CN111768381A - Part defect detection method and device and electronic equipment - Google Patents
Part defect detection method and device and electronic equipment Download PDFInfo
- Publication number
- CN111768381A CN111768381A CN202010605838.9A CN202010605838A CN111768381A CN 111768381 A CN111768381 A CN 111768381A CN 202010605838 A CN202010605838 A CN 202010605838A CN 111768381 A CN111768381 A CN 111768381A
- Authority
- CN
- China
- Prior art keywords
- defect
- size
- algorithm
- images
- result
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Abstract
The application discloses a method and a device for predicting part defects and electronic equipment, and relates to the fields of artificial intelligence, deep learning, cloud computing and computer vision, in particular to the aspect of industrial quality inspection. The specific implementation scheme is as follows: acquiring planar images of a plurality of planes of the part; cutting the plane image according to a reference size to obtain a plurality of sub-images; inputting a plurality of subgraphs into a defect prediction model trained in advance to obtain a defect prediction result of each subgraph; and inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises defect types and defect grades. Through the scheme, on one hand, the real defects of the parts can be accurately identified, so that the accuracy of the defect detection result is improved, and the probability of whether the defects of the parts are judged by mistake is reduced. On the other hand, the input data of the defect prediction model is convenient to process, and the labor cost can be reduced.
Description
Technical Field
The application relates to the fields of artificial intelligence, deep learning, cloud computing and computer vision, in particular to the field of industrial quality inspection.
Background
With the development of artificial intelligence technology, the computer vision is used for landing on various scenes, such as industrial quality inspection, power inspection, unmanned driving, intelligent retail and the like, and the artificial intelligence technology is relied on to complete specific tasks. At present, the combination of industrial production and deep learning tends to be great, on one hand, the performance and the stability can be ensured, and on the other hand, the labor cost can be greatly reduced. Generally, the requirements of different industrial scenes need to be customized in design scheme, but a set of requirements which can be expanded into the same or similar industrial scenes can be extracted by performing abstract representation on the requirements set forth by the scenes.
Aiming at the problem of whether a part product has flaws or defects in an industrial quality inspection scene, the prior art provides a part defect detection method based on deep learning, the defects are identified based on an end-to-end deep learning model, and marking information accurate to the pixel level is relied on, so that the required labor cost is high, and the condition that the defects of all planes of the part have correlation cannot be processed.
Disclosure of Invention
The application provides a method and a device for detecting defects of parts and electronic equipment.
According to a first aspect of the present application, a method for detecting a defect of a component is provided, which includes:
acquiring planar images of a plurality of planes of the part;
cutting the plane image according to a reference size to obtain a plurality of sub-images;
inputting a plurality of subgraphs into a defect prediction model trained in advance to obtain a defect prediction result of each subgraph;
and inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises defect types and defect grades.
According to a second aspect of the present application, there is provided a component defect detecting apparatus, including:
the acquisition module is used for acquiring plane images of a plurality of planes of the part;
the cutting module is used for cutting the plane image according to the reference size to obtain a plurality of sub-images;
the prediction module is used for inputting a plurality of sub-images into a defect prediction model trained in advance to obtain a defect prediction result of each sub-image;
and the recognition module is used for inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises a defect type and a defect grade.
According to a third aspect of the present application, there is provided an electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to cause the at least one processor to perform a method provided by any one of the embodiments of the present application.
According to a fourth aspect of the present application, there is provided a non-transitory computer readable storage medium storing computer instructions, wherein the computer instructions are configured to cause a computer to perform the method provided by any one of the embodiments of the present application.
Through the scheme, on one hand, the real defects of the parts can be accurately identified, so that the accuracy of the defect detection result is improved, and the probability of whether the defects of the parts are judged by mistake is reduced. On the other hand, the input data of the defect prediction model is convenient to process, and the labor cost can be reduced.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present application, nor do they limit the scope of the present application. Other features of the present application will become apparent from the following description.
Drawings
The drawings are included to provide a better understanding of the present solution and are not intended to limit the present application. Wherein:
FIG. 1 is a flow chart of a method for detecting defects of a component according to a first embodiment of the present application;
fig. 2 is a flowchart of a method for detecting a defect of a component according to a first embodiment of the present application, in which a planar image is cut into a plurality of sub-images;
FIG. 3 is a flowchart of a method for detecting defects of a component according to a first embodiment of the present application;
fig. 4 is a schematic diagram of a planar image cut into a plurality of sub-images according to a part defect detection method according to a first embodiment of the present application;
fig. 5 is a schematic diagram of a planar image cut to obtain a plurality of sub-image samples according to the component defect detection method according to the first embodiment of the present application;
FIG. 6 is a schematic diagram of a part defect detection method according to a first embodiment of the present application;
FIG. 7 is a block diagram of a component defect detection apparatus according to a second embodiment of the present application;
fig. 8 is a block diagram of a cutting module of a part defect detecting apparatus according to a second embodiment of the present application;
fig. 9 is a block diagram of an electronic device for implementing the component defect detection method according to the embodiment of the present application.
Detailed Description
The following description of the exemplary embodiments of the present application, taken in conjunction with the accompanying drawings, includes various details of the embodiments of the application for the understanding of the same, which are to be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 shows a flowchart of a part defect detection method according to an embodiment of the present application.
As shown in fig. 1, the method for detecting the defect of the component includes:
step S101: planar images of a plurality of planes of the part are acquired. Illustratively, the part can be a part which needs to be subjected to defect detection in an industrial quality inspection scene, and the outer surface of the part has a plurality of planes. The planar image may be any form of planar image acquired by an optical device, such as an RGB (Red Green Blue) format image or a grayscale image. By acquiring the planar images of a plurality of planes of the part, the three-dimensional part can be converted into a plurality of two-dimensional planar images, and the defect detection is carried out on the part by a two-dimensional image processing technology.
Step S102: and cutting the plane image according to the reference size to obtain a plurality of sub-images. Because the three-dimensional size of the part can be any size, the obtained planar image of each plane of the part is irregular in shape and inconsistent in size, and therefore the planar image needs to be cut according to the reference size to obtain a subgraph with a uniform size meeting the input requirement of the defect prediction model.
Step S103: and inputting the multiple subgraphs into a defect prediction model trained in advance to obtain a defect prediction result of each subgraph. The defect prediction model can be obtained by training at least one of other algorithms such as a target detection algorithm, a segmentation algorithm and an identification algorithm.
Step S104: and inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises defect types and defect grades. The defect identification model can be obtained by training at least one of other algorithms such as an identification algorithm and a retrieval algorithm.
The defect types may include types of mainstream defects on an industrial quality inspection line, such as several tens of defect types, e.g., cracks, bumps, scratches, weld marks, deformations, burrs, and stock shortages. The defect level may be divided according to specific situations, such as the length, width, depth and area of the defect region.
The part defect detection method provided by the embodiment of the application can be used for detecting whether defects or flaws exist in a part product in an industrial quality inspection scene. The method comprises the steps of obtaining planar images of multiple planes of a part, cutting the planar images of the multiple planes to obtain multiple sub-images, inputting the multiple sub-images into a defect prediction model to obtain defect prediction results of the sub-images, inputting the defect prediction results of the sub-images into a defect recognition model to further obtain the defect recognition results of the sub-images, and judging the defect types and the defect grades of the sub-images according to the defect recognition results. The embodiment of the application solves the labor cost problem and the time problem of pure manual quality inspection, effectively promotes the development of the industrial quality inspection field based on computer vision, and is beneficial to the development of the national 'new capital construction' strategy.
In addition, in a scene that the defect types and the defect grades of the parts need to be distinguished, the defect prediction results are identified through the defect identification model, so that the identification precision of the defects in the subgraphs of the planes of the parts is improved, the defect conditions of the parts are judged by integrating the defect identification results of the planes, the accurate identification of the real defects of the parts can be realized, the precision of the defect detection results is improved, and the probability of whether the defects of the parts are judged by mistake is reduced.
In addition, the planar images of the planes of the parts are directly obtained, the planar images are subjected to cutting preprocessing, and then the defect prediction model can be input for defect prediction.
In one embodiment, as shown in fig. 2, step S102 includes:
step S201: under the condition that the height size and the width size of the plane image are not in integral multiple relation, expanding the plane image so that the height size and the width size of the expanded plane image are in integral multiple relation;
step S202: and cutting the plane image by taking the smaller size of the height size and the width size of the expanded plane image as a reference size to obtain a plurality of subgraphs.
In one example, as shown in fig. 4, step S102 may adopt an adaptive method to perform cutting on the plane image. Therefore, the size of each plane self-adaptive cutting graph can be arranged according to the performance of the machine, so that the accuracy of the defect prediction of the sub-graph of the part by the defect prediction model is improved, and the training cost of the defect prediction model can be controlled.
Specifically, the height dimension and the width dimension of the plane image are acquired, and whether the height dimension and the width dimension of the plane image are in integral multiple relation is judged. In the case where the height size and the width size of the planar image are in an integral multiple relationship, the smaller one of the height size and the width size of the planar image may be taken as a reference size, and then the planar image may be cut according to the reference size to divide the planar image into a plurality of sub-images of uniform size. The height and width of the subgraph can be the same, and the reference size can be an integral multiple of the length of the subgraph. For example, in the example shown in fig. 4, the height and width dimensions of the planar image are the same, the reference dimension may be the side length of the planar image, the side length of the sub-image to be cut may be 1/2 of the reference dimension, and four sub-images with the same size are obtained after the planar image is cut.
When the height and width dimensions of the planar image are not in an integral multiple relationship, the size having a larger value among the height and width dimensions of the planar image is expanded so that the size having a larger value becomes an integral multiple of the size having a smaller value after the expansion. Then, the size with the smaller value is cut as a reference size. The subsequent cutting method is the same as that in the case where the height and width of the midplane image are integer multiples, and therefore, the description thereof is omitted.
In one embodiment, the training samples of the defect prediction model are the same size as the reference size.
In one example, as shown in fig. 5, planar objects of various sizes are obtained, the positions of defects in the planar image are manually calibrated, and then the manually calibrated defect regions in the planar image are cut into uniform sizes according to a reference size, so as to obtain a plurality of sub-image training samples. And the plurality of sub-graph training samples are input into the defect prediction model to train the defect prediction model.
In one embodiment, the defect prediction model is constructed using at least one of an object detection algorithm, a semantic segmentation algorithm, and an identification algorithm.
In one example, the defect prediction model may be constructed using a multi-target detection algorithm or a single-target detection algorithm. Specifically, the cut subgraph is input into a defect prediction model, and a candidate frame which is possibly a defect in the subgraph is obtained through a multi-target detection algorithm or a single-target detection algorithm. And then judging whether the candidate frame is a defect or not according to the credibility and the category of the prediction frame, and finally outputting a defect prediction result. The single-target detection algorithm can be a single-stage target detection method represented by YOLO, and comprises a YOLO (you Only Look one) series method. For example, the YOLOv2 series algorithm, the YOLO9000 series algorithm, the YOLOv3 series algorithm, and the like may be used. The multi-target detection algorithm may be a two-stage target detection algorithm represented by fast RCNN. For example, an R-CNN algorithm, SPPNet algorithm, Fast-RCNN algorithm, FPN algorithm, etc.
In another example, the defect prediction model may be constructed using an instance segmentation algorithm or a semantic segmentation algorithm. Specifically, the cut subgraph is input into a defect prediction model, and a binarization prediction graph aiming at a defect area in the subgraph can be obtained through an example segmentation algorithm or a semantic segmentation algorithm. And then calculating pixel points of the defects and the proportion of the defects in the whole subgraph to predict whether the region is the defects or not and predict the defect grade of the defects to obtain a defect prediction result. In other examples of the present application, the algorithm used by the defect prediction model may also be at least one of region-based image segmentation algorithm, edge detection-based image segmentation algorithm, genetic algorithm-based image segmentation algorithm, clustering-based image segmentation algorithm, graph-based image segmentation algorithm, and Mask R-CNN and R-FCN.
In other examples, the defect prediction model may also be constructed using a recognition algorithm. Specifically, the cut subgraph is input into a defect prediction model, the prediction category of the defect in the subgraph can be obtained through a recognition algorithm, then whether the image is defective or not can be judged according to the prediction reliability, and finally a defect prediction result is output.
In one embodiment, the defect identification model may be constructed using a classification identification algorithm or a search algorithm.
In one example, the defect identification model is constructed using a classification identification algorithm and the defect prediction model is constructed using an object detection algorithm. Specifically, the defect prediction result output by the defect prediction model comprises a candidate frame, the candidate frame is input into the defect recognition model, the defects in the candidate frame are finely classified through a classification recognition algorithm, the defect category or the defect grade corresponding to the defects of the subgraph in the candidate frame is obtained, and the defect recognition result is output.
In another example, the defect identification model may be constructed using a search algorithm. Specifically, the defect prediction result output by the defect prediction model comprises a candidate frame, the candidate frame is input into the defect identification model, the defects in the candidate frame are searched in the index library through a search algorithm, the labels closest to the defects in the candidate frame in the index library are obtained, and the defect identification result is obtained. The index library may be obtained by performing feature extraction on the candidate frame sample in a training stage of the defect identification model.
The part defect detection method of the embodiment of the application carries out defect prediction and defect identification on the plane images of all planes of the part. When a defect is detected in one or more planes of the component, the whole component is determined to have a defect, which may result in over-killing the component that is actually good due to a determination error, and therefore, the defect recognition results of the planes of the component need to be fused.
In one embodiment, as shown in fig. 3, the method for detecting a defect of a component further includes:
step S301: counting the defect condition of each plane of the part according to the defect identification result of the subgraph;
step S302: and fusing the defect conditions of each plane by adopting at least one of a multi-threshold fusion algorithm, a voting fusion algorithm and a weight voting fusion algorithm to obtain a defect detection result.
In one example, the defect conditions of the planes are fused through a multi-threshold fusion algorithm, and a defect detection result is obtained. Because the probability of the defects of different planes of the part is different and the defect conditions of all the planes are different, different thresholds are set for different planes of the part by counting the probability of the defects of all the planes of the part, the yield over-killing rate of all the planes of one part in the step S103 and the step S104 and the like, and then the defect identification results of a plurality of planes of the part are fused according to the thresholds of all the planes to obtain the defect detection result.
In another example, the defect conditions of the planes are fused through a weight voting fusion algorithm, and a defect detection result is obtained. It can be understood that, since the importance levels of the respective planes of the component are different, different weights are given to the respective planes for the importance levels of the respective planes of the component, and then the voting result is multiplied by the weights as a final defect detection result.
The defect conditions of all planes are fused to obtain a defect detection result, so that the accurate identification of the real defects of the parts can be further improved, and the probability of the condition of recalling the parts due to misjudgment is reduced.
The method for detecting the defect of the component part according to the embodiment of the present application is described below as a specific example with reference to fig. 6.
As shown in fig. 6, first, optical scheme recording is performed on the planes of all the components, for example, by taking a picture with a camera or infrared imaging, and a plane image of each plane of the components is acquired. And then carrying out self-adaptive processing on the plane images to obtain a plurality of subgraphs of each plane image.
And inputting a plurality of subgraphs into the defect prediction model as subgraph samples in a training stage aiming at the defect prediction model, and training the defect prediction model.
And inputting each subgraph into the defect prediction model in a prediction stage aiming at the defect prediction model to obtain a defect prediction result of each subgraph. And then, post-processing the defect prediction result of each subgraph, namely inputting the defect prediction result of each subgraph into a trap recognition model to recognize the defects of the subgraph in the defect prediction result to obtain the defect recognition result of the subgraph, wherein the defect recognition result comprises defect types and defect grades. And finally, fusing the defect identification results of the sub-images of the planes to obtain a defect detection result of the part, and judging whether the part has defects or not according to the defect detection result.
As shown in fig. 7, in one embodiment, the present application further provides a component defect detecting apparatus 400, including:
an obtaining module 410, configured to obtain planar images of multiple planes of the component;
a cutting module 420, configured to cut the planar image according to a reference size to obtain multiple sub-images;
the prediction module 430 is configured to input the multiple sub-images into a defect prediction model trained in advance, so as to obtain a defect prediction result of each sub-image;
and the recognition module 440 is configured to input the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each sub-graph, where the defect recognition result includes a defect type and a defect grade.
In one embodiment, as shown in fig. 8, the cutting module 420 further comprises:
an image expansion submodule 421 configured to expand the planar image so that the height size and the width size of the expanded planar image are in an integer multiple relationship when the height size and the width size of the planar image are not in an integer multiple relationship;
and the cutting sub-module 422 is configured to cut the planar image by using a smaller one of the height and width dimensions of the expanded planar image as a reference dimension, so as to obtain a plurality of sub-images.
In one embodiment, the defect prediction model is constructed using at least one of an object detection algorithm, a semantic segmentation algorithm, and an identification algorithm.
In one embodiment, as shown in fig. 7, the component defect detecting apparatus 400 further includes:
the statistical module 450 is used for counting the defect condition of each plane of the part according to the defect identification result of the subgraph;
and the fusion module 460 is configured to fuse the defect conditions of each plane by using at least one of a multi-threshold fusion algorithm, a voting fusion algorithm, and a weighted voting fusion algorithm to obtain a defect detection result.
In one embodiment, the training samples of the defect prediction model are the same size as the reference size.
According to an embodiment of the present application, an electronic device and a readable storage medium are also provided.
Fig. 9 is a block diagram of an electronic device according to the method for detecting a defect of a component according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.
As shown in fig. 9, the electronic apparatus includes: one or more processors 501, memory 502, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). Fig. 9 illustrates an example of one processor 501.
The memory 502, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the part defect detection method in the embodiments of the present application. The processor 501 executes various functional applications and data processing of the server by executing non-transitory software programs, instructions and modules stored in the memory 502, that is, implements the component defect detection method in the above method embodiment.
The memory 502 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of the electronic device of the part defect detection method, and the like. Further, the memory 502 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 502 may optionally include a memory remotely located from the processor 501, and these remote memories may be connected to the electronic device of the part defect detection method through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The electronic device of the part defect detecting method may further include: an input device 503 and an output device 504. The processor 501, the memory 502, the input device 503 and the output device 504 may be connected by a bus or other means, and fig. 9 illustrates the connection by a bus as an example.
The input device 503 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the electronic apparatus of the part defect detecting method, such as an input device of a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointing stick, one or more mouse buttons, a track ball, a joystick, etc. The output devices 504 may include a display device, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibrating motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present application may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present application can be achieved, and the present invention is not limited herein.
The above-described embodiments should not be construed as limiting the scope of the present application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (12)
1. A method for detecting defects of parts comprises the following steps:
acquiring planar images of a plurality of planes of the part;
cutting the plane image according to a reference size to obtain a plurality of sub-images;
inputting a plurality of subgraphs into a defect prediction model trained in advance to obtain a defect prediction result of each subgraph;
and inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises defect types and defect grades.
2. The method of claim 1, wherein cutting the planar image by a reference size to obtain a plurality of subgraphs comprises:
under the condition that the height size and the width size of the plane image are not in integral multiple relation, expanding the plane image so that the height size and the width size of the expanded plane image are in integral multiple relation;
and cutting the plane image by taking the smaller size of the height size and the width size of the expanded plane image as the reference size to obtain a plurality of sub-images.
3. The method of claim 1, the defect prediction model being constructed using at least one of an object detection algorithm, a semantic segmentation algorithm, and an identification algorithm.
4. The method of claim 1, further comprising:
counting the defect condition of each plane of the part according to the defect identification result of the subgraph;
and fusing the defect condition of each plane by adopting at least one of a multi-threshold fusion algorithm, a voting fusion algorithm and a weight voting fusion algorithm to obtain a defect detection result.
5. The method of any of claims 1-4, wherein a size of a training sample of the defect prediction model is the same as the reference size.
6. A component defect detecting apparatus comprising:
the acquisition module is used for acquiring plane images of a plurality of planes of the part;
the cutting module is used for cutting the plane image according to the reference size to obtain a plurality of sub-images;
the prediction module is used for inputting a plurality of sub-images into a defect prediction model trained in advance to obtain a defect prediction result of each sub-image;
and the recognition module is used for inputting the defect prediction result into a defect recognition model trained in advance to obtain a defect recognition result of each subgraph, wherein the defect recognition result comprises a defect type and a defect grade.
7. The apparatus of claim 6, wherein the cutting module further comprises:
the image expansion submodule is used for expanding the plane image under the condition that the height size and the width size of the plane image are not in integral multiple relation, so that the height size and the width size of the expanded plane image are in integral multiple relation;
and the cutting submodule is used for cutting the plane image by taking the smaller size of the height size and the width size of the expanded plane image as the reference size to obtain a plurality of sub-images.
8. The apparatus of claim 6, the defect prediction model is constructed using at least one of an object detection algorithm, a semantic segmentation algorithm, and an identification algorithm.
9. The apparatus of claim 6, further comprising:
the statistical module is used for counting the defect condition of each plane of the part according to the defect identification result of the subgraph;
and the fusion module is used for fusing the defect conditions of each plane by adopting at least one of a multi-threshold fusion algorithm, a voting fusion algorithm and a weight voting fusion algorithm to obtain a defect detection result.
10. The apparatus of any of claims 6-9, wherein a size of a training sample of the defect prediction model is the same as the reference size.
11. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 5.
12. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010605838.9A CN111768381A (en) | 2020-06-29 | 2020-06-29 | Part defect detection method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010605838.9A CN111768381A (en) | 2020-06-29 | 2020-06-29 | Part defect detection method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111768381A true CN111768381A (en) | 2020-10-13 |
Family
ID=72723002
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010605838.9A Pending CN111768381A (en) | 2020-06-29 | 2020-06-29 | Part defect detection method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111768381A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112287922A (en) * | 2020-10-29 | 2021-01-29 | 北京百度网讯科技有限公司 | Meter reading method, device, equipment and storage medium |
| CN112581463A (en) * | 2020-12-25 | 2021-03-30 | 北京百度网讯科技有限公司 | Image defect detection method and device, electronic equipment, storage medium and product |
| CN112734690A (en) * | 2020-12-17 | 2021-04-30 | 郑州金惠计算机系统工程有限公司 | Surface defect detection method and device and computer readable storage medium |
| CN112907543A (en) * | 2021-02-24 | 2021-06-04 | 胡志雄 | Product appearance defect detection method based on random defect model |
| CN113160176A (en) * | 2021-04-23 | 2021-07-23 | 歌尔股份有限公司 | Defect detection method and device |
| CN113222967A (en) * | 2021-05-28 | 2021-08-06 | 长江存储科技有限责任公司 | Wafer detection method and system |
| CN113256624A (en) * | 2021-06-29 | 2021-08-13 | 中移(上海)信息通信科技有限公司 | Continuous casting round billet defect detection method and device, electronic equipment and readable storage medium |
| CN113256615A (en) * | 2021-06-22 | 2021-08-13 | 常州微亿智造科技有限公司 | Method and device for detecting linear defects |
| CN113327241A (en) * | 2021-06-16 | 2021-08-31 | 南昌航空大学 | Visual detection method and system for surface defects of bearing end face |
| CN113674203A (en) * | 2021-07-14 | 2021-11-19 | 歌尔股份有限公司 | Defect detection model training method and device and defect detection method and device |
| CN113706490A (en) * | 2021-08-19 | 2021-11-26 | 中科苏州智能计算技术研究院 | Wafer defect detection method |
| CN113902742A (en) * | 2021-12-08 | 2022-01-07 | 中导光电设备股份有限公司 | TFT-LCD detection-based defect true and false judgment method and system |
| CN114820543A (en) * | 2022-05-07 | 2022-07-29 | 苏州方石科技有限公司 | Defect detection method and device |
| CN115239644A (en) * | 2022-07-05 | 2022-10-25 | 港珠澳大桥管理局 | Concrete defect identification method and device, computer equipment and storage medium |
| CN116167992A (en) * | 2023-02-02 | 2023-05-26 | 北京百度网讯科技有限公司 | Defect detection method, defect detection device, electronic device and computer readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108460761A (en) * | 2018-03-12 | 2018-08-28 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN108921811A (en) * | 2018-04-03 | 2018-11-30 | 阿里巴巴集团控股有限公司 | Detect method and apparatus, the article damage detector of article damage |
| CN109658419A (en) * | 2018-11-15 | 2019-04-19 | 浙江大学 | The dividing method of organella in a kind of medical image |
| CN109949286A (en) * | 2019-03-12 | 2019-06-28 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
| CN110176064A (en) * | 2019-05-24 | 2019-08-27 | 武汉大势智慧科技有限公司 | A kind of photogrammetric main object automatic identifying method for generating threedimensional model |
| CN111046926A (en) * | 2019-11-26 | 2020-04-21 | 山东浪潮人工智能研究院有限公司 | Computer vision image classification integrated learning method |
| CN111209905A (en) * | 2020-04-23 | 2020-05-29 | 北京慧智数据科技有限公司 | Defect shielding license plate recognition method based on combination of deep learning and OCR technology |
-
2020
- 2020-06-29 CN CN202010605838.9A patent/CN111768381A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108460761A (en) * | 2018-03-12 | 2018-08-28 | 北京百度网讯科技有限公司 | Method and apparatus for generating information |
| CN108921811A (en) * | 2018-04-03 | 2018-11-30 | 阿里巴巴集团控股有限公司 | Detect method and apparatus, the article damage detector of article damage |
| CN109658419A (en) * | 2018-11-15 | 2019-04-19 | 浙江大学 | The dividing method of organella in a kind of medical image |
| CN109949286A (en) * | 2019-03-12 | 2019-06-28 | 北京百度网讯科技有限公司 | Method and apparatus for output information |
| CN110176064A (en) * | 2019-05-24 | 2019-08-27 | 武汉大势智慧科技有限公司 | A kind of photogrammetric main object automatic identifying method for generating threedimensional model |
| CN111046926A (en) * | 2019-11-26 | 2020-04-21 | 山东浪潮人工智能研究院有限公司 | Computer vision image classification integrated learning method |
| CN111209905A (en) * | 2020-04-23 | 2020-05-29 | 北京慧智数据科技有限公司 | Defect shielding license plate recognition method based on combination of deep learning and OCR technology |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112287922A (en) * | 2020-10-29 | 2021-01-29 | 北京百度网讯科技有限公司 | Meter reading method, device, equipment and storage medium |
| CN112734690A (en) * | 2020-12-17 | 2021-04-30 | 郑州金惠计算机系统工程有限公司 | Surface defect detection method and device and computer readable storage medium |
| CN112581463B (en) * | 2020-12-25 | 2024-02-27 | 北京百度网讯科技有限公司 | Image defect detection method and device, electronic equipment, storage medium and product |
| CN112581463A (en) * | 2020-12-25 | 2021-03-30 | 北京百度网讯科技有限公司 | Image defect detection method and device, electronic equipment, storage medium and product |
| CN112907543A (en) * | 2021-02-24 | 2021-06-04 | 胡志雄 | Product appearance defect detection method based on random defect model |
| CN112907543B (en) * | 2021-02-24 | 2024-03-26 | 胡志雄 | Product appearance defect detection method based on random defect model |
| CN113160176A (en) * | 2021-04-23 | 2021-07-23 | 歌尔股份有限公司 | Defect detection method and device |
| CN113222967A (en) * | 2021-05-28 | 2021-08-06 | 长江存储科技有限责任公司 | Wafer detection method and system |
| CN113327241A (en) * | 2021-06-16 | 2021-08-31 | 南昌航空大学 | Visual detection method and system for surface defects of bearing end face |
| CN113256615A (en) * | 2021-06-22 | 2021-08-13 | 常州微亿智造科技有限公司 | Method and device for detecting linear defects |
| CN113256615B (en) * | 2021-06-22 | 2021-10-19 | 常州微亿智造科技有限公司 | Method and device for detecting linear defects |
| CN113256624A (en) * | 2021-06-29 | 2021-08-13 | 中移(上海)信息通信科技有限公司 | Continuous casting round billet defect detection method and device, electronic equipment and readable storage medium |
| CN113674203A (en) * | 2021-07-14 | 2021-11-19 | 歌尔股份有限公司 | Defect detection model training method and device and defect detection method and device |
| CN113706490B (en) * | 2021-08-19 | 2024-01-12 | 中科苏州智能计算技术研究院 | Wafer defect detection method |
| CN113706490A (en) * | 2021-08-19 | 2021-11-26 | 中科苏州智能计算技术研究院 | Wafer defect detection method |
| CN113902742B (en) * | 2021-12-08 | 2022-05-20 | 中导光电设备股份有限公司 | TFT-LCD detection-based defect true and false judgment method and system |
| CN113902742A (en) * | 2021-12-08 | 2022-01-07 | 中导光电设备股份有限公司 | TFT-LCD detection-based defect true and false judgment method and system |
| CN114820543A (en) * | 2022-05-07 | 2022-07-29 | 苏州方石科技有限公司 | Defect detection method and device |
| CN115239644A (en) * | 2022-07-05 | 2022-10-25 | 港珠澳大桥管理局 | Concrete defect identification method and device, computer equipment and storage medium |
| CN115239644B (en) * | 2022-07-05 | 2024-04-05 | 港珠澳大桥管理局 | Concrete defect identification method, device, computer equipment and storage medium |
| CN116167992A (en) * | 2023-02-02 | 2023-05-26 | 北京百度网讯科技有限公司 | Defect detection method, defect detection device, electronic device and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111768381A (en) | Part defect detection method and device and electronic equipment | |
| CN111860506A (en) | Method and device for recognizing characters | |
| CN111693534A (en) | Surface defect detection method, model training method, device, equipment and medium | |
| CN111598164A (en) | Method and device for identifying attribute of target object, electronic equipment and storage medium | |
| CN112507949A (en) | Target tracking method and device, road side equipment and cloud control platform | |
| EP3852008A2 (en) | Image detection method and apparatus, device, storage medium and computer program product | |
| CN112633276A (en) | Training method, recognition method, device, equipment and medium | |
| CN111967490A (en) | Model training method for map detection and map detection method | |
| CN112508003A (en) | Character recognition processing method and device | |
| CN112668586A (en) | Model training method, image processing device, storage medium, and program product | |
| CN110717933A (en) | Post-processing method, device, equipment and medium for moving object missed detection | |
| CN111814633A (en) | Method, device, equipment and storage medium for detecting display scene | |
| CN111783760A (en) | Character recognition method and device, electronic equipment and computer readable storage medium | |
| CN115861400B (en) | Target object detection method, training device and electronic equipment | |
| CN111783639A (en) | Image detection method and device, electronic equipment and readable storage medium | |
| CN114419035B (en) | Product identification method, model training device and electronic equipment | |
| CN113326796B (en) | Object detection method, model training method and device and electronic equipment | |
| CN111563541B (en) | Training method and device of image detection model | |
| CN111832658B (en) | Point-of-interest information processing method and device, electronic equipment and storage medium | |
| CN112749701A (en) | Method for generating license plate contamination classification model and license plate contamination classification method | |
| CN111696095A (en) | Method and device for detecting surface defects of object | |
| CN111967299B (en) | Unmanned aerial vehicle inspection method, unmanned aerial vehicle inspection device, unmanned aerial vehicle inspection equipment and storage medium | |
| CN111753960B (en) | Model training and image processing method and device, electronic equipment and storage medium | |
| CN112183484B (en) | Image processing method, device, equipment and storage medium | |
| CN113887394A (en) | Image processing method, device, equipment and storage medium |
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 |