Power tool warehouse-in and warehouse-out management method and system based on machine vision and medium

Abstract

The application discloses a power tool warehouse-in and warehouse-out management method and system based on machine vision and a medium, and the method comprises the following steps: acquiring warehouse-in and warehouse-out video images of power tools; inputting the warehouse-in and warehouse-out video images of the power tools into a LightTrack target tracking and identification network, performing multi-target identification on each frame of image in the warehouse-in and warehouse-out video images of the power tools, outputting single-frame sequence images and target trajectory graphs of target labeling, and storing the single-frame sequence images and the target trajectory graphs in a database; performing face identity recognition by using a key block sparse representation face recognition method according to the single-frame sequence images of target labeling, and performing power tool color and size identification; performing target personnel and power tool warehouse-in and warehouse-out judgment and power tool and user consistency identification according to the target trajectory graphs; and registering power tool use information in a power tool management system. The application improves the intelligent management level and the warehouse-in and warehouse-out management efficiency of the power tools.

Description

Technical Field

[0001] This invention relates to the field of power safety and warehouse management, specifically to a method and system for managing the entry and exit of power tools based on machine vision. Background Technology

[0002] Currently, the management of power safety tools entering and leaving the warehouse mostly relies on facial recognition or fingerprint identification for staff, and RFID radio frequency chips for tool registration. This combination of technologies achieves warehouse management. However, this method requires active worker participation. Due to RFID chip failure or worker negligence, tools are often missed or misregistered, resulting in insufficient intelligent management and low efficiency in warehouse entry and exit.

[0003] In view of the above, this application is hereby submitted. Summary of the Invention

[0004] The purpose of this invention is to provide a machine vision-based method, system, and medium for managing the entry and exit of electrical tools. This invention aims to improve the intelligent management level of tools, reduce management costs, and increase the efficiency of managing the entry and exit of electrical tools. It employs machine vision technology, using video recognition to determine the entry and exit of personnel and identify their identities. It identifies the model and entry / exit status of electrical tools and uses intelligent algorithms for trajectory analysis to intelligently identify the use of electrical tools by personnel. The entry and exit of tools are then registered in the tool management system, thereby improving the level of intelligent, unmanned management and the efficiency of entry and exit management.

[0005] This invention is achieved through the following technical solution:

[0006] In a first aspect, the present invention provides a machine vision-based method for managing the entry and exit of electrical tools, the method comprising:

[0007] Acquire video images of electrical tools entering and leaving the warehouse;

[0008] The video images of electrical tools entering and leaving the warehouse are input into the LightTrack target tracking and recognition network. Multi-target recognition is performed on each frame of the video images of electrical tools entering and leaving the warehouse. The single-frame sequence images with target annotation and the target trajectory map are output and stored in the database.

[0009] Based on the single-frame sequence of images labeled with the target, the key block sparse representation face recognition method is used to identify face identity and to identify the color and size of electrical tools.

[0010] Based on the target trajectory map, determine the entry and exit of target personnel and electrical tools, and identify whether the electrical tools and the personnel who issued them are consistent; and register the information on the issuance of electrical tools in the electrical tool management system.

[0011] Furthermore, multi-target recognition is performed on each frame of the video images of electrical tools entering and leaving the warehouse. The multi-target objects include personnel targets and electrical tool targets. Electrical tool targets include safety helmets, insulated shoes, insulated gloves, safety belts, insulated rods, voltage testers, grounding wires, etc.

[0012] Furthermore, the steps for facial identity recognition using the key block sparse representation method are as follows:

[0013] Match single-sample faces with facial feature templates in the database to identify four key facial regions: left eye, right eye, nose, and mouth.

[0014] The above four regions are represented using sparse representation: k = 1, 2, 3, 4, corresponding to the left eye, right eye, nose, and mouth respectively; A k For sparse dictionaries, y is the sparsity coefficient. k For feature region images;

[0015] Determine the sparse feature vector α of the key facial region of the i-th worker in the database. ik The feature vectors of the four key facial regions are combined to form a matrix γ. i =[α i,1 α i,2 α i,3 α i,4 This information is stored in the database for easy comparison and retrieval later.

[0016] A facial database is collected using a camera, and the SDM (Supervised Descent Method) algorithm is used to detect faces and locate feature points.

[0017] Four key facial regions were extracted through feature template matching; a sparse dictionary A was established. i Solve for the feature vectors β1, β2, β3, and β4 of the four feature regions of the face, and form the feature matrix β = [β1 β2 β3 β4].

[0018] make Solve That is, k is e i The minimum value of i, e i To reflect the similarity between the collected faces and the faces in the database; e iThe minimum value is represented by E. E is compared with a set threshold T: if E is greater than T, the person is identified as a non-staff member and is prohibited from entering; otherwise, the staff member is identified as the i-th staff member.

[0019] Furthermore, the sparse dictionary is constructed using the Orthogonal Matching Pursuit (OMP) method. k .

[0020] Furthermore, color and size identification of electrical tools is performed, including:

[0021] Represent the target region using the RGB color model and calculate the average RGB value of the target region; based on the average RGB value s of the target region... i Determine whether the color of the electrical appliance matches the target color a. j Whether they are consistent; the formula for judgment is: Among them, T c The tolerance range for color change is defined as follows: if the above formula is satisfied, the color of the electrical appliance is considered to be consistent with the target color.

[0022] The depth information H of the target is obtained by a depth camera. The length of the power tool in the image is l. The actual length L of the power tool is calculated by the focal length f of the depth camera. The calculation formula is: L = lH / f.

[0023] Furthermore, based on the target trajectory map, the entry and exit of target personnel and electrical tools are determined, including:

[0024] If the following formula is satisfied, the item is considered to be put into storage. The formula for determining whether an item is put into storage is:

[0025]

[0026] If the following formula is satisfied, the goods are deemed ready for shipment. The shipment determination formula is:

[0027]

[0028]

[0029] In the formula, (x k,i ,y k,i L represents the two-dimensional trajectory of target k identified by the LightTrack target tracking and recognition network. k The coordinates of the two-dimensional trajectory point at the i-th time point, recorded in chronological order of the trajectory time; (x k,i+1 ,y k,i+1 θ represents the coordinates of the two-dimensional trajectory point at the next time point after the i-th time point; ki For personnel k in (x k,i ,y k,i The direction angle of motion of the position; (m)i n i θ represents the coordinates of the standard trajectory route L of the two-dimensional image of the target entering and leaving the database; out θ in These are the trajectory direction angles for the outbound and inbound directions, respectively; T1 and T2 are threshold values.

[0030] If the inbound and outbound judgment formulas are not met, then it is considered another case.

[0031] Furthermore, whether the electrical tools and the personnel who received them are consistent should be identified, including:

[0032] By tracking the movement of staff, video frames showing staff entering and leaving the warehouse can be determined.

[0033] In the same video frame, bounding boxes of workers and electrical tools are generated based on the LightTrack target tracking and recognition network. The bounding boxes of workers and tools are used to determine whether the electrical tools were picked up by a certain worker.

[0034] Furthermore, the information on the requisition of electrical tools includes the personnel who requisitioned (or put into storage), the time of requisition, and the model of the tool.

[0035] Secondly, the present invention also provides a machine vision-based power tool inbound and outbound management system, which uses the aforementioned machine vision-based power tool inbound and outbound management method; the system includes:

[0036] The acquisition unit is used to acquire video images of electrical tools entering and leaving the warehouse;

[0037] The multi-target recognition unit is used to input video images of power tools entering and leaving the warehouse into the LightTrack target tracking and recognition network, perform multi-target recognition on each frame of the video images of power tools entering and leaving the warehouse, output single-frame sequence images with target annotation and target trajectory map, and store them in the database.

[0038] The face recognition unit is used to perform face recognition based on a single-frame sequence of images with target annotations, using the key block sparse representation face recognition method.

[0039] The electrical tool identification unit is used to identify the color and size of electrical tools based on a single-frame sequence of images labeled with the target.

[0040] The entry and exit judgment unit is used to judge the entry and exit of target personnel and electrical tools and to identify whether the electrical tools and the personnel who received them are consistent, based on the target trajectory map.

[0041] The requisition information registration unit is used to register the requisition information of electrical tools and equipment in the electrical tool and equipment management system.

[0042] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described machine vision-based method for managing the entry and exit of electrical tools.

[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0044] This invention relates to a machine vision-based method, system, and medium for managing the entry and exit of electrical tools. To improve the intelligent management level of tools, reduce management costs, and increase the efficiency of tool entry and exit management, this invention employs machine vision technology. It uses video recognition technology to determine the entry and exit of personnel and identify their identities. The invention identifies the model and entry / exit status of electrical tools and uses intelligent algorithms for trajectory analysis to intelligently identify the use of electrical tools by personnel. The entry and exit of tools are then registered in the tool management system, thereby improving the level of intelligent, unmanned management and the efficiency of tool entry and exit management. Attached Figure Description

[0045] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0046] Figure 1 This is a flowchart of the machine vision-based method for managing the entry and exit of electrical tools and equipment according to the present invention.

[0047] Figure 2 This is a detailed flowchart of the machine vision-based method for managing the entry and exit of electrical tools and equipment according to the present invention.

[0048] Figure 3 This is a single-frame target recognition result image of the LightTrack target output of the present invention;

[0049] Figure 4 This is a target motion trajectory diagram of the LightTrack method of the present invention;

[0050] Figure 5 This is an illustration of the facial recognition process in this invention;

[0051] Figure 6 This is a structural block diagram of the power tool inbound and outbound management system based on machine vision according to the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0053] Currently, the management of power safety tools entering and leaving the warehouse mostly relies on facial recognition or fingerprint identification for staff, and RFID radio frequency chips for tool registration. This combination of technologies achieves warehouse management. However, this method requires active worker participation. Due to RFID chip failure or worker negligence, tools are often missed or misregistered, resulting in insufficient intelligent management and low efficiency in warehouse entry and exit.

[0054] Therefore, this invention designs a machine vision-based method, system, and medium for managing the entry and exit of electrical tools. To improve the intelligent management level of tools, reduce management costs, and increase the efficiency of managing the entry and exit of electrical tools, this invention employs machine vision technology. It uses video recognition technology to determine the entry and exit of personnel and identify their identities. It identifies the model and entry / exit status of electrical tools and uses intelligent algorithms for trajectory analysis to intelligently identify the use of electrical tools by personnel. The entry and exit of tools are then registered in the tool management system, thereby improving the intelligent, unmanned management level and efficiency of the entry and exit of tools.

[0055] Example 1

[0056] like Figure 1 and Figure 2 As shown, the present invention provides a machine vision-based method for managing the entry and exit of electrical tools and equipment. The method includes:

[0057] Step 1: Obtain video images of electrical tools entering and leaving the warehouse;

[0058] Step 2: Input the video images of the entry and exit of electrical tools into the LightTrack target tracking and recognition network, perform multi-target recognition on each frame of the video images of the entry and exit of electrical tools, output single-frame sequence images of target annotation and target trajectory map, and store them in the database;

[0059] The LightTrack target tracking and recognition network is an existing technology. Specifically, LightTrack uses Neural Architecture Search (NAS) to achieve a lightweight and efficient target tracker with less computation and fewer parameters, making it easy to deploy on resource-constrained mobile chips and achieve fast operation.

[0060] The multi-target objects identified by the LightTrack target tracking and recognition network include personnel targets and electrical tool targets. Electrical tool targets include safety helmets, insulated shoes, insulated gloves, safety belts, insulated rods, voltage testers, grounding wires, etc.

[0061] The LightTrack target tracking and recognition network takes video images of electrical tools entering and leaving the warehouse as input, performs multi-target recognition on each frame of the image, and thus realizes the tracking of the motion trajectory of moving targets.

[0062] Based on the needs of subsequent analysis, the output of the LightTrack target tracking and recognition network is divided into two parts: one part is a single-frame sequence for target identification and annotation, which is used for face ID identification and tool model recognition. Figure 3 The system outputs a single-frame target recognition result for LightTrack targets; another part of the output is a target motion trajectory map, used for determining the entry and exit of equipment. Figure 4 This is a trajectory diagram of the LightTrack target motion.

[0063] Step 3: Based on the single-frame sequence of images labeled with the target, the key block sparse representation face recognition method is used to identify the face identity and to identify the color and size of the power tools.

[0064] Specifically, the LightTrack target tracking and recognition network only achieves face tracking and recognition, but does not enable employee identification via face. Further facial recognition for personnel entering and exiting the premises is required. Employee identification uses a single-sample face recognition network (i.e., only one standard face image of each employee is in the database) to reduce computational load.

[0065] First, this invention employs a key block sparse representation face recognition method for facial identification, with the following specific steps:

[0066] Step 31: Match the single-sample face with the facial feature template in the database to determine four key facial regions: left eye, right eye, nose, and mouth.

[0067] Step 32: Apply sparse representation to the above four regions respectively: k = 1, 2, 3, 4, corresponding to the left eye, right eye, nose, and mouth respectively; A k For sparse dictionaries, y is the sparsity coefficient. k For feature region images;

[0068] The sparse dictionary is constructed using the Orthogonal Matching Pursuit (OMP) method. k .

[0069] Step 33: Determine the sparse feature vector α of the key facial region of the i-th worker in the database. ik The feature vectors of the four key facial regions are combined to form a matrix γ. i =[α i,1 α i,2 αi,3 α i,4 This information is stored in the database for easy comparison and retrieval later.

[0070] Step 34: Collect a face database using a camera, and use the SDM algorithm (Supervised Descent Method) to detect faces and locate feature points;

[0071] Step 35: Extract four key facial regions through feature template matching; use the established sparse dictionary A i Solve for the feature vectors β1, β2, β3, and β4 of the four feature regions of the face, and form the feature matrix β = [β1 β2 β3 β4].

[0072] Step 36: Let Solve That is, k is e i The minimum value of i, e i To reflect the similarity between the collected faces and the faces in the database; e i The minimum value is represented by E. E is compared with a set threshold T: if E is greater than T, the person is identified as a non-staff member and is prohibited from entering; otherwise, the staff member is identified as the i-th staff member.

[0073] like Figure 5 As shown, (a) is the captured face, (b) is the face feature template, and (c) is the four identified key facial regions.

[0074] Second, perform color and size identification of electrical tools, including:

[0075] (1) Color information recognition

[0076] Tools and equipment are distinguished by color. For example, red and yellow safety helmets are suitable for different workplaces and groups of people. Color recognition involves representing the color of a target area in the RGB color space, representing the target area using the RGB color model, and calculating the average RGB value s of the target area. i = [s1 s2 s3]; based on the average RGB value of the target region s i =[s1 s2 s3], determine whether the color of the electrical tool matches the target color a. j Whether [a1 a2 a3] are consistent can be determined using the following formula in the color space: Among them, T c The tolerance range for color change is defined as follows: if the above formula is satisfied, the color of the electrical appliance is considered to be consistent with the target color.

[0077] (2) Size recognition

[0078] For some electrical tools, different models have different lengths and sizes. For example, a voltage tester has a different length depending on the signal it's used to generate. Taking a voltage tester as an example, the depth information H of the target is obtained through a depth camera, and the length of the electrical tool in the image is l. The actual length L of the electrical tool is calculated using the focal length f of the depth camera; the calculation formula is: L = lH / f. The size of the voltage tester can be determined using L.

[0079] Step 4: Based on the target trajectory map, determine the entry and exit of target personnel and electrical tools, and identify whether the electrical tools and the personnel who issued them are consistent; and register the electrical tool issuance information in the electrical tool management system.

[0080] First, based on the target trajectory map, determine the entry and exit of target personnel and electrical tools from the warehouse.

[0081] The two-dimensional trajectory L of target k identified by the LightTrack target tracking and recognition network. k The coordinates of the two-dimensional trajectory points are recorded in chronological order of time as (i represents the i-th time point) (x k,i ,y k,i The coordinates of the two-dimensional trajectory point at the next time step are (x... k,i+1 ,y k,i+1 Personnel k is in (x k,i ,y k,i The direction angle of motion of the position θ ki

[0082]

[0083] Establish a standard trajectory route L for the target's entry and exit from the warehouse (the entry and exit trajectories are the same). The coordinates of the standard trajectory route L are (m... i n i Simultaneously, establish the direction angle θ of the inbound and outbound trajectory. in θ out (usually θ) in =-θ out );

[0084] If the following formula is satisfied, the item is considered to be put into storage. The formula for determining whether an item is put into storage is:

[0085]

[0086] If the following formula is satisfied, the goods are deemed ready for shipment. The shipment determination formula is:

[0087]

[0088] In the formula, T1 and T2 are threshold values;

[0089] If the inbound and outbound judgment formulas are not met, then it is considered another case.

[0090] Second, verify the consistency between the electrical tools and the person who issued them, that is, determine whether the tools were issued to the correct person. This includes:

[0091] By tracking the movement of staff, video frames showing staff entering and leaving the warehouse can be determined.

[0092] In the same video frame, bounding boxes of workers and electrical tools are generated based on the LightTrack target tracking and recognition network. The bounding boxes of workers and tools are used to determine whether the electrical tools were picked up by a certain worker.

[0093] Assuming the number of video frames for worker k entering and leaving the warehouse is G (G is the number of video frames acquired from the start to the end of the trajectory), the bounding box of worker k identified in the g-th frame is K. kg Electrical tools Z j The target bounding box in the g-th frame is M. j,g (j = 1, 2, ..., N), where N is the number of tools and implements identified in the frame.

[0094] Determine the Z power tool j The basis for determining whether staff member K received the item is:

[0095] (1). Initialize J = 1, g = 1;

[0096] (2). If the conditions are met Then J increases by 1; otherwise, J remains unchanged.

[0097] (3). If g ≤ G, increment g by 1 and return to step (2);

[0098] (4). Then determine the electrical equipment Z j It was issued to staff member K.

[0099] Third, the information on the issuance of electrical tools and equipment should be registered in the electrical tool and equipment management system. The information on the issuance of electrical tools and equipment includes the personnel who issued (or put into storage) the tools and equipment, the issuance time, and the model of the tools and equipment.

[0100] Example 2

[0101] like Figure 6 As shown, the difference between this embodiment and Embodiment 1 is that this embodiment provides a machine vision-based power tool inbound / outbound management system, which uses the machine vision-based power tool inbound / outbound management method of Embodiment 1; the system includes:

[0102] The acquisition unit is used to acquire video images of electrical tools entering and leaving the warehouse;

[0103] The multi-target recognition unit is used to input video images of power tools entering and leaving the warehouse into the LightTrack target tracking and recognition network, perform multi-target recognition on each frame of the video images of power tools entering and leaving the warehouse, output single-frame sequence images with target annotation and target trajectory map, and store them in the database.

[0104] The face recognition unit is used to perform face recognition based on a single-frame sequence of images with target annotations, using the key block sparse representation face recognition method.

[0105] The electrical tool identification unit is used to identify the color and size of electrical tools based on a single-frame sequence of images labeled with the target.

[0106] The entry and exit judgment unit is used to judge the entry and exit of target personnel and electrical tools and to identify whether the electrical tools and the personnel who received them are consistent, based on the target trajectory map.

[0107] The requisition information registration unit is used to register the requisition information of electrical tools and equipment in the electrical tool and equipment management system.

[0108] The execution process of each unit can be carried out according to the steps of the machine vision-based power tool entry and exit management method in Example 1, and will not be described in detail in this example.

[0109] Meanwhile, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned machine vision-based method for managing the entry and exit of electrical tools.

[0110] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0111] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0112] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0113] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0114] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A machine vision-based method for managing the entry and exit of electrical tools and equipment, characterized in that, The method includes: Acquire video images of electrical tools entering and leaving the warehouse; The video images of the power tools entering and leaving the warehouse are input into the LightTrack target tracking and recognition network. Multi-target recognition is performed on each frame of the video images of the power tools entering and leaving the warehouse. The single-frame sequence images of the target annotation and the target trajectory map are output and stored in the database. Based on the single-frame sequence of images labeled with the target, the key block sparse representation face recognition method is used to identify the face identity and to identify the color and size of the power tools. Based on the target trajectory map, determine the entry and exit of target personnel and electrical tools, and identify whether the electrical tools and the personnel who issued them are consistent; and register the electrical tool issuance information in the electrical tool management system. Based on the target trajectory map, determine the entry and exit of target personnel and electrical tools, including: If the following formula is satisfied, the item is considered to be put into storage. The formula for determining whether an item is put into storage is: and ; If the following formula is satisfied, the goods are deemed ready for shipment. The shipment determination formula is: and ; ； In the formula, (x k,i ,y k,i ( ) represents the two-dimensional trajectory of target k identified by the LightTrack target tracking and recognition network. The coordinates of the two-dimensional trajectory point at the i-th time point, recorded in chronological order of the trajectory time; (x k,i+1 ,y k,i+1 () represents the coordinates of the two-dimensional trajectory point at the time following the i-th time point; For personnel k in (x k,i ,y k,i The direction angle of motion of the position; (m) i n i The coordinates of the standard trajectory route L of the two-dimensional image of the target entering and leaving the database are shown. , These are the trajectory direction angles for the outbound and inbound directions, respectively. and For the threshold; Whether the electrical tools and the personnel who issued them are consistent includes: By tracking the movement of staff, video frames showing staff entering and leaving the warehouse can be determined. In the same video frame, bounding boxes of workers and electrical tools are generated based on the LightTrack target tracking and recognition network. The bounding boxes of workers and tools are used to determine whether the electrical tools were picked up by a certain worker.

2. The method for managing the entry and exit of electrical tools based on machine vision according to claim 1, characterized in that, Multi-target recognition is performed on each frame of the video images of the entry and exit of the power tools. The multi-target objects include personnel targets and power tool targets. The power tool targets include safety helmets, insulated shoes, insulated gloves, safety belts, insulated rods, voltage testers, and grounding wires.

3. The method for managing the entry and exit of electrical tools based on machine vision according to claim 1, characterized in that, The steps for facial identity recognition using the key block sparse representation facial recognition method are as follows: Match single-sample faces with facial feature templates in the database to identify four key facial regions: left eye, right eye, nose, and mouth. The above four regions are represented using sparse representation: k = 1, 2, 3, 4, corresponding to the left eye, right eye, nose, and mouth respectively; For sparse dictionaries, The sparsity coefficient is . For feature region images; Determine the sparse feature vector of the key facial region of the i-th worker in the database. The feature vectors of the four key facial regions are combined into a matrix. And store it in the database; A facial database is collected using a camera, and the SDM algorithm is used to detect faces and locate feature points. Four key facial regions were extracted through feature template matching; a sparse dictionary A was established. i Solve for the feature vectors of the four feature regions of the collected face. , , , , forming the feature matrix ; make Solve That is, k is The smallest value of i, To reflect the similarity between the collected faces and the faces in the database; The minimum value is represented by E. E is compared with a set threshold T: if E is greater than T, the person is identified as a non-staff member and is prohibited from entering; otherwise, the staff member is identified as the i-th staff member.

4. The machine vision-based method for managing the entry and exit of electrical tools and equipment according to claim 3, characterized in that, The sparse dictionary is constructed by solving the orthogonal matching pursuit method. .

5. The method for managing the entry and exit of electrical tools based on machine vision according to claim 1, characterized in that, Perform color and size identification of electrical tools, including: Represent the target region using the RGB color model and calculate the average RGB values ​​of the target region; based on the average RGB values ​​of the target region... Determine whether the color of the electrical tool matches the target color. Whether they are consistent; the formula for judgment is: ,in, The tolerance range for color change is defined as follows: if the above formula is satisfied, the color of the electrical appliance is considered to be consistent with the target color. The depth information H of the target is obtained by a depth camera. The length of the power tool in the image is l. The actual length L of the power tool is calculated by the focal length f of the depth camera. The calculation formula is: L = lH / f.

6. The method for managing the entry and exit of electrical tools based on machine vision according to claim 1, characterized in that, The information on the requisition of electrical tools includes the personnel who requisitioned the tools, the time of requisition, and the model of the tools.

7. A machine vision-based warehouse management system for electrical tools and equipment, characterized in that, The system uses the machine vision-based power tool inbound and outbound management method as described in any one of claims 1 to 6; the system includes: The acquisition unit is used to acquire video images of electrical tools entering and leaving the warehouse; The multi-target recognition unit is used to input the video images of the power tools entering and leaving the warehouse into the LightTrack target tracking and recognition network, perform multi-target recognition on each frame of the video images of the power tools entering and leaving the warehouse, output single-frame sequence images of target annotation and target trajectory map, and store them in the database. The face recognition unit is used to perform face recognition based on the single-frame sequence image labeled by the target, using the key block sparse representation face recognition method. The electrical tool identification unit is used to identify the color and size of electrical tools based on the single-frame sequence image of the target annotation; The entry and exit judgment unit is used to judge the entry and exit of target personnel and electrical tools and to identify whether the electrical tools and the personnel who received them are consistent, based on the target trajectory map. The requisition information registration unit is used to register the requisition information of electrical tools and equipment in the electrical tool and equipment management system.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the machine vision-based method for managing the entry and exit of electrical tools as described in any one of claims 1 to 6.

Images