual conditions and expert experience, we choose the weight coefficient as follows:
G＝0.4、H＝0.4、C＝0.3、D＝0.2、E＝0.1、F＝0.5；
now, we perform feature extraction on the multi-frame image B after image processing to calculate PG 1 And PG 2 And compared to standard defect features:
assume that the calculated PG 1 0.25 PG 2 0.1.
Then weComparative calculated PG 1 And PG 2 And standard defect characteristics, as follows:
PG 1 compared with standard Qp: a defect state is considered to be better with a difference of 0.25 from 0.2. PG 1 And standard Lw 1 Compared with the prior art: a large difference between 0.25 and 0.15 is considered a poor defect condition. PG 1 Compared with standard Fs: a large difference between 0.25 and 0.1 is considered a poor defect condition. PG 1 Compared with standard Wg: a large difference between 0.25 and 0.1 is considered a poor defect condition. PG 2 Compared with standard Rd: a defect state is considered to be good, with 0.1 being not much different from 0.12. PG 2 And standard Lw 2 Compared with the prior art: a defect state is considered to be better with a small difference between 0.1 and 0.08.
The following conclusion can be drawn by combining the comparison and evaluation results:
PG 1 and PG 2 In (2), PG 1 There is a large deviation, mainly due to Lw 1 The values of Fs and Wg are larger, while PG 2 Closer to the standard defect features, it means that local temperature anomalies and surface thermal cracks are relatively less.
And (3) according to the evaluation result, making a defect compensation scheme:
for Lw 1 The larger Fs and Wg may allow for targeted repair and cleaning measures, such as removal of corrosions and dirt, repair of cracks, etc. For the case of local temperature anomalies and less surface thermal cracking, improvements in the heat sink material, structure or design may be considered to improve the overall performance and reliability of the heat sink. The specific defect compensation scheme needs to be further formulated according to actual conditions and professional knowledge so as to ensure the normal operation and the optimized performance of the radiator.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (9)
- A method for identifying surface defects of an IGBT power module radiator is characterized by comprising the following steps: comprises the steps of,shooting and acquiring a plurality of frames of images on the surface of the radiator by adopting a high-resolution industrial camera, acquiring a plurality of frames of images A, and acquiring infrared imaging on the surface of the radiator in real time by adopting an infrared thermal imaging camera to acquire a plurality of frames of images B;performing image processing on the multi-frame image A and the multi-frame image B, wherein the image processing comprises noise removal, contrast increase, graying processing and filtering processing; and adjusting and calculating a focal size proportion value f (x, y), and dynamically extracting image features according to the focal size proportion value f (x, y);establishing an identification model, extracting bubble characteristics, crack characteristics, corrosion characteristics and dirt characteristics in a multi-frame image A after image processing, and analyzing and calculating the characteristics to obtain a bubble distribution coefficient Qp and a first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs and fouling distribution coefficient Wg; extracting features of the multi-frame image B after image processing, extracting local temperature abnormal features and surface thermal crack distribution features, and obtaining an abnormal feature distribution coefficient Rd and a second crack distribution coefficient Lw 2 The method comprises the steps of carrying out a first treatment on the surface of the The first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 Matching, if the numbers are the same, the identification effect is correct, and if the deviation exceeds a preset deviation value, the identification is needed again;labeling the corresponding features and classifications of each image according to the existing images and the extracted features, and storing the labeled features and classifications as a total data set SJ;dividing the marked total data set SJ into a training set and a testing set, training by adopting a cross-validation method, and optimizing model parameters to minimize prediction errors;the bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Correlating the corrosion distribution coefficient Fs and the dirt distribution coefficient Wg to obtain a first defect evolution evaluation coefficient PG 1 The abnormal characteristic distribution coefficient Rd and the second crack distribution coefficient Lw are calculated 2 Associated, obtain the second defect evolution assessment coefficient PG 2 Estimating the first defect evolution estimation coefficient PG 1 And a second defectNotch evolution evaluation coefficient PG 2 Fitting and calculating to obtain comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) The formula of (2) is:wherein m is expressed as a first defect evolution estimation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 W represents a preset error threshold value, and n represents an error threshold value factor;the comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) And comparing the defect characteristic with the standard defect characteristic to obtain a corresponding defect compensation scheme.
- 2. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: the formula for calculating the focal point size proportion value f (x, y) through a two-dimensional Gaussian function is as follows:wherein x and y are represented as the image abscissa and ordinate, and α is the gaussian kernel; f is the focal size proportion of (x, y), a mask is used for scanning each pixel of the image in specific operation, and the value of a pixel point at the center of the mask is the weighted average gray value of pixels in the neighborhood determined by the mask; a Z multiplied by Z mask is adopted, Z is more than or equal to 4, and the obtained filtering effect provides an image with higher image quality for edge detection.
- 3. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: the bubble distribution coefficient Qp, the first crack distribution coefficient Lw 1 The corrosion distribution coefficient Fs and the fouling distribution coefficient Wg are calculated by the following formulas:wherein: np represents the number of detected bubbles in the image, N represents the total number of areas in the image or the total number of pixels of the heat sink surface; the bubble distribution coefficient Qp represents the duty ratio of bubbles on the surface of the radiator; nw1 represents the number of the first type of cracks detected in the image, and the first crack distribution coefficient Lw 1 Representing the duty cycle of a first type of crack on the surface of the radiator; ns represents the number of corrosion areas detected in the image, and the corrosion distribution coefficient Fs represents the duty ratio of the corrosion areas on the radiator surface; ng represents the number of dirt areas detected in the image, and the dirt distribution coefficient Wg represents the duty cycle of the dirt areas on the radiator surface.
- 4. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: setting the image B as I (x, y), wherein (x, y) represents pixel coordinates in the image; in order to extract the local temperature anomaly characteristic, the anomaly characteristic distribution coefficient Rd is obtained by calculation using the following formula:wherein Σ represents summing all pixels of the image B, μ represents the average temperature of the image B, σ represents the standard deviation of the image B; the abnormal characteristic distribution coefficient Rd represents the degree of deviation of the temperature of each pixel point in the image from the average temperature.
- 5. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: setting the image B as I (x, y), namely a single-channel gray level image, and obtaining the second crack distribution coefficient Lw by adopting the following formula to calculate in order to extract the surface thermal crack distribution characteristics 2 ：Where, Σ represents summing all pixels of image B,a laplace operator representing the image B; second crack distribution coefficient Lw 2 Representing the quadratic square of the gray value rate of change of each pixel in the image.
- 6. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: calculating a first crack distribution coefficient Lw by adopting an absolute error matching method 1 And a second crack distribution coefficient Lw 2 Absolute error, if the error is smaller than the preset deviation value, the identification effect is considered to be correct;or, calculating the first crack distribution coefficient Lw by adopting a relative error matching method 1 And a second crack distribution coefficient Lw 2 Relative error, i.e. the difference in values divided by the first crack distribution coefficient Lw 1 And a second crack distribution coefficient Lw 2 If the relative error is smaller than the preset deviation value, the identification effect is considered to be correct;or, calculating the first crack distribution coefficient Lw by adopting a root mean square error matching method 1 And a second crack distribution coefficient Lw 2 The root mean square error between the two, i.e. the average of the sum of squares of the differences at each pixel point, if the root mean square error is less than a preset deviation value, the recognition effect is consideredCorrect.
- 7. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: the first defect evolution evaluation coefficient PG 1 And a second defect evolution evaluation coefficient PG 2 The method is obtained by calculation through the following association formula:wherein G, H, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw, respectively 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 The weight coefficient value of (2) is adjusted and set by a user; beta is expressed as a modified natural number.
- 8. The method for identifying the surface defects of the radiator of the IGBT power module according to claim 1, wherein the method comprises the following steps: setting a group of standard defect characteristics according to professional knowledge and experimental data, wherein the characteristics represent the characteristics which the radiator should have in an ideal state;the comprehensive evaluation coefficient PG (PG) 1 ，PG 2 ) Comparing the defect state with standard defect characteristics to obtain an evaluation result, wherein the evaluation result is used for evaluating the defect state and deviation degree of the surface of the radiator;according to the evaluation result, a targeted defect compensation scheme is formulated; the defect remedy includes repair or cleaning measures for bubbles, cracks, corrosion and dirt defects, and also includes improvements suggestions for radiator materials, structures or designs.
- 9. The method for identifying surface defects of an IGBT power module radiator as claimed in claim 7, which is characterized in thatThe method is characterized in that: setting standard defect characteristics, wherein the standard defect characteristics comprise standard Qp:0.2 (20%), standard Lw 1 :0.15 (15%), standard Fs:0.1 (10%), standard Wg:0.1 (10%), standard Rd:0.12 (12%), standard Lw 2 :0.08 (8%); G. h, C, D, E, F is the bubble distribution coefficient Qp and the first crack distribution coefficient Lw 1 Corrosion distribution coefficient Fs, fouling distribution coefficient Wg, abnormal characteristic distribution coefficient Rd and second crack distribution coefficient Lw 2 G=0.4, h=0.4, c=0.3, d=0.2, e=0.1, f=0.5;if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Qp: identifying a poor bubble defect state;if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 1 : identifying a first crack defect condition;if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Fs: identifying a poor corrosion defect state;if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Wg: identifying a poor fouling defect condition;if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Rd: identifying abnormal characteristics and poor defect state; if the integrated evaluation coefficient PG (PG) 1 ，PG 2 ) > standard Lw 2 The second crack defect state is identified as poor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311097425.4A CN116843680B (en) | 2023-08-29 | 2023-08-29 | IGBT power module radiator surface defect identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311097425.4A CN116843680B (en) | 2023-08-29 | 2023-08-29 | IGBT power module radiator surface defect identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116843680A CN116843680A (en) | 2023-10-03 |
| CN116843680B true CN116843680B (en) | 2023-12-15 |
Family
ID=88163758
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311097425.4A Active CN116843680B (en) | 2023-08-29 | 2023-08-29 | IGBT power module radiator surface defect identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116843680B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117274257B (en) * | 2023-11-21 | 2024-01-23 | 山西集智数据服务有限公司 | Automatic classification system for book looks based on machine vision |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106290394A (en) * | 2016-09-30 | 2017-01-04 | 华南理工大学 | A kind of cpu heat aluminium extruded forming defect detecting system and detection method |
| CN115409824A (en) * | 2022-09-06 | 2022-11-29 | 长沙理工大学 | Silicon wafer surface defect detection method based on deep convolutional neural network |
| CN115965599A (en) * | 2022-12-28 | 2023-04-14 | 黄山谷捷股份有限公司 | IGBT power module radiator surface defect identification method |
-
2023
- 2023-08-29 CN CN202311097425.4A patent/CN116843680B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106290394A (en) * | 2016-09-30 | 2017-01-04 | 华南理工大学 | A kind of cpu heat aluminium extruded forming defect detecting system and detection method |
| CN115409824A (en) * | 2022-09-06 | 2022-11-29 | 长沙理工大学 | Silicon wafer surface defect detection method based on deep convolutional neural network |
| CN115965599A (en) * | 2022-12-28 | 2023-04-14 | 黄山谷捷股份有限公司 | IGBT power module radiator surface defect identification method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116843680A (en) | 2023-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116843680B (en) | IGBT power module radiator surface defect identification method | |
| Fujita et al. | A robust automatic crack detection method from noisy concrete surfaces | |
| US9582872B2 (en) | Optical film defect detection method and system thereof | |
| Spataru et al. | Automatic detection and evaluation of solar cell micro-cracks in electroluminescence images using matched filters | |
| WO2021168733A1 (en) | Defect detection method and apparatus for defect image, and computer-readable storage medium | |
| CN108846397B (en) | Automatic detection method for cable semi-conducting layer based on image processing | |
| CN111242123A (en) | Power equipment fault diagnosis method based on infrared image | |
| WO2020173177A1 (en) | Object color difference defect detection method, device, computer device, and storage medium | |
| CN106197612A (en) | A kind of transparent bottled liquid-level detecting method based on machine vision | |
| CN116777911A (en) | Double-substrate radiator surface defect detection system based on image recognition | |
| CN112017109B (en) | Online ferrographic video image bubble elimination method | |
| CN113850749A (en) | Method for training defect detector | |
| CN113240667A (en) | Automobile mold plane crack detection method based on image processing | |
| Yeh et al. | A wavelet-based approach in detecting visual defects on semiconductor wafer dies | |
| CN117173187A (en) | Intelligent valve fault detection system | |
| US8004564B1 (en) | Automated systems and methods for testing infrared cameras | |
| CN114965316A (en) | Hyperspectral multi-parameter based water quality monitoring method and computer program product | |
| CN111638218A (en) | Method for detecting surface defects of coating | |
| Srivatava et al. | Non-linear quality evaluation index for mammograms | |
| CN117333467B (en) | Image processing-based glass bottle body flaw identification and detection method and system | |
| CN116778520A (en) | Mass license data quality inspection method | |
| CN107705284B (en) | Bayesian small sample learning-based surface defect detection method | |
| JP4948238B2 (en) | Distribution analysis method and apparatus, abnormal equipment estimation method and apparatus, program for causing computer to execute distribution analysis method or abnormal equipment estimation method, and computer-readable recording medium recording the program | |
| CN117078679B (en) | Automatic assembly line production detection method for cooling fan based on machine vision | |
| Zhu et al. | Inspection and classification of wheat quality using image processing |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |