Specification checking method and device based on artificial intelligence

Abstract

The embodiment of the application discloses a kind of normative checking method and device based on artificial intelligence, first receive and respond target detection instruction, select one or more target detection objects based on the target detection instruction, in order to better detect the operation article or operation action corresponding to target detection object, the target detection area corresponding to these target detection objects is acquired in advance, and the two are bound to form a normative check target;Image data is collected by image acquisition equipment, including static picture data and dynamic video tracking data, the data obtained after image data is preprocessed is more accurate, which helps to better extract feature points and then accurately identify targets;Embodiment greatly improves the monitoring feasibility and flexibility of student experiment process based on artificial intelligence, helps to accurately judge the operability of student experiment, and improves efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a specification checking method and device based on artificial intelligence. BACKGROUND

[0002] At present, the state is developing primary and middle school physics and chemistry experiments to improve students' practical ability, so experiments are not only a verification tool for scientific theory, but also a necessary link for students to absorb knowledge and improve practical ability. However, the laboratory environment is complex and changeable, and various instruments and equipment are involved, which has a strong requirement on the operation ability of students. The traditional experimental teaching and operation management mode often has the following problems: first, due to the complexity of the instrument and the cumbersome operation process, especially for beginners, it is easy to cause misoperation. For example, the assembly of experimental equipment does not meet the specifications, which may cause experimental failure or inaccurate results, and even cause safety accidents in some cases. Secondly, for teachers, the timeliness and accuracy of teaching are limited in the face of a large number of students' experimental operation. Teachers cannot effectively supervise and guide in each experimental link, resulting in low experimental learning quality of students. SUMMARY

[0003] In view of the defects, the embodiment of the present application discloses a specification checking method and device based on artificial intelligence, which can help improve the safety and standardization of experimental operation through intelligent automatic detection.

[0004] The first aspect of the embodiment of the present application discloses a specification checking method based on artificial intelligence, comprising:

[0005] In response to a target detection instruction, a target detection object is selected, a target detection area corresponding to the target detection object is obtained, and the target detection area and the target detection object are associated to obtain a specification checking target;

[0006] Image data of the specification checking target is collected, the image data includes picture data and video tracking data, the image data is preprocessed, feature points are extracted from the preprocessed image data, and an identification result of the specification checking target is obtained according to the feature points, the identification result includes a detection standard, a position and a type.

[0007] As an optional implementation manner, in the first aspect of the embodiment of the present application, the pre-processing of the image data comprises:

[0008] The image data is denoised based on bilateral filtering;

[0009] The denoised image data is contrast adjusted;

[0010] Adjust the size of the image data, and identify an interference area in the image data, and crop the interference area.

[0011] As an optional implementation, in the first aspect of the embodiment of the present application, the feature points are extracted from the preprocessed image data, including:

[0012] The preprocessed image data is input into a preset YOLO deep learning model to extract the feature points of the image data.

[0013] As an optional implementation, in the first aspect of the embodiment of the present application, the feature points are extracted from the preprocessed image data, further including:

[0014] The preprocessed image data is input into a preset hand key point detection model and an eye key point detection model to extract the feature points.

[0015] As an optional implementation, in the first aspect of the embodiment of the present application, the identification result of the specification check target is obtained according to the feature points, including:

[0016] The image data is divided into a plurality of identification areas, and the code corresponding to each identification area is obtained according to the feature points.

[0017] The lead connection information between each two identification areas is obtained to create a directed graph adjacency list, and the preset current flow direction of each identification area is obtained, and a node is added in the directed graph adjacency list based on the preset current flow direction.

[0018] Based on the directed graph adjacency list after adding the node, it is judged whether there is a current loop in the image data to obtain the circuit connection information of the image data, and a corresponding identification result is generated according to the circuit connection information.

[0019] As an optional implementation, in the first aspect of the embodiment of the present application, further including:

[0020] When the visual tracking data detects that the specification check target moves, the moving image of the specification check target is extracted at a preset interval to generate a change path, and it is judged whether to generate a warning information according to the change path.

[0021] As an optional implementation, in the first aspect of the embodiment of the present application, when the visual tracking data detects that the specification check target moves, the moving image of the specification check target is extracted at a preset interval to generate a change path, including:

[0022] When the visual tracking data detects that the specification check target moves, the video tracking data is cut according to a preset interval to divide into a plurality of moving images;

[0023] Identify and mark the standard inspection target in each moving image, and obtain the position information of the standard inspection target in each video segment;

[0024] Generate the change path of the standard inspection target based on the position information of the standard inspection target of the two adjacent moving images.

[0025] As an optional implementation, in the first aspect of the embodiment of the present application, the judging whether to generate the early warning information according to the change path comprises:

[0026] Obtaining the change type of the standard inspection target according to the change path, and selecting the corresponding standard inspection target transformation reference according to the change type;

[0027] Comparing the change path with the standard inspection target transformation reference to calculate the matching degree between the change path and the standard inspection target transformation reference, and generating the early warning information when the matching degree is lower than a preset value.

[0028] The second aspect of the embodiment of the present application discloses a standard inspection device based on artificial intelligence, comprising:

[0029] The instruction response module is used for selecting a target detection object in response to a target detection instruction, obtaining a target detection area corresponding to the target detection object, and associating the target detection area and the target detection object to obtain a standard inspection target.

[0030] The image acquisition module is used for acquiring image data of the standard inspection target, wherein the image data comprises picture data and video tracking data, pre-processing the image data, extracting feature points from the pre-processed image data, and obtaining an identification result of the standard inspection target according to the feature points, wherein the identification result comprises a detection standard, a position and a type.

[0031] As an optional implementation, in the second aspect of the embodiment of the present application, the pre-processing the image data comprises:

[0032] The image data is de-noised based on bilateral filtering;

[0033] The de-noised image data is contrast-adjusted;

[0034] The size of the image data is adjusted, and an interference area in the image data is identified, and the interference area is cropped.

[0035] As an optional implementation, in the second aspect of the embodiment of the present application, the feature points are extracted from the pre-processed image data, comprising:

[0036] The preprocessed image data is input into a preset YOLO deep learning model to extract feature points of the image data.

[0037] As an optional implementation, in the second aspect of the embodiment of the present application, the feature points are extracted from the preprocessed image data, and the method further comprises:

[0038] The preprocessed image data is input into a preset hand key point detection model and an eye key point detection model to extract feature points.

[0039] As an optional implementation, in the second aspect of the embodiment of the present application, the identification result of the specification inspection target is obtained according to the feature points, and the method further comprises:

[0040] The image data is divided into a plurality of identification regions, and the encoding corresponding to each identification region is obtained according to the feature points.

[0041] The lead connection information between each two identification regions is obtained to create a directed graph adjacency list, and the preset current flow direction of each identification region is obtained, and a node is added in the directed graph adjacency list based on the preset current flow direction.

[0042] Based on the directed graph adjacency list after the node is added, it is judged whether there is a current loop in the image data, so as to obtain the circuit connection information of the image data, and a corresponding identification result is generated according to the circuit connection information.

[0043] As an optional implementation, in the second aspect of the embodiment of the present application, the method further comprises that a target comparison module is configured to, when the movement of the specification inspection target is detected based on the visual tracking data, extract the moving image of the specification inspection target at a preset interval to generate a change path, and judge whether to generate a warning information according to the change path.

[0044] As an optional implementation, in the second aspect of the embodiment of the present application, when the movement of the specification inspection target is detected based on the visual tracking data, the moving image of the specification inspection target is extracted at a preset interval to generate a change path, and the method further comprises:

[0045] When the movement of the specification inspection target is detected based on the visual tracking data, the video tracking data is cut according to a preset interval to divide into a plurality of moving images;

[0046] The specification inspection target in each moving image is identified and marked, and the position information of the specification inspection target in each video segment is obtained.

[0047] The change path of the specification inspection target is generated based on the position information of the specification inspection target of adjacent two moving images.

[0048] As an optional implementation, in the second aspect of the embodiment of the present application, the judging whether to generate the early warning information according to the change path comprises:

[0049] The change path is compared with the specification check target transformation reference, and a matching degree between the change path and the specification check target transformation reference is calculated, and when the matching degree is lower than a preset value, the early warning information is generated.

[0050] The change path is compared with the specification check target transformation reference, and a matching degree between the change path and the specification check target transformation reference is calculated, and when the matching degree is lower than a preset value, the early warning information is generated.

[0051] The third aspect of the embodiment of the present application discloses an electronic device, comprising: a memory storing executable program codes; a processor coupled with the memory; the processor invokes the executable program codes stored in the memory, and is used for executing the specification check method based on artificial intelligence disclosed in the first aspect of the embodiment of the present application.

[0052] The fourth aspect of the embodiment of the present application discloses a computer readable storage medium storing a computer program, wherein the computer program makes a computer execute the specification check method based on artificial intelligence disclosed in the first aspect of the embodiment of the present application.

[0053] Compared with the prior art, the embodiment of the present application has the following beneficial effects:

[0054] In the embodiment of the present application, first, a target detection instruction is received and responded, one or more target detection objects are selected based on the target detection instruction, in order to better detect the operation articles or operation actions corresponding to the target detection objects, the target detection regions corresponding to the target detection objects are acquired in advance, and the target detection objects and the target detection regions are bound to form specification check targets; image data including static picture data and dynamic video tracking data is collected through an image collection device, the image data is preprocessed to obtain data with higher accuracy, which is helpful to better extract feature points and then accurately identify targets; on the basis of artificial intelligence, the embodiment greatly improves the monitoring feasibility and flexibility of student experiment processes, helps to accurately judge the operability of student experiments, greatly saves manpower, and improves efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0056] Figure 1is a flow diagram of a specification checking method based on artificial intelligence disclosed by an embodiment of the present application.

[0057] Figure 2 is a structural diagram of a specification checking device based on artificial intelligence provided by an embodiment of the present application.

[0058] Figure 3 is a structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0060] It should be noted that the terms "first", "second", "third", "fourth" and the like in the specification and claims of the present application are used to distinguish different objects, rather than to describe a specific order. The terms "include" and "have" and any variations thereof in the embodiments of the present application are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product or device.

[0061] The embodiments of the present application disclose a specification checking method and device based on artificial intelligence, an electronic device and a storage medium. In the embodiments, a target detection instruction is first received and responded, one or more target detection objects are selected based on the target detection instruction, in order to better detect operation articles or operation actions corresponding to the target detection objects, target detection regions corresponding to the target detection objects are acquired in advance, and the target detection objects and the target detection regions are bound to form specification checking targets; image data including static picture data and dynamic video tracking data is collected through an image collection device, the image data is preprocessed to obtain data with higher accuracy, which is helpful to better extract feature points and then accurately identify targets, generate a dynamic change path of the targets, and determine whether early warning is needed according to the change path; the embodiments greatly improve the monitoring feasibility and flexibility of student experiment processes on the basis of artificial intelligence, help to accurately judge the operability of student experiments, greatly save manpower, and improve efficiency.

[0062] Embodiment one

[0063] Please refer to Figure 1 , Figure 1is a flowchart of the specification checking method based on artificial intelligence disclosed by the embodiments of the present application. Among them, the execution subject of the method described in the embodiments of the present application is an execution subject composed of software or / and hardware, which can receive relevant information through wired or / and wireless mode, and can send certain instructions. Of course, it can also have certain processing function and storage function. The execution subject can control multiple devices, such as remote physical servers or cloud servers and related software, or it can be a local host or server and related software that performs related operations on a device placed somewhere. In some scenarios, it can also control multiple storage devices, which can be placed in the same place or different places. As shown in Figure 1 The specification checking method based on artificial intelligence includes the following steps:

[0064] 101, in response to the target detection instruction, selecting a target detection object, obtaining a target detection area corresponding to the target detection object, and associating the target detection area and the target detection object to obtain a specification checking target.

[0065] In the embodiments, the target detection instruction can be user input, preset rules or other external systems, etc. According to the instruction, one or more target detection objects can be selected. Usually, the target detection object is a person, and when applied in a student laboratory scene, the target detection object is usually a student. At the same time, the target detection area of the target detection object is also obtained. The target detection area is closely related to the target detection object, for example, the target detection area can be a body part in the target detection object, or an instrument operated by the target detection object. The target detection area can be one or more. For example, when the target detection area is an experimental instrument, there can be multiple experimental instruments of the same type and placed in different positions, but one of them is for the target detection object to operate, so the target detection area is bound to the target detection object.

[0066] 102, collect image data of the specification checking target, the image data includes picture data and video tracking data, pre-process the image data, extract feature points from the pre-processed image data, and obtain the identification result of the specification checking target according to the feature points, the identification result includes detection standard, position and type.

[0067] In the embodiments, computer vision technology is used for image acquisition and preprocessing, which is used for image acquisition and preprocessing to ensure that the image quality meets the requirements of subsequent analysis; the training and optimization of deep learning model, the performance of the model directly affects the recognition accuracy of the dressing feature; the judgment of the detection result, according to the preset safety dressing rule, accurately evaluate the dressing compliance; the formulation and implementation of safety dressing rule, formulate reasonable safety dressing rule, and ensure its effective implementation.

[0068] The image data of the inspection target is collected, which includes picture data and video tracking data, providing rich information about the appearance and motion of the target object. The image data is preprocessed to remove noise, enhance features, adjust contrast, etc., to improve the accuracy of subsequent feature extraction. Specifically, the image data in this embodiment is preprocessed, including: denoising the image data based on bilateral filtering; adjusting the contrast of the denoised image data; adjusting the size of the image data, and identifying the interference area in the image data, and cropping the interference area.

[0069] Bilateral filtering is a nonlinear, edge-preserving smoothing filtering method. It considers the weight of both the spatial domain (i.e. the physical distance between pixels) and the pixel value domain (i.e. the difference between pixel values). This method can effectively remove noise while preserving the edge information of the image. When operating, set the parameters of bilateral filtering, including the Gaussian standard deviation of the spatial domain (σ_s) and the Gaussian standard deviation of the pixel value domain (σ_r), apply the bilateral filtering algorithm to the image data to obtain the denoised image. Contrast adjustment is a common step in image processing, which aims to enhance the brightness difference between different regions in the image, making the image clearer and easier to identify. In operation, the global or local contrast of the image is calculated, and the brightness value of the image is adjusted according to the contrast calculation result to increase or decrease the contrast. Histogram equalization, adaptive contrast enhancement, etc. are used to realize contrast adjustment. In image data, interference areas may include irrelevant backgrounds, obstructions, noise, etc., which may affect the results of subsequent image analysis. Therefore, it is necessary to identify and crop these interference areas. Image segmentation, edge detection, color space analysis, etc. are used to identify interference areas. According to the identification result, the position and range of the interference area are determined. The image is cropped to remove the interference area and retain the target area.

[0070] Further, the denoising method in the embodiment can also be implemented through a convolutional neural network or a generative adversarial network, etc. to learn more complex noise patterns through a large amount of data training, to achieve higher quality denoising. A denoising network is constructed, and the image with noise and the clean image are used as training data to train the denoising network. After training, the denoising network is used to denoise the image data. For contrast enhancement and color correction processing, multiple regions in the image can be identified based on image segmentation technology, and different contrast enhancement and color correction strategies are applied to each region.

[0071] Then feature points are extracted from the preprocessed image data, including: inputting the preprocessed image data into a preset YOLO deep learning model to extract feature points of the image data.

[0072] Further comprising: inputting the pre-processed image data into a preset hand key point detection model and an eye key point detection model to extract feature points.

[0073] As an application scenario, in a student laboratory, the main task is to capture experimental data and machine status, and monitor the safety of the placement and use of equipment in the laboratory. In image acquisition, the installation position of the camera ensures coverage of the entire area to be monitored, and the image of each piece of experimental equipment can be captured, thereby achieving comprehensive monitoring of the experimental process. In this application scenario, the preprocessing step is used to improve the accuracy and efficiency of subsequent target detection, including using filtering techniques to remove noise in the image, ensuring the clarity and quality of the image. At the same time, in order to adapt to the input size requirements of the model, the image data is scaled and adjusted in size to reduce computational complexity and improve processing speed. In the recognition stage, a deep learning model YOLO is used to accurately detect experimental instruments such as test tubes, beakers, and microscopes. Through this deep learning model, different types of instruments can be automatically identified, and their categories and position coordinates can be extracted. After instrument recognition, the position information of each instrument and its key components is obtained, which is the basis for subsequent normative judgment. According to the preset instrument assembly specifications, the method will comprehensively judge the identified experimental instruments. The judgment content includes whether the instruments are placed in the specified position, whether the components of the instruments are connected stably, and whether the instruments have deviations in the horizontal and vertical directions. These judgments not only help to ensure the normativity of experimental operations, but also improve the safety of experiments, avoiding accidents caused by improper placement of instruments.

[0074] Further, it can further include step 103: based on the visual tracking data, when the normative inspection target is detected to move, extracting the moving image of the normative inspection target at a preset interval to generate a change path, and judging whether to generate a warning information according to the change path.

[0075] In the above application scenario, target tracking and recognition technology is introduced, that is, the target is tracked through video and tracking algorithm, which can track the key points of instruments or operators in the laboratory in real time. This technology can accurately identify in the video stream, ensuring that the state of the instrument is continuously monitored throughout the experiment, and potential problems are discovered and corrected in a timely manner. ByteTrack technology can maintain accurate tracking of targets in complex experimental environments through efficient data association algorithms, effectively reducing recognition errors caused by factors such as changes in lighting and occlusion.

[0076] In another scenario, the application and the laboratory jointly detect the key points of the students' body parts. Specifically, the YOLO deep learning algorithm is also applicable to fast and accurate target detection. In addition to detecting experimental instruments, the resnet convolutional neural network is used to detect hand key points, and the key points can be set in advance at different positions on the face, such as a total of 21 key points, which has high processing accuracy and speed, and is suitable for real-time application. For face key point detection, retinaface can be used to realize 5-point face key point detection, which has high processing accuracy and speed, and is suitable for real-time application.

[0077] In the application scenario of body part detection, the center point coordinates of the narrow-mouthed bottle and the graduated cylinder are calculated, the geometric relationship is used to judge whether the bottle mouth of the narrow-mouthed bottle is located directly above the graduated cylinder, the hand key points (such as finger tips) are detected whether they are in the boundary box of the narrow-mouthed bottle, and the hand is ensured to be correctly contacted, the relative position between the eye key points (such as the corners of the eyes) and the water level of the graduated cylinder is calculated, the distance is judged, and the student is ensured to visually observe the water level.

[0078] In this step, the identification result of the standard inspection target is obtained according to the feature points, including: dividing the image data into multiple recognition areas, obtaining the code corresponding to each recognition area according to the feature points; obtain the connection information of the guide line between each two recognition areas to create a directed graph adjacency list, and obtain the preset current flow direction of each recognition area, add nodes in the directed graph adjacency list based on the preset current flow direction; judge whether there is a current loop in the image data based on the directed graph adjacency list after adding nodes, to obtain the circuit connection information of the image data, and generate the corresponding identification result according to the circuit connection information.

[0079] The above steps are applied in another application scenario, namely electrical experiment. The traditional electrical experiment mainly relies on manual judgment when checking the circuit connection. This has many problems, such as in the experiment examination, a large number of invigilators are needed to check the circuit connection of each student one by one, which greatly increases the workload of manual scoring, and manual checking may cause misjudgment due to subjective factors or fatigue, affecting the fairness of the examination. At the same time, in daily teaching, when students perform experiments, teachers may have difficulty in fully considering the operation of each student due to the large number of students, and cannot timely find the errors in the circuit connection of the students, which is not conducive to the effective development of teaching. In addition, the traditional manual detection method is low in efficiency and cannot quickly and accurately judge the circuit conditions of a large number of students. Moreover, manual detection can only be carried out at a specific time and place, and cannot realize real-time monitoring and feedback. With the continuous development of science and technology, although some auxiliary detection means have appeared, there are still many deficiencies. For example, some simple electronic detection devices have single functions and can only detect some basic circuit parameters, and need to be connected to the circuit used by students for experiments, which affects the normal experiment process of students, and cannot comprehensively and accurately judge the correctness of the circuit connection. The embodiment is applied in this scene, which can use machine vision recognition technology to recognize the equipment and analyze the connection state of the terminal and the wire in real time. First, high-definition cameras are used to obtain the connection images of the equipment, terminal and wire in the middle school physics electrical experiment. These cameras are carefully arranged in the experimental area to ensure that the images of each key part can be captured comprehensively and clearly. Second, advanced image recognition algorithms are used to process the obtained images. By extracting and analyzing the shape, color, position and other characteristics of the equipment, terminal and wire in the image, their position and state in the image are accurately recognized. Then, with the help of powerful graph algorithms, the connection relationship of the recognized equipment, terminal and wire is calculated and judged. This graph algorithm can accurately analyze whether the current circuit connection meets the experimental requirements according to the pre-set circuit connection rules and patterns. In this application scenario, taking the detection of the battery, switch, bulb and ammeter forming a series circuit as an example, a terminal device for collecting video is placed on the student's experiment table, which has a camera to collect the circuit connection of the student in real time from the top of the experiment table. The video stream is transmitted to an artificial intelligence analysis server through the network, which is used for artificial intelligence analysis and processing of the circuit connection state in the video stream to judge the circuit connection. First, the equipment in the video is recognized, including but not limited to battery, ammeter, switch, small bulb, etc. Through the target detection algorithm, the position, size and category of the equipment in the video are recognized, and the positive and negative terminal posts of the equipment in the video are also detected.By identifying the wire head position and size of the wire in the video, the wire used ensures that the colors are different, such as (red, green, blue, black, etc.), and each wire head of each color is a category. Each terminal is numbered to facilitate subsequent algorithm description. Model each terminal, and according to the intersection of the wire head rectangle frame detected by artificial intelligence and the terminal rectangle frame, determine whether the two terminals are connected by a wire, and create a directed graph adjacency table. According to the internal current flow direction allowed by the equipment, add nodes to the directed graph adjacency table. For example, the internal current flow direction of the ammeter is only from the positive terminal to the negative terminal, the switch and the bulb do not limit the current flow direction, and the current flow direction of the battery is from the negative terminal to the positive terminal. Starting from the positive terminal of the battery, the depth-first traversal of the graph is performed, and the existence of a loop in the graph is detected, which indicates that the circuit is connected in a loop. There is only one loop in the graph, which indicates that the circuit is a series circuit. The connection of the series circuit can be known by the order of the depth-first traversal, and the connection is converted into a graph node name: battery positive terminal- ammeter positive terminal- ammeter negative terminal- bulb positive terminal- bulb negative terminal- switch positive terminal- switch negative terminal- battery negative terminal.

[0080] In this step, when the standard inspection target is detected to move based on the visual tracking data, the moving images of the standard inspection target are extracted at a preset interval to generate a change path, including: when the standard inspection target is detected to move based on the visual tracking data, the video tracking data is cut at a preset interval to divide into several moving images; identify and mark the standard inspection target in each moving image to obtain the position information of the standard inspection target in each video segment; and generate a change path of the standard inspection target based on the position information of the standard inspection target of adjacent two moving images. By generating a change path, for example, the opening action of a bottle cap, from opening to completely opening, to the placement position of the bottle cap, for example, the bottle cap should be placed at position A to be standard, but the student places it at position B, which may not be safe. Whether it meets the standard can be judged by the change path.

[0081] Further, whether to generate a warning information is judged according to the change path, including: obtaining the change type of the standard inspection target according to the change path, selecting a corresponding standard inspection target transformation reference according to the change type; comparing the change path with the standard inspection target transformation reference, calculating the matching degree between the change path and the standard inspection target transformation reference, and generating a warning information when the matching degree is lower than a preset value.

[0082] Embodiment two

[0083] Please refer to Figure 2 , Figure 2is a structural schematic view of a specification checking device based on artificial intelligence disclosed by the embodiment of the present application. As shown in the figure, Figure 2 The specification checking device based on artificial intelligence can include: an instruction response module 201, an image acquisition module 202, wherein the instruction response module 201 is used for selecting a target detection object in response to a target detection instruction, acquiring a target detection area corresponding to the target detection object, and associating the target detection area and the target detection object to obtain a specification checking target; the image acquisition module 202 is used for acquiring image data of the specification checking target, the image data including picture data and video tracking data, pre-processing the image data, extracting feature points from the pre-processed image data, and obtaining an identification result of the specification checking target according to the feature points, the identification result including a detection standard, a position and a type.

[0084] Further, it can also include a target comparison module 203: for detecting the movement of the specification checking target based on the visual tracking data, extracting the moving image of the specification checking target at a preset interval to generate a change path, and judging whether to generate a warning information according to the change path.

[0085] In the image acquisition module 202, the image data is pre-processed, including: denoising the image data based on bilateral filtering; adjusting the contrast of the denoised image data; adjusting the size of the image data, and identifying the interference area in the image data, and cropping the interference area.

[0086] Further, the feature points are extracted from the pre-processed image data, including: inputting the pre-processed image data into a preset YOLO deep learning model to extract the feature points of the image data. As a preferred, it also includes: inputting the pre-processed image data into a preset hand key point detection model and an eye key point detection model to extract the feature points.

[0087] According to the feature points to obtain the identification result of the specification checking target, including: dividing the image data into a plurality of identification areas, obtaining the corresponding code of each identification area according to the feature points; obtaining the connection information of the guide line between each two identification areas to create a directed graph adjacency list, and obtaining the preset current flow direction of each identification area, adding nodes in the directed graph adjacency list based on the preset current flow direction; judging whether there is a current loop in the image data based on the directed graph adjacency list after adding nodes, to obtain the circuit connection information of the image data, and generating the corresponding identification result according to the circuit connection information.

[0088] In the target comparison module 203, when the standard inspection target movement is detected based on the visual tracking data, the moving images of the standard inspection target are extracted at a preset interval to generate a change path, including: when the standard inspection target movement is detected based on the visual tracking data, the video tracking data is cut at a preset interval to divide into several moving images; the standard inspection target in each moving image is identified and marked to obtain the standard inspection target position information in each video segment; and the change path of the standard inspection target is generated based on the standard inspection target position information of adjacent two moving images.

[0089] Further, whether to generate a warning information is judged according to the change path, including: the change type of the standard inspection target is obtained according to the change path, the corresponding standard inspection target transformation reference is selected according to the change type; the change path and the standard inspection target transformation reference are compared to calculate the matching degree between the change path and the standard inspection target transformation reference, and when the matching degree is lower than a preset value, a warning information is generated.

[0090] Embodiment three

[0091] Please refer to Figure 3 , Figure 3 is a structural schematic diagram of an electronic device disclosed in an embodiment of the present application. The electronic device can be a computer, a server, etc., and of course, under certain circumstances, it can also be a smart device such as a mobile phone, a tablet computer, and a monitoring terminal, and an image acquisition device with processing function. As shown in the figure, Figure 3 The electronic device can include:

[0092] a memory 301 storing executable program codes;

[0093] a processor 302 coupled with the memory 301;

[0094] The processor 302 calls the executable program codes stored in the memory 301 to execute part or all of the steps of the standard inspection method based on artificial intelligence in embodiment one.

[0095] The present application discloses a computer readable storage medium storing a computer program, wherein the computer program causes a computer to execute part or all of the steps of the standard inspection method based on artificial intelligence in embodiment one.

[0096] The present application also discloses a computer program product, wherein when the computer program product runs on a computer, it causes the computer to execute part or all of the steps of the standard inspection method based on artificial intelligence in embodiment one.

[0097] The application also discloses an application publishing platform, which is used for publishing the computer program product.

[0098] In various embodiments of the present application, it should be understood that the size of the serial number of the processes does not mean the inevitable sequence of execution, and the execution sequence of the processes should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0099] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0100] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0101] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-accessible memory. Based on this understanding, the technical solutions of the present application or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product stored in a memory includes a plurality of parts or all steps of the method for enabling a computer device (which can be a personal computer, a server or a network device, etc., and specifically can be a processor in the computer device) to execute the method described in each embodiment of the present application.

[0102] In the embodiments provided by the present application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that the determination of B according to A does not mean that B is determined only according to A, but also can be determined according to A and / or other information.

[0103] Those skilled in the art can understand that part or all of the steps in the various methods of the embodiments can be completed by instructing the relevant hardware by a program, and the program can be stored in a computer readable storage medium, including Read-Only Memory (ROM), Random Access Memory (RAM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), One-time Programmable Read-Only Memory (OTPROM), Electrically-Erasable Programmable Read-Only Memory (EEPROM), Compact Disc Read-Only Memory (CD-ROM) or other optical disk memories, magnetic disk memories, magnetic tape memories, or any other computer readable medium capable of carrying or storing data.

[0104] The above discloses the specification checking method, device, electronic equipment and storage medium based on artificial intelligence according to the embodiments of the present application in detail. The principle and implementation mode of the present application are described by applying specific examples in this paper. The above embodiment is only used to help understand the method and core idea of the present application. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. An AI-based specification checking method, characterized by, include: In response to the target detection command, a target detection object is selected, the target detection area corresponding to the target detection object is obtained, and the target detection area and the target detection object are associated to obtain the standard inspection target; Image data of the standardized inspection target is collected, including image data and video tracking data. The image data is preprocessed, and feature points are extracted from the preprocessed image data to obtain the identification result of the standardized inspection target based on the feature points. The identification result includes detection criteria, location, and type. The step of obtaining the identification result of the standardized inspection target based on the feature points includes: The image data is divided into multiple recognition regions, and the code corresponding to each recognition region is obtained based on the feature points. Obtain the wire connection information between every two identified regions to create a directed graph adjacency list, and obtain the preset current flow direction for each identified region. Based on the preset current flow direction, add nodes to the directed graph adjacency list. Based on the adjacency list of the directed graph after adding nodes, it is determined whether there is a current loop in the image data, so as to obtain the circuit connection information of the image data, and generate the corresponding recognition result according to the circuit connection information. Also includes: When the movement of the target under inspection is detected based on visual tracking data, the movement image of the target under inspection is extracted at preset intervals to generate a change path, and a warning message is generated based on the change path.

2. The standardization inspection method according to claim 1, characterized in that, The preprocessing of the image data includes: The image data is denoised using bilateral filtering; Adjust the contrast of the denoised image data; Adjust the size of the image data, identify interference areas in the image data, and crop the interference areas.

3. The standardization inspection method according to claim 1, characterized in that, Feature points are extracted from the preprocessed image data, including: The preprocessed image data is input into a preset YOLO deep learning model to extract feature points from the image data.

4. The standardization inspection method according to claim 3, characterized in that, Extracting feature points from preprocessed image data also includes: The preprocessed image data is input into the preset hand keypoint detection model and eye keypoint detection model to extract feature points.

5. The standardization inspection method according to claim 1, characterized in that, When the movement of the target under inspection is detected based on visual tracking data, the movement image of the target under inspection is extracted at preset intervals to generate a change path, including: When the visual tracking data detects movement of the target being inspected, the video tracking data is segmented at preset intervals to divide it into several moving images. Identify and mark the standardized inspection targets in each moving image segment, and obtain the location information of the standardized inspection targets in each video segment; The change path of the target is generated based on the target location information of two adjacent moving images.

6. The standardization inspection method according to claim 1, characterized in that, The step of determining whether to generate a warning message based on the change path includes: Based on the change path, the change type of the standard inspection target is obtained, and the corresponding standard inspection target transformation benchmark is selected based on the change type. The change path is compared with the standard inspection target transformation benchmark to calculate the matching degree between the change path and the standard inspection target transformation benchmark. When the matching degree is lower than a preset value, an early warning message is generated.

7. A normative inspection device based on artificial intelligence, characterized in that, include: Command response module: Used to select target detection objects in response to target detection commands, obtain the target detection area corresponding to the target detection object, and associate the target detection area and the target detection object to obtain the standardized inspection target; Image acquisition module: used to acquire image data of the standard inspection target, the image data including picture data and video tracking data, preprocess the image data, extract feature points from the preprocessed image data, and obtain the recognition result of the standard inspection target based on the feature points, the recognition result including detection criteria, location and type; The step of obtaining the identification result of the standardized inspection target based on the feature points includes: The image data is divided into multiple recognition regions, and the code corresponding to each recognition region is obtained based on the feature points. Obtain the wire connection information between every two identified regions to create a directed graph adjacency list, and obtain the preset current flow direction for each identified region. Based on the preset current flow direction, add nodes to the directed graph adjacency list. Based on the adjacency list of the directed graph after adding nodes, it is determined whether there is a current loop in the image data, so as to obtain the circuit connection information of the image data, and generate the corresponding recognition result according to the circuit connection information. Also includes: When the movement of the target under inspection is detected based on visual tracking data, the movement image of the target under inspection is extracted at preset intervals to generate a change path, and a warning message is generated based on the change path.

8. An electronic device, characterized in that, include: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the artificial intelligence-based normative inspection method according to any one of claims 1 to 5.

Images