Industrial defect detection method and system based on artificial intelligence

Abstract

The invention is suitable for the technical field of industrial defect detection, and provides an artificial intelligence-based industrial defect detection method and system. According to the invention, historical detection data and a standard product image are obtained, a detection stable area is determined, and a standard area image is intercepted; carrying out detection adjustment based on artificial intelligence, and obtaining a detection shooting image; intercepting a corresponding area image, carrying out similarity comparison calculation, and judging whether a similarity detection condition exists or not; and when the similar detection conditions exist, defect detection is performed on the detection shooting image, and detection marking is performed. A stable detection area can be determined, a standard area image can be intercepted, in the product detection process, a corresponding area image is intercepted, similarity comparison calculation is carried out, and defect detection and marking are carried out on a detection shot image when similar detection conditions exist, so that verification before defect detection is realized, a wrong detection result is avoided, and the detection efficiency is improved. Accuracy and reliability of industrial defect detection are effectively improved, and normal production of products is ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial defect detection, and in particular, relates to an industrial defect detection method and system based on artificial intelligence. Background Art

[0002] Industrial defect detection is the use of a series of technical means to conduct comprehensive and full-process inspection of products during the manufacturing process to identify possible defects, including cracks, scratches, stains, dimensional deviations, pores, inclusions, circuit problems, and other forms.

[0003] Image detection of industrial defects is one of the common technical means of industrial defect detection.

[0004] In the prior art, image detection of industrial defects mainly relies on direct image comparison. However, in special circumstances such as changes in light brightness, light obstruction, changes in product placement, etc., it may cause originally defective products to be detected as defective products, resulting in erroneous detection results. As a result, the accuracy and reliability of industrial defect detection cannot be guaranteed, which in turn affects the normal production of products. Summary of the invention

[0005] The purpose of the embodiments of the present invention is to provide an industrial defect detection method and system based on artificial intelligence, aiming to solve the technical problems existing in the prior art mentioned in the background technology.

[0006] The embodiment of the present invention is implemented as follows:

[0007] An industrial defect detection method based on artificial intelligence, the method specifically comprises the following steps:

[0008] Acquire historical inspection data and standard product images of target industrial products, perform inspection stability analysis, determine the inspection stability area from the standard product images, and intercept the standard area image;

[0009] Determine the current inspection product, perform inspection adjustment on the current inspection product based on artificial intelligence, perform inspection shooting, and obtain the inspection shooting image;

[0010] From the detection shot image, intercept the corresponding area image of the detection stable area, perform similarity comparison calculation on the corresponding area image and the standard area image, and determine whether similar detection conditions are met;

[0011] When similar inspection conditions are met, defect inspection is performed on the inspection shot image based on the standard product image, and inspection marking is performed on the current inspection product.

[0012] As a further limitation of the technical solution of the embodiment of the present invention, the acquisition of historical detection data and standard product images of the target industrial product, the detection stability analysis, the determination of the detection stability area from the standard product image, and the interception of the standard area image specifically include the following steps:

[0013] Receive testing planning information and determine target industrial products;

[0014] Acquire historical inspection data and standard product images of the target industrial product;

[0015] Based on the standard product image, performing a detection stability analysis on the historical detection data, and determining a detection stability area from the standard product image;

[0016] In the standard product image, a standard area image of the detection stable area is cut out.

[0017] As a further limitation of the technical solution of the embodiment of the present invention, the determining of the current inspection product, the inspection adjustment of the current inspection product based on artificial intelligence, and the inspection shooting, and the acquisition of the inspection shooting image specifically include the following steps:

[0018] Obtaining detection adjustment parameters of the target industrial product;

[0019] Determine the current testing product in real time;

[0020] Based on artificial intelligence, according to the detection adjustment parameters, the detection adjustment is performed on the current detection product;

[0021] Perform detection shooting and obtain detection shooting images.

[0022] As a further limitation of the technical solution of the embodiment of the present invention, the method of intercepting the corresponding area image of the detection stable area from the detection captured image, performing similarity comparison calculation on the corresponding area image and the standard area image, and determining whether similar detection conditions are met specifically includes the following steps:

[0023] From the detected captured image, intercepting a region image corresponding to the detected stable region;

[0024] Perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value;

[0025] Compare the similarity comparison value with a preset standard threshold value and record the comparison result;

[0026] According to the comparison result, it is determined whether similar detection conditions exist.

[0027] As a further limitation of the technical solution of the embodiment of the present invention, the performing similarity comparison calculation on the corresponding area image and the standard area image to obtain the similarity comparison value specifically includes the following steps:

[0028] Gray-scale the corresponding region image to generate a corresponding gray-scale image;

[0029] Gray-scale the standard region image to generate a standard gray-scale image;

[0030] A similarity comparison calculation is performed on the corresponding grayscale image and the standard grayscale image to obtain a similarity comparison value.

[0031] As a further limitation of the technical solution of the embodiment of the present invention, the calculation formula of the similarity comparison value is:

[0032]

[0033] Among them, N is the similarity comparison value, d represents the corresponding grayscale image, s represents the standard grayscale image, and u d is the average brightness corresponding to the grayscale image, u s is the average brightness corresponding to the standard grayscale image, σ d is the standard deviation of the corresponding grayscale image, σ s is the standard deviation corresponding to the standard grayscale image, σ ds is the covariance between the corresponding grayscale image and the standard grayscale image, L is the value range of the pixel value, k1 is the preset first adjustment coefficient, and k2 is the preset second adjustment coefficient.

[0034] As a further limitation of the technical solution of the embodiment of the present invention, when there are similar detection conditions, based on the standard product image, defect detection is performed on the detection image, and detection marking is performed on the current detection product, specifically including the following steps:

[0035] When similar inspection conditions are met, defect inspection is performed on the inspection image based on the standard product image to obtain a defect inspection result;

[0036] Identify the detected image to obtain the current product identity of the currently detected product;

[0037] According to the current product identity and the defect detection result, the current detection product is detected and marked.

[0038] An industrial defect detection system based on artificial intelligence, the system includes a detection stability analysis module, a detection adjustment shooting module, a similarity comparison judgment module and a defect detection marking module, wherein:

[0039] A detection stability analysis module is used to obtain historical detection data and standard product images of target industrial products, perform detection stability analysis, determine the detection stability area from the standard product image, and intercept the standard area image;

[0040] The detection adjustment and shooting module is used to determine the current detection product, perform detection adjustment on the current detection product based on artificial intelligence, and perform detection shooting to obtain the detection shooting image;

[0041] A similarity comparison judgment module is used to intercept the corresponding area image of the detection stable area from the detection shooting image, perform similarity comparison calculation on the corresponding area image and the standard area image, and judge whether similar detection conditions are met;

[0042] The defect detection and marking module is used to perform defect detection on the detection shot image based on the standard product image when similar detection conditions exist, and to perform detection marking on the current detection product.

[0043] As a further limitation of the technical solution of the embodiment of the present invention, the detection stability analysis module specifically includes:

[0044] A product determination unit, used to receive the inspection planning information and determine the target industrial product;

[0045] A data acquisition unit, used to acquire historical inspection data and standard product images of the target industrial product;

[0046] a stability analysis unit, configured to perform a detection stability analysis on the historical detection data based on the standard product image, and determine a detection stability area from the standard product image;

[0047] The standard interception unit is used to intercept the standard area image of the detection stability area in the standard product image.

[0048] As a further limitation of the technical solution of the embodiment of the present invention, the similarity comparison judgment module specifically includes:

[0049] A corresponding interception unit, used for intercepting a corresponding area image of the detection stable area from the detection shot image;

[0050] A similarity comparison calculation unit is used to perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value;

[0051] A comparison recording unit, used for comparing the similarity comparison value with a preset standard threshold value and recording the comparison result;

[0052] The condition judgment unit is used to judge whether there are similar detection conditions according to the comparison result.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] The embodiment of the present invention obtains historical detection data and standard product images, determines the detection stable area, and intercepts the standard area image; based on artificial intelligence, performs detection adjustment and obtains the detection shot image; intercepts the corresponding area image, performs similarity comparison calculation, and determines whether there are similar detection conditions; when there are similar detection conditions, performs defect detection on the detection shot image and performs detection marking. It is possible to determine the detection stable area, intercept the standard area image, intercept the corresponding area image during the product detection process, perform similarity comparison calculation, and perform defect detection and marking on the detection shot image when there are similar detection conditions, thereby achieving verification before defect detection, avoiding the generation of erroneous detection results, effectively improving the accuracy and reliability of industrial defect detection, and ensuring the normal production of products. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A flow chart of an industrial defect detection method based on artificial intelligence provided by an embodiment of the present invention is shown;

[0056] Figure 2 A flow chart showing the stability analysis of the detection in the method provided by the embodiment of the present invention is shown;

[0057] Figure 3 A flowchart of detection, adjustment and shooting in the method provided by an embodiment of the present invention is shown;

[0058] Figure 4 A flowchart of determining whether similar detection conditions exist in the method provided by an embodiment of the present invention is shown;

[0059] Figure 5 A flowchart of similarity comparison calculation in the method provided by an embodiment of the present invention is shown;

[0060] Figure 6 A flowchart of defect detection and marking in a method provided by an embodiment of the present invention is shown;

[0061] Figure 7 An application architecture diagram of an artificial intelligence-based industrial defect detection system provided by an embodiment of the present invention is shown;

[0062] Figure 8 It shows a structural block diagram of a detection stability analysis module in a system provided by an embodiment of the present invention;

[0063] Fig. 9 The structure block diagram of the similarity comparison judgment module in the system provided by the embodiment of the present invention is shown. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] It is understandable that in the prior art, image detection of industrial defects mainly relies on direct image comparison. However, in special circumstances such as changes in light brightness, light obstruction, changes in product placement, etc., it may cause originally defective products to be detected as defective products, resulting in erroneous detection results, resulting in the accuracy and reliability of industrial defect detection not being guaranteed, which in turn affects the normal production of products.

[0066] To solve the above problems, an industrial defect detection method and system based on artificial intelligence disclosed in an embodiment of the present invention performs detection stability analysis by acquiring historical detection data and standard product images of target industrial products, determining the detection stability area from the standard product image, and intercepting the standard area image; determining the current detection product, based on artificial intelligence, performing detection adjustment on the current detection product, and performing detection shooting to obtain the detection shooting image; intercepting the corresponding area image of the detection stability area from the detection shooting image, performing similarity comparison calculation on the corresponding area image and the standard area image, and judging whether there are similar detection conditions; when there are similar detection conditions, based on the standard product image, performing defect detection on the detection shooting image, and performing detection marking on the current detection product. It is possible to determine the detection stability area, intercept the standard area image, intercept the corresponding area image during the product detection process, perform similarity comparison calculation, and perform defect detection and marking on the detection shooting image when there are similar detection conditions, thereby realizing verification before defect detection, avoiding the generation of erroneous detection results, effectively improving the accuracy and reliability of industrial defect detection, and ensuring the normal production of products.

[0067] Specifically, Figure 1 A flow chart of an artificial intelligence-based industrial defect detection method provided by an embodiment of the present invention is shown.

[0068] In a preferred embodiment of the present invention, an industrial defect detection method based on artificial intelligence comprises the following steps:

[0069] Step S101: Obtain historical inspection data and standard product images of a target industrial product, perform inspection stability analysis, determine an inspection stability region from the standard product image, and capture a standard region image.

[0070] In an embodiment of the present invention, detection planning information is received, and target industrial products that need to be inspected for industrial defects are determined. By obtaining the internal category code of the target industrial product, data matching is performed in a preset internal database according to the internal category code, and historical detection data and standard product images of the target industrial product are obtained. Based on the standard product image, the historical detection data is analyzed, and areas with historical detection anomalies are marked and counted on the standard product image. From the standard product image, an area that has never had a historical detection anomaly or the number of historical detection anomalies is less than a preset anomaly threshold is selected and marked as a detection stable area. Then, in the standard product image, a standard area image of the detection stable area is captured.

[0071] It is understandable that the target industrial product is a product category.

[0072] Furthermore, Figure 2 A flow chart of performing detection stability analysis in the method provided in an embodiment of the present invention is shown.

[0073] Specifically, in another preferred embodiment provided by the present invention, the step of acquiring historical detection data and a standard product image of a target industrial product, performing detection stability analysis, determining a detection stability area from the standard product image, and intercepting a standard area image specifically comprises the following steps:

[0074] Step S1011: Receive inspection planning information and determine target industrial products.

[0075] Step S1012: Acquire historical inspection data and standard product images of the target industrial product.

[0076] Step S1013: Based on the standard product image, perform a detection stability analysis on the historical detection data, and determine a detection stability area from the standard product image.

[0077] Step S1014: intercepting the standard area image of the detection stability area in the standard product image.

[0078] Furthermore, the industrial defect detection method based on artificial intelligence also includes the following steps:

[0079] Step S102: determine the current inspection product, perform inspection adjustment on the current inspection product based on artificial intelligence, perform inspection shooting, and obtain the inspection shooting image.

[0080] In an embodiment of the present invention, detection adjustment parameters of a target industrial product are obtained, and during the process of industrial defect detection of the target industrial product, the current detection product is determined in real time, and based on artificial intelligence and in accordance with the detection adjustment parameters, the current detection product is clamped, spatially positioned, and / or rotated in the direction thereof, and the current detection product is placed in a preset detection seat, and then the current detection product is inspected and photographed to obtain an inspection photographed image.

[0081] It can be understood that the current inspection product is a specific product entity that performs industrial defect detection in the product category of the target industrial product.

[0082] Furthermore, Figure 3 A flow chart of detection, adjustment and shooting in the method provided by an embodiment of the present invention is shown.

[0083] Specifically, in another preferred embodiment of the present invention, the determining of the currently inspected product, adjusting the currently inspected product based on artificial intelligence, and performing inspection shooting, and obtaining the inspection shooting image specifically include the following steps:

[0084] Step S1021: Acquire detection adjustment parameters of the target industrial product.

[0085] Step S1022: determine the current detected product in real time.

[0086] Step S1023: Based on artificial intelligence, the current detection product is adjusted according to the detection adjustment parameters.

[0087] Step S1024: perform detection shooting to obtain the detection shooting image.

[0088] Furthermore, the industrial defect detection method based on artificial intelligence also includes the following steps:

[0089] Step S103: intercepting the corresponding region image of the detection stable region from the detection shot image, performing similarity comparison calculation on the corresponding region image and the standard region image to determine whether similar detection conditions are met.

[0090] In the embodiment of the present invention, in the detection shot image, the corresponding area image of the detection stable area is intercepted, and the corresponding area image and the standard area image are both grayed to generate a corresponding gray image and a standard gray image. By performing pixel value analysis on the corresponding gray image and the standard gray image, similarity comparison calculation is performed to obtain a similarity comparison value, and then the similarity comparison value is compared with a preset standard threshold, and the comparison result is recorded. Then, according to the comparison result, when the similarity comparison value is greater than the standard threshold, it is determined that the similarity detection condition exists; and when the similarity comparison value is not greater than the standard threshold, it is determined that the similarity detection condition does not exist. Specifically, the calculation formula of the similarity comparison value is:

[0091]

[0092] Among them, N is the similarity comparison value, d represents the corresponding grayscale image, s represents the standard grayscale image, and u d is the average brightness corresponding to the grayscale image, u s is the average brightness corresponding to the standard grayscale image, σ d is the standard deviation of the corresponding grayscale image, σ s is the standard deviation corresponding to the standard grayscale image, σ ds is the covariance between the corresponding grayscale image and the standard grayscale image, L is the value range of the pixel value, k1 is the preset first adjustment coefficient, and k2 is the preset second adjustment coefficient.

[0093] Furthermore, Figure 4 A flow chart of determining whether similar detection conditions exist in the method provided by an embodiment of the present invention is shown.

[0094] Specifically, in another preferred embodiment of the present invention, the corresponding area image of the detection stable area is intercepted from the detection captured image, and the corresponding area image and the standard area image are subjected to similarity comparison calculation to determine whether similar detection conditions are met. Specifically, the following steps are included:

[0095] Step S1031: intercepting an area image corresponding to the detection stable area from the detection captured image.

[0096] Step S1032: perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value.

[0097] Furthermore, Figure 5 A flow chart of similarity comparison calculation in the method provided by an embodiment of the present invention is shown.

[0098] Specifically, in another preferred embodiment of the present invention, the similarity comparison calculation of the corresponding area image and the standard area image to obtain the similarity comparison value specifically includes the following steps:

[0099] Step S10321, grayscale the corresponding area image to generate a corresponding grayscale image.

[0100] Step S10322: grayscale the standard area image to generate a standard grayscale image.

[0101] Step S10323: perform similarity comparison calculation on the corresponding grayscale image and the standard grayscale image to obtain a similarity comparison value.

[0102] Furthermore, the step of intercepting the corresponding region image of the detection stable region from the detection captured image, performing similarity comparison calculation on the corresponding region image and the standard region image, and determining whether similar detection conditions are met also includes the following steps:

[0103] Step S1033: compare the similarity comparison value with a preset standard threshold, and record the comparison result.

[0104] Step S1034: determine whether similar detection conditions exist according to the comparison result.

[0105] Furthermore, the industrial defect detection method based on artificial intelligence also includes the following steps:

[0106] Step S104: When similar detection conditions are met, defect detection is performed on the detection shot image based on the standard product image, and detection marking is performed on the current detection product.

[0107] In an embodiment of the present invention, under similar inspection conditions, defect detection is performed on the inspection image based on the standard product image to obtain the defect detection result, and the inspection image is identified to obtain the current product identity of the current inspection product, and then the current inspection product is marked according to the current product identity and the defect detection result, thereby realizing background marking of industrial defect detection.

[0108] It can be understood that the current product identity is the internal identity code of the current tested product.

[0109] Furthermore, Figure 6 A flow chart of defect detection and marking in the method provided by an embodiment of the present invention is shown.

[0110] Specifically, in another preferred embodiment provided by the present invention, when similar detection conditions are present, defect detection is performed on the detection captured image based on the standard product image, and detection marking is performed on the current detection product, specifically including the following steps:

[0111] Step S1041: When similar detection conditions are met, defect detection is performed on the detection shot image based on the standard product image to obtain a defect detection result.

[0112] Step S1042: Identify the detection image to obtain the current product identity of the current detection product.

[0113] Step S1043: perform detection marking on the current detected product according to the current product identity and the defect detection result.

[0114] Furthermore, Figure 7 An application architecture diagram of an artificial intelligence-based industrial defect detection system provided by an embodiment of the present invention is shown.

[0115] Specifically, in another preferred embodiment provided by the present invention, an industrial defect detection system based on artificial intelligence includes:

[0116] The detection stability analysis module 101 is used to obtain historical detection data and standard product images of target industrial products, perform detection stability analysis, determine the detection stability area from the standard product image, and capture the standard area image.

[0117] In an embodiment of the present invention, the detection stability analysis module 101 receives detection planning information, determines the target industrial products that need to be inspected for industrial defects, obtains the internal category code of the target industrial products, performs data matching in a preset internal database according to the internal category code, obtains the historical detection data and the standard product image of the target industrial products, analyzes the historical detection data based on the standard product image, marks and counts the areas with historical detection anomalies on the standard product image, selects an area from the standard product image that has never had a historical detection anomaly or the number of historical detection anomalies is less than a preset anomaly threshold, and marks it as a detection stable area, and then captures the standard area image of the detection stable area in the standard product image.

[0118] Furthermore, Figure 8 It shows a structural block diagram of the detection stability analysis module 101 in the system provided by an embodiment of the present invention.

[0119] Specifically, in another preferred embodiment provided by the present invention, the detection stability analysis module 101 specifically includes:

[0120] The product determination unit 1011 is used to receive the detection planning information and determine the target industrial product.

[0121] The data acquisition unit 1012 is used to acquire the historical detection data and standard product images of the target industrial product.

[0122] The stability analysis unit 1013 is used to perform a detection stability analysis on the historical detection data based on the standard product image, and determine a detection stability area from the standard product image.

[0123] The standard interception unit 1014 is used to intercept the standard area image of the detection stability area in the standard product image.

[0124] Furthermore, the industrial defect detection system based on artificial intelligence also includes:

[0125] The detection adjustment and shooting module 102 is used to determine the current detection product, perform detection adjustment on the current detection product based on artificial intelligence, and perform detection shooting to obtain the detection shooting image.

[0126] In an embodiment of the present invention, the detection adjustment shooting module 102 obtains the detection adjustment parameters of the target industrial product, determines the current detection product in real time during the process of industrial defect detection of the target industrial product, and based on artificial intelligence, clamps, adjusts the spatial position and / or rotates the current detection product according to the detection adjustment parameters, places the current detection product in a preset detection seat, and then detects and shoots the current detection product to obtain a detection shot image.

[0127] The similarity comparison judgment module 103 is used to intercept the corresponding area image of the detection stable area from the detection shooting image, perform similarity comparison calculation on the corresponding area image and the standard area image, and judge whether similar detection conditions are met.

[0128] In the embodiment of the present invention, the similarity comparison judgment module 103 intercepts the corresponding area image of the detection stable area in the detection captured image, and grayscales the corresponding area image and the standard area image to generate a corresponding grayscale image and a standard grayscale image, and performs pixel value analysis on the corresponding grayscale image and the standard grayscale image to perform similarity comparison calculation, obtain a similarity comparison value, and then compare the similarity comparison value with a preset standard threshold, record the comparison result, and then, based on the comparison result, when the similarity comparison value is greater than the standard threshold, determine that the similarity detection condition exists; and when the similarity comparison value is not greater than the standard threshold, determine that the similarity detection condition does not exist. Specifically, the calculation formula of the similarity comparison value is:

[0129]

[0130] Among them, N is the similarity comparison value, d represents the corresponding grayscale image, s represents the standard grayscale image, and u d is the average brightness corresponding to the grayscale image, u s is the average brightness corresponding to the standard grayscale image, σ dis the standard deviation of the corresponding grayscale image, σ s is the standard deviation corresponding to the standard grayscale image, σ ds is the covariance between the corresponding grayscale image and the standard grayscale image, L is the value range of the pixel value, k1 is the preset first adjustment coefficient, and k2 is the preset second adjustment coefficient.

[0131] Furthermore, Fig. 9 It shows a structural block diagram of the similarity comparison judgment module 103 in the system provided by the embodiment of the present invention.

[0132] Specifically, in another preferred embodiment provided by the present invention, the similarity comparison judgment module 103 specifically includes:

[0133] The corresponding interception unit 1031 is used to intercept the corresponding area image of the detection stable area from the detection shot image.

[0134] The similarity comparison calculation unit 1032 is used to perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value.

[0135] The comparison recording unit 1033 is used to compare the similarity comparison value with a preset standard threshold value and record the comparison result.

[0136] The condition judgment unit 1034 is used to judge whether there are similar detection conditions according to the comparison result.

[0137] Furthermore, the industrial defect detection system based on artificial intelligence also includes:

[0138] The defect detection and marking module 104 is used to perform defect detection on the detection image based on the standard product image when similar detection conditions exist, and to perform detection marking on the current detection product.

[0139] In an embodiment of the present invention, under similar detection conditions, the defect detection marking module 104 performs defect detection on the inspection image based on the standard product image, obtains the defect detection result, and identifies the inspection image to obtain the current product identity of the current inspection product, and then performs inspection marking on the current inspection product based on the current product identity and the defect detection result, thereby realizing background marking of industrial defect detection.

[0140] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0141] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0142] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. An industrial defect detection method based on artificial intelligence, characterized in that: The method specifically comprises the following steps: Acquire historical inspection data and standard product images of target industrial products, perform inspection stability analysis, determine the inspection stability area from the standard product images, and intercept the standard area image; Determine the current inspection product, perform inspection adjustment on the current inspection product based on artificial intelligence, perform inspection shooting, and obtain the inspection shooting image; From the detection shot image, intercept the corresponding area image of the detection stable area, perform similarity comparison calculation on the corresponding area image and the standard area image, and determine whether similar detection conditions are met; When similar inspection conditions are met, defect inspection is performed on the inspection shot image based on the standard product image, and inspection marking is performed on the current inspection product.

2. The industrial defect detection method based on artificial intelligence according to claim 1 is characterized in that: The method of acquiring historical detection data and a standard product image of a target industrial product, performing detection stability analysis, determining a detection stability region from the standard product image, and intercepting a standard region image specifically includes the following steps: Receive testing planning information and determine target industrial products; Acquire historical inspection data and standard product images of the target industrial product; Based on the standard product image, performing a detection stability analysis on the historical detection data, and determining a detection stability area from the standard product image; In the standard product image, a standard area image of the detection stable area is cut out.

3. The industrial defect detection method based on artificial intelligence according to claim 1 is characterized in that: The step of determining the current inspection product, adjusting the inspection of the current inspection product based on artificial intelligence, and performing inspection shooting, and obtaining the inspection shooting image specifically includes the following steps: Obtaining detection adjustment parameters of the target industrial product; Determine the current testing product in real time; Based on artificial intelligence, according to the detection adjustment parameters, the detection adjustment is performed on the current detection product; Perform detection shooting and obtain detection shooting images.

4. The industrial defect detection method based on artificial intelligence according to claim 1 is characterized in that: The method of intercepting the corresponding area image of the detection stable area from the detection shot image, performing similarity comparison calculation on the corresponding area image and the standard area image, and judging whether similar detection conditions are met specifically comprises the following steps: From the detected captured image, intercepting a region image corresponding to the detected stable region; Perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value; Compare the similarity comparison value with a preset standard threshold value and record the comparison result; According to the comparison result, it is determined whether similar detection conditions exist.

5. The artificial intelligence-based industrial defect detection method according to claim 4, characterized in that: The similarity comparison calculation of the corresponding area image and the standard area image to obtain the similarity comparison value specifically includes the following steps: Gray-scale the corresponding region image to generate a corresponding gray-scale image; Gray-scale the standard region image to generate a standard gray-scale image; A similarity comparison calculation is performed on the corresponding grayscale image and the standard grayscale image to obtain a similarity comparison value.

6. The artificial intelligence-based industrial defect detection method according to claim 5, characterized in that: The calculation formula of the similarity comparison value is: Among them, N is the similarity comparison value, d represents the corresponding grayscale image, s represents the standard grayscale image, and u d is the average brightness corresponding to the grayscale image, u s is the average brightness corresponding to the standard grayscale image, σ d is the standard deviation of the corresponding grayscale image, σ s is the standard deviation corresponding to the standard grayscale image, σ ds is the covariance between the corresponding grayscale image and the standard grayscale image, L is the value range of the pixel value, k1 is the preset first adjustment coefficient, and k2 is the preset second adjustment coefficient.

7. The artificial intelligence-based industrial defect detection method according to claim 1, characterized in that: The method of performing defect detection on the inspection image based on the standard product image when similar inspection conditions are met and marking the current inspection product specifically includes the following steps: When similar inspection conditions are met, defect inspection is performed on the inspection image based on the standard product image to obtain a defect inspection result; Identify the detected image to obtain the current product identity of the currently detected product; According to the current product identity and the defect detection result, the current detection product is detected and marked.

8. An industrial defect detection system based on artificial intelligence, characterized in that: The system includes a detection stability analysis module, a detection adjustment shooting module, a similarity comparison judgment module and a defect detection marking module, wherein: A detection stability analysis module is used to obtain historical detection data and standard product images of target industrial products, perform detection stability analysis, determine the detection stability area from the standard product image, and intercept the standard area image; The detection adjustment and shooting module is used to determine the current detection product, perform detection adjustment on the current detection product based on artificial intelligence, and perform detection shooting to obtain the detection shooting image; A similarity comparison judgment module is used to intercept the corresponding area image of the detection stable area from the detection shooting image, perform similarity comparison calculation on the corresponding area image and the standard area image, and judge whether similar detection conditions are met; The defect detection and marking module is used to perform defect detection on the detection shot image based on the standard product image when similar detection conditions exist, and to perform detection marking on the current detection product.

9. The artificial intelligence-based industrial defect detection system according to claim 8, characterized in that: The detection stability analysis module specifically includes: A product determination unit, used to receive the inspection planning information and determine the target industrial product; A data acquisition unit, used to acquire historical inspection data and standard product images of the target industrial product; A stability analysis unit, configured to perform a detection stability analysis on the historical detection data based on the standard product image, and determine a detection stability area from the standard product image; The standard interception unit is used to intercept the standard area image of the detection stability area in the standard product image.

10. The artificial intelligence-based industrial defect detection system according to claim 8, characterized in that: The similarity comparison judgment module specifically includes: A corresponding interception unit, used for intercepting a corresponding area image of the detection stable area from the detection shot image; A similarity comparison calculation unit is used to perform similarity comparison calculation on the corresponding area image and the standard area image to obtain a similarity comparison value; A comparison recording unit, used for comparing the similarity comparison value with a preset standard threshold value and recording the comparison result; The condition judgment unit is used to judge whether there are similar detection conditions according to the comparison result.

Images