Factory intelligent safety inspection system integrating computer vision and VR (virtual reality)
Through the technology of integrating computer vision and VR, automatic planning of inspection routes and detection of abnormalities has been solved, and the problems of low patrol efficiency and safety caused by artificial dependence in the existing technology have been achieved, achieving more efficient and accurate intelligent security patrol in the factory.
Patent Information
- Application Number
- CN202411700537.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2025-03-18
AI Technical Summary
The existing factory safety inspection methods rely on manual sensitivity and fail to effectively plan inspection items, frequency and routes, resulting in low safety and low inspection efficiency.
The intelligent security inspection system in the factory area is adopted that integrates computer vision and VR. Through area division, benchmarking factory area determination, inspection parameter analysis and wireless communication modules, the inspection route is automatically planned, abnormalities are detected and alarms are sent.
It improves the efficiency and accuracy of intelligent safety inspection in the factory, reduces manual errors, and enhances the safety and inspection efficiency of the factory.
Smart Images
Figure CN119647761A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of safety inspection technology, and in particular to an intelligent factory safety inspection system integrating computer vision and VR. Background Art
[0002] At present, computer vision is a field of artificial intelligence, which refers to enabling computers and systems to obtain meaningful information from images, videos and other visual inputs, and to take actions or provide suggestions based on that information. If artificial intelligence gives computers the ability to think, then computer vision gives them the ability to discover, observe and understand. VR fusion technology is a technology that combines virtual reality with other technologies or application areas. This fusion aims to create a more immersive and interactive experience and is widely used in multiple fields. With the increasing development of people's production technology, conventional factory safety inspections can no longer meet people's requirements, so factory intelligent safety inspections based on the fusion of computer vision and VR are crucial.
[0003] The existing factory safety inspection method refers to the staff using handheld equipment, such as smoke detectors, temperature sensors, gas detectors, etc., to inspect various areas within the factory, monitor the environmental parameters of various places within the factory in real time, and then determine whether there are any abnormalities. However, the existing factory safety inspection method is too dependent on the staff's sensitivity to safety issues, and does not take into account the inspection items, inspection frequency and inspection route planning of the factory, resulting in low safety in the factory and low safety inspection efficiency in the dust-wiping functional area. There is room for improvement. Summary of the invention
[0004] In order to improve the efficiency of intelligent safety inspection in factory areas by integrating computer vision and VR, the present application provides an intelligent safety inspection system in factory areas by integrating computer vision and VR.
[0005] A factory intelligent safety inspection system that integrates computer vision and VR adopts the following technical solutions:
[0006] A factory intelligent safety inspection system integrating computer vision and VR, including:
[0007] Plant area division module: configured to divide the area according to the usage function of each location in the plant area to be tested to obtain the plant area division result;
[0008] Benchmarking plant area determination module: configured to analyze the benchmarking plant area of the plant area to be tested according to the production product information of the plant area to be tested to obtain a benchmarking plant area data pool;
[0009] The first inspection parameter analysis module is configured to obtain the first inspection parameter information of the plant to be tested by analyzing the historical safety problems of each benchmark plant and the plant to be tested in the benchmark plant data pool;
[0010] The second inspection parameter analysis module is configured to determine the inspection time of each area of the plant to be tested according to the work and rest time of the plant to be tested, the plant area division result and the first inspection parameter information to obtain the second inspection parameter information of the plant to be tested;
[0011] The third inspection parameter analysis module is configured to plan the factory inspection route based on the location of the factory equipment and the location of the factory personnel to obtain the third inspection parameter information of the factory to be tested;
[0012] The plant inspection module is configured to detect the plant area to be tested based on the first inspection parameter information, the second inspection parameter information and the third inspection parameter information of the plant area to be tested, determine whether there is an abnormal situation, if there is an abnormal situation, determine the abnormal type to obtain abnormal type information, and locate the location where the abnormal situation occurs to obtain abnormal location information;
[0013] The wireless communication module is configured to send the abnormality type information and the abnormality location information to a remote user terminal.
[0014] Preferably, the factory area to be tested is preliminarily divided into regions according to the use functions of each location in the factory area to be tested to obtain a first region division result, wherein the use functions of each location in the factory area to be tested refer to processing function, storage function, office management function and employee life function;
[0015] The regions in the first region division result are further divided into multiple sub-regions to obtain the second region division result, wherein the region with processing function in the tested plant area is further divided into multiple sub-regions according to the scale of working equipment, the region with warehousing and logistics function in the tested plant area is further divided into multiple sub-regions according to the scale of stored items, the region with office management function in the tested plant area is further divided into multiple sub-regions according to the office management content, and the region with employee living function in the tested plant area is further divided into multiple sub-regions according to the number of employees and the living location of each employee;
[0016] The first area division result and the second area division result are combined to form a plant area division result.
[0017] Preferably, a plant area database is obtained, wherein the plant area database includes a plurality of sample plant areas and sample production product information of each sample plant area;
[0018] Obtaining information about the products produced in the plant to be tested, judging the characteristics of the products produced in the plant to be tested based on the information about the products produced in the plant to be tested, and then labeling the products produced in the plant to be tested to obtain product label information;
[0019] Determine the characteristics of the sample production product information of each sample factory in the factory database, and then label the sample production products of each sample factory in the factory database to obtain sample label information;
[0020] Determine priorities of multiple product labels of products produced in the tested plant area according to the product label information to obtain product label priority information, wherein the product label priority information includes a first product label priority and a second product label priority;
[0021] Matching the first product label priority with the label information of each sample plant area sample in the plant area database, determining the label inclusion, and obtaining the first label similarity coefficient AY based on the label inclusion, wherein the higher the label inclusion, the higher the first similarity coefficient AY;
[0022] Matching the second product label priority with the label information of each sample plant area sample in the plant area database, determining the label inclusion, and obtaining the second label similarity coefficient AE based on the label inclusion, wherein the higher the label inclusion, the higher the second similarity coefficient AE;
[0023] According to the first similarity coefficient AY and the second similarity coefficient AE, based on the comparison similarity function AD=a1×AY+a2×AE, the similarity between each sample plant area in the plant area database and the plant area to be tested is analyzed to obtain the comparison similarity coefficient AD of each sample plant area, wherein a1 and a2 are proportional factors, and a1 and a2 are both greater than 0 and a1>a2;
[0024] The comparison similarity coefficient AD of each sample plant area is compared with the preset comparison similarity threshold. If the comparison similarity coefficient AD of the sample plant area is greater than the preset comparison similarity threshold, it is marked as a benchmark plant area, and a benchmark plant area data pool is created, which includes multiple benchmark plant areas.
[0025] Preferably, the historical safety situation of the plant to be tested is obtained to obtain first historical safety information, wherein the first historical safety information includes first historical safety problem type information of the plant to be tested and first historical safety problem occurrence frequency information BY of each type of historical safety problem;
[0026] Acquire the historical safety situation of each benchmarking plant in the benchmarking plant data pool to obtain second historical safety information, wherein the second historical safety information includes second historical safety problem type information of each benchmarking plant in the benchmarking plant data pool and second historical safety problem occurrence frequency information BE of each type of historical safety problem;
[0027] Determine the inspection item information of the plant area to be tested according to the first historical safety problem type information of the plant area to be tested and the second historical safety problem type information of each benchmark plant area in the benchmark plant area data pool;
[0028] According to the first historical safety problem occurrence frequency information BY and the second historical safety problem occurrence frequency information BE, the inspection frequency coefficient BP of the factory inspection item to be tested is calculated based on the inspection frequency function BP=b1×BY+b2×BE, where b1 and b2 are proportional factors, b1 and b2 are both greater than 0 and b1>b2;
[0029] Based on the inspection frequency coefficient BP of the inspection item in the plant area to be tested, the inspection frequency information of each inspection item in the plant area to be tested is obtained, wherein the larger the inspection frequency coefficient BP of the inspection item in the plant area to be tested is, the higher the inspection frequency information of each inspection item in the plant area to be tested is;
[0030] The inspection item information of the plant area to be tested and the inspection frequency information of each inspection item of the plant area to be tested are combined to form the first inspection parameter information of the plant area to be tested.
[0031] Preferably, the work and rest time table of the factory area to be tested is obtained to obtain the work and rest time information of the factory area;
[0032] According to the inspection item information in the first inspection parameter information and the plant area division result, the inspection items that need to be inspected in each area of the plant area to be tested are determined, and the regional inspection item information of each area of the plant area to be tested is obtained;
[0033] Obtaining the production management scale information of each area of the plant to be tested, and according to the inspection frequency information of each inspection item of the plant to be tested in the first inspection parameter information, determining the detection frequency of the regional inspection item information of each area of the plant to be tested based on the production management scale information of each area of the plant to be tested to obtain the regional detection frequency information of the regional inspection items of each area of the plant to be tested;
[0034] According to the factory work and rest time information of the factory to be tested, it is determined whether each area of the factory to be tested is in use. If it is in use, the inspection time of each area of the factory to be tested during all periods is determined based on the factory work and rest time information and the area detection frequency information to obtain the second inspection parameter information.
[0035] Preferably, a drone equipped with a camera flies around the plant to be tested, the camera photographs each area of the plant to be tested to obtain photographed image information, and a digital twin model of the plant to be tested is created based on the photographed image information;
[0036] Obtain the location of the production equipment in the factory to obtain the location information of the production equipment in the factory;
[0037] Perform real-time positioning operations on personnel in the plant to be tested to obtain personnel location information at each moment, and based on the personnel location information at each moment, count the number of personnel in each area of the plant to be tested in real time to obtain regional personnel number information CS. Based on VR fusion technology, display the personnel location information at each moment in each area of the digital twin model of the plant to be tested in real time to obtain regional personnel flow information CL;
[0038] Determine whether there is a correlation between the regional inspection items of each area of the plant to be tested and the production equipment in the plant based on the regional inspection item information of the plant to be tested;
[0039] If there is no association, the route for sequentially detecting each area of the tested plant is marked as the first plant inspection route. If there is an association, the priority of each area of the tested plant is determined based on the route planning function CG=c1×CS+c2×CL according to the regional personnel quantity information CS and the regional personnel flow information CL of each area of the tested plant, and the regional priority coefficient CG is obtained, where c1 and c2 are proportional factors and are both greater than 0. Based on the regional priority coefficient CG, the inspection routes of each area of the tested plant are planned to obtain the second plant inspection route, that is, the higher the regional priority coefficient CG, the higher the priority coefficient CG is.
[0040] The first factory inspection route or the second factory inspection route is the third inspection parameter information.
[0041] Preferably, according to the inspection item information of the plant area to be tested, the inspection frequency information of each inspection item of the plant area to be tested, the second inspection parameter information and the third inspection parameter information, the drone is equipped with a camera, a sensor unit and a positioning unit to inspect each area of the plant area to be tested, and it is determined whether there is an abnormality. If there is an abnormality, the abnormality type is determined to obtain abnormality type information;
[0042] Based on the positioning unit, a positioning operation is performed on the location where the abnormal situation exists to obtain abnormal location information.
[0043] Preferably, a wireless communication module is obtained and signal connection links are established between the wireless communication module and the camera, the sensor unit and the positioning unit respectively;
[0044] The wireless communication module obtains the abnormality type information and the abnormality location information and performs cache processing;
[0045] The abnormality type information and the abnormality location information are sent to a remote user terminal based on a wireless communication module.
[0046] In summary, the present application includes at least one of the following beneficial technical effects:
[0047] 1. The characteristics of the products produced in the tested plant area are judged by the product information of the tested plant area, and then the products produced in the tested plant area are labeled to obtain product label information. The characteristics of the sample production product information are judged, and then the sample production products of each sample plant area are labeled to obtain sample label information. The priority of each product label is judged according to the product label information to obtain product label priority information. The product label priority information is matched with the sample label information, and then the similarity between each sample plant area and the tested plant area in the plant area database is analyzed to obtain the comparison similarity coefficient AD of each sample plant area, and then the benchmark plant area is obtained and a benchmark plant area data pool is created, which provides a suitable benchmark plant area for the subsequent intelligent safety inspection of the tested plant area, thereby improving the efficiency of intelligent safety inspection of the plant area integrated with computer vision and VR;
[0048] 2. With the help of a drone equipped with a camera, fly around the plant to be tested, and obtain image information by shooting with the camera. Create a digital twin model of the plant to be tested based on the image information, obtain the location information of the plant's production equipment, and obtain the regional personnel quantity information CS and regional personnel flow information CL based on the personnel location information of the plant to be tested at each moment. Determine whether there is a correlation between the regional inspection items of each area of the plant to be tested and the production equipment in the plant. If there is no correlation, mark the route that is inspected in turn in each area of the plant to be tested as the first plant inspection route. If there is a correlation, determine the priority of each area of the plant to be tested based on the regional personnel quantity information CS and regional personnel flow information CL of each area of the plant to be tested to obtain the regional priority coefficient CG, and then plan the inspection route based on the regional priority coefficient CG, thereby improving the accuracy of the inspection route of the plant to be tested, and thereby improving the efficiency of intelligent safety inspection of the plant based on the fusion of computer vision and VR. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 This is a module diagram of the factory intelligent safety inspection system that mainly reflects the integration of computer vision and VR in this embodiment.
[0050] Figure numerals: 1. Plant area division module; 2. Benchmarking plant area determination module; 3. First patrol inspection parameter analysis module; 4. Second patrol inspection parameter analysis module; 5. Third patrol inspection parameter analysis module; 6. Plant area patrol module; 7. Wireless communication module. DETAILED DESCRIPTION
[0051] The present application is further described in detail below in conjunction with the accompanying drawings.
[0052] The embodiment of the present application discloses a factory intelligent safety inspection system integrating computer vision and VR.
[0053] Reference Figure 1A factory area intelligent safety inspection system integrating computer vision and VR includes a factory area division module, a benchmark factory area determination module, a first inspection parameter analysis module, a second inspection parameter analysis module, a third inspection parameter analysis module, a factory area inspection module and a wireless communication module.
[0054] Reference Figure 1 , Plant area division module: It is configured to divide the area according to the usage function of each location in the plant to be tested to obtain the plant area division result.
[0055] Among them, the specific implementation method of the plant area division module is as follows:
[0056] The factory area to be tested is preliminarily divided into regions according to the use functions of each location in the factory area to be tested to obtain a first region division result, wherein the use functions of each location in the factory area to be tested refer to processing function, warehousing function, office management function and employee life function.
[0057] In actual applications, the area responsible for processing functions refers to the area used for various production processes and processing operations; the area responsible for warehousing functions refers to the area used for storage of raw materials and temporary storage of products; the area responsible for office management functions refers to the area used for daily management of the factory, decision making, production scheduling and quality control, including offices, conference rooms, control centers, etc.; the area responsible for employee living functions refers to employee dormitories, canteens, entertainment facilities and other areas.
[0058] Each area in the first area division result is further divided into multiple sub-areas to obtain a second area division result, wherein the area with processing function in the tested factory area is further divided into multiple sub-areas according to the scale of working equipment, the area with warehousing and logistics function in the tested factory area is further divided into multiple sub-areas according to the scale of stored items, the area with office management function in the tested factory area is further divided into multiple sub-areas according to the office management content, and the area with employee living function in the tested factory area is further divided into multiple sub-areas according to the number of employees and the living locations of each employee.
[0059] The first area division result and the second area division result are combined to form a plant area division result.
[0060] Reference Figure 1 , Benchmarking plant area determination module: configured to analyze the benchmarking plant areas of the plant area to be tested according to the production product information of the plant area to be tested to obtain a benchmarking plant area data pool.
[0061] Among them, the specific implementation method of the benchmarking plant area determination module is as follows:
[0062] A plant area database is obtained, wherein the plant area database includes multiple sample plants and sample production product information of each sample plant area.
[0063] The production product information of the plant to be tested is obtained, and the production product characteristics of the plant to be tested are judged according to the production product information of the plant to be tested, and then the production products of the plant to be tested are labeled to obtain product label information.
[0064] Let's take an example to illustrate: for example, if the product of the factory to be tested is shoes, then the material, type, size range, applicable population and other characteristics of the shoes are determined, and then the shoes are labeled according to the characteristics of the shoes. For example, if the material of the shoes is leather, the type is short boots, the size range is 35-40, and the target population is female, then the shoes are labeled with leather labels, short boots labels, 35-40 labels, and female labels.
[0065] The characteristics of the sample production product information of each sample factory in the factory database are determined, and then labels are affixed to the sample production products of each sample factory in the factory database to obtain sample label information.
[0066] The priorities of multiple product labels of the products produced in the tested factory are determined according to the product label information to obtain product label priority information, where the product label priority information includes a first product label priority and a second product label priority.
[0067] Let's take an example to illustrate: for example, the product produced by the factory to be tested is shoes, and the shoes are affixed with leather labels, short boot labels, size 35-40 labels, and women's labels. In this case, the leather labels and short boot labels are marked as the first product label priority, and the size 35-40 labels and women's labels are marked as the second product label priority.
[0068] The first product label priority is matched with the label information of each sample plant sample in the plant database, the label inclusion is determined, and the first label similarity coefficient AY is obtained based on the label inclusion, wherein the higher the label inclusion, the higher the first similarity coefficient AY.
[0069] Let's take an example to illustrate: for example, the products produced by the factory to be tested are shoes, and the first product label priority is leather labels and short boots labels. If the sample label information of the sample factory contains leather labels and short boots labels, the label inclusion degree is 100%; if the sample label information of the sample factory contains leather labels or short boots labels, the label inclusion degree is 50%; if the sample label information of the sample factory does not contain leather labels and short boots labels, the label inclusion degree is 0%.
[0070] The second product label priority is matched with the label information of each sample plant sample in the plant database, the label inclusion is determined, and the second label similarity coefficient AE is obtained based on the label inclusion, wherein the higher the label inclusion, the higher the second similarity coefficient AE.
[0071] Let's take an example to illustrate: for example, the product produced by the factory to be tested is shoes, and the second product label priority is the size 35-40 label and the female label. If the sample label information of the sample factory contains the size 35-40 label and the female label, the label inclusion degree is 100%; if the sample label information of the sample factory contains the size 35-40 label or the female label, the label inclusion degree is 50%; if the sample label information of the sample factory does not contain the size 35-40 label and does not contain the female label, the label inclusion degree is 0%.
[0072] According to the first similarity coefficient AY and the second similarity coefficient AE, calculation is performed based on the comparison similarity function AD=a1×AY+a2×AE, and the similarity between each sample factory area in the factory area database and the factory area to be tested is analyzed respectively to obtain the comparison similarity coefficient AD of each sample factory area, wherein a1 and a2 are proportional factors, and both a1 and a2 are greater than 0 and a1>a2.
[0073] The comparison similarity coefficient AD of each sample plant area is compared with the preset comparison similarity threshold. If the comparison similarity coefficient AD of the sample plant area is greater than the preset comparison similarity threshold, it is marked as a benchmark plant area, and a benchmark plant area data pool is created, which includes multiple benchmark plant areas.
[0074] In specific applications, the production product characteristics of the plant to be tested are judged through the production product information of the plant to be tested, and then the production products of the plant to be tested are labeled to obtain product label information. By judging the characteristics of the sample production product information, the sample production products of each sample plant are labeled to obtain sample label information. The priority of each product label is judged according to the product label information to obtain product label priority information. The product label priority information is matched with the sample label information, and then the similarity between each sample plant in the plant database and the plant to be tested is analyzed to obtain the comparison similarity coefficient AD of each sample plant. Then, the benchmark plant is obtained and a benchmark plant data pool is created, which provides a suitable benchmark plant for the subsequent intelligent safety inspection of the plant to be tested, thereby improving the efficiency of the intelligent safety inspection of the plant integrating computer vision and VR.
[0075] Reference Figure 1 , the first inspection parameter analysis module: is configured to analyze the historical safety problems of each benchmarking factory area and the factory area to be tested in the benchmarking factory area data pool to obtain the first inspection parameter information of the factory area to be tested.
[0076] The specific execution method of the first inspection parameter analysis module is as follows:
[0077] The historical safety situation of the plant to be tested is obtained to obtain first historical safety information, where the first historical safety information includes first historical safety problem type information of the plant to be tested and first historical safety problem occurrence frequency information BY of each type of historical safety problem.
[0078] The historical safety conditions of each benchmark plant in the benchmark plant data pool are obtained to obtain the second historical safety information, which includes the second historical safety problem type information of each benchmark plant in the benchmark plant data pool and the second historical safety problem occurrence frequency information BE of each type of historical safety problem.
[0079] According to the first historical safety problem type information of the plant area to be tested and the second historical safety problem type information of each benchmark plant area in the benchmark plant area data pool, the inspection item information of the plant area to be tested is determined.
[0080] Let's take an example to illustrate: for example, the first historical safety problem type information of the plant to be tested includes line fire problems and warehouse fire problems, and the second historical safety problem type information of the benchmark plant includes line fire problems and pipeline leakage problems. Then the inspection items of the plant to be tested are line fire detection, warehouse fire detection and pipeline leakage detection.
[0081] According to the first historical safety problem occurrence frequency information BY and the second historical safety problem occurrence frequency information BE, the inspection frequency coefficient BP of the factory inspection item to be tested is calculated based on the inspection frequency function BP=b1×BY+b2×BE, where b1 and b2 are proportional factors, both b1 and b2 are greater than 0 and b1>b2.
[0082] The inspection frequency information of each inspection item in the plant area to be tested is obtained based on the inspection frequency coefficient BP of the inspection item in the plant area to be tested, wherein the larger the inspection frequency coefficient BP of the inspection item in the plant area to be tested is, the higher the inspection frequency information of each inspection item in the plant area to be tested is.
[0083] The inspection item information of the plant area to be tested and the inspection frequency information of each inspection item of the plant area to be tested are combined to form the first inspection parameter information of the plant area to be tested.
[0084] Reference Figure 1 The second inspection parameter analysis module is configured to determine the inspection time of each area of the plant to be tested according to the work and rest time of the plant to be tested, the plant area division result and the first inspection parameter information to obtain the second inspection parameter information of the plant to be tested.
[0085] The specific execution method of the second inspection parameter analysis module is as follows:
[0086] Obtain the work and rest time table of the factory to be tested to obtain the work and rest time information of the factory.
[0087] According to the inspection item information in the first inspection parameter information and the plant area division result, the inspection items that need to be inspected in each area of the plant to be tested are determined, and the regional inspection item information of each area of the plant to be tested is obtained.
[0088] Obtain the production management scale information of each area of the factory to be tested, and according to the inspection frequency information of each inspection item of the factory to be tested in the first inspection parameter information, determine the inspection frequency of the regional inspection item information of each area of the factory to be tested based on the production management scale information of each area of the factory to be tested, to obtain the regional inspection frequency information of the regional inspection items of each area of the factory to be tested.
[0089] Let's take an example to illustrate: the regional inspection project information of the area in the plant to be tested that is responsible for the warehousing function is the warehouse fire problem, and the inspection frequency information of the warehouse fire inspection project in the plant to be tested is once every half an hour. If the production management scale information of the area in the plant to be tested that is responsible for the warehousing function, that is, the warehouse storage capacity, is half of the maximum storage capacity, then the inspection frequency information of the warehouse fire inspection project in the area in the plant to be tested that is responsible for the warehousing function is once per hour.
[0090] According to the factory work and rest time information of the factory to be tested, it is determined whether each area of the factory to be tested is in use. If it is in use, the inspection time of each area of the factory to be tested during all periods is determined based on the factory work and rest time information and the area detection frequency information to obtain the second inspection parameter information.
[0091] Reference Figure 1 The third inspection parameter analysis module is configured based on computer vision technology and VR fusion technology to plan the factory inspection route according to the location of factory equipment and factory personnel to obtain the third inspection parameter information of the factory to be tested.
[0092] The specific implementation of the third inspection parameter analysis module is as follows:
[0093] Based on the drone equipped with a camera flying around the factory area to be tested, the camera takes pictures of various areas of the factory area to be tested to obtain image information, and a digital twin model of the factory area to be tested is created based on the image information.
[0094] Get the location of the factory production equipment to obtain the location information of the factory production equipment.
[0095] Real-time positioning operations are performed on personnel in the factory area to be tested to obtain personnel location information at each moment. Based on the personnel location information at each moment, the number of personnel in each area of the factory area to be tested is counted in real time to obtain regional personnel number information CS. Based on VR fusion technology, the personnel location information at each moment is displayed in real time in each area of the digital twin model of the factory area to be tested to obtain regional personnel flow information CL.
[0096] According to the regional inspection item information of the plant to be tested, it is determined whether there is a correlation between the regional inspection items of each area of the plant to be tested and the production equipment in the plant.
[0097] If there is no association, the route that detects each area of the factory to be tested in turn will be marked as the first factory inspection route. If there is a correlation, the priority of each area of the factory to be tested is determined based on the route planning function CG=c1×CS+c2×CL according to the regional personnel quantity information CS and the regional personnel flow information CL of each area of the factory to be tested, and the regional priority coefficient CG is obtained, where c1 and c2 are proportional factors and are both greater than 0. Based on the regional priority coefficient CG, the inspection routes of each area of the factory to be tested are planned to obtain the second factory inspection route, that is, the higher the regional priority coefficient CG, the higher the priority of the area to be tested.
[0098] The first factory area inspection route or the second factory area inspection route is the third inspection parameter information.
[0099] In specific applications, a drone equipped with a camera flies around the factory area to be tested, and the camera is used to capture image information. A digital twin model of the factory area to be tested is created based on the captured image information, and the location information of the factory production equipment is obtained. According to the personnel location information of the factory area at each moment, the regional personnel quantity information CS and the regional personnel flow information CL are obtained. It is determined whether there is a correlation between the regional inspection items of each area of the factory area to be tested and the factory production equipment. If there is no correlation, the route inspected in each area of the factory area to be tested in turn is marked as the first factory inspection route. If there is a correlation, the priority of each area of the factory area to be tested is determined according to the regional personnel quantity information CS and the regional personnel flow information CL of each area of the factory area to be tested, and the regional priority coefficient CG is obtained. Then, the inspection route is planned according to the regional priority coefficient CG, which improves the accuracy of the inspection route of the factory area to be tested, thereby improving the efficiency of intelligent safety inspection of the factory area based on the fusion of computer vision and VR.
[0100] Reference Figure 1 The factory inspection module is configured to detect the factory area to be tested based on the first inspection parameter information, the second inspection parameter information and the third inspection parameter information of the factory area to be tested, and determine whether there is an abnormal situation. If there is an abnormal situation, determine the abnormal type to obtain abnormal type information, and locate the location where the abnormal situation occurs to obtain abnormal location information.
[0101] Among them, the specific implementation method of the factory inspection module is as follows:
[0102] According to the inspection item information of the factory area to be tested, the inspection frequency information of each inspection item in the factory area to be tested, the second inspection parameter information and the third inspection parameter information, the camera, sensor unit and positioning unit are installed on the drone to inspect each area of the factory area to be tested, and it is determined whether there is any abnormal situation. If there is an abnormal situation, the type of abnormality is determined to obtain the abnormality type information.
[0103] Based on the positioning unit, a positioning operation is performed on the location where the abnormal situation exists to obtain abnormal location information.
[0104] Reference Figure 1 , the wireless communication module is configured to send the abnormality type information and the abnormality location information to the remote user terminal.
[0105] Among them, the specific implementation method of the wireless communication module is as follows:
[0106] A wireless communication module is obtained and signal connection links are established between the wireless communication module and the camera, the sensor unit and the positioning unit respectively.
[0107] The wireless communication module obtains the abnormal type information and abnormal location information and performs cache processing.
[0108] The abnormal type information and abnormal location information are sent to the remote user terminal based on the wireless communication module.
[0109] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.
Claims
1. A factory intelligent safety inspection system integrating computer vision and VR, characterized in that: The following steps are involved: Plant area division module: configured to divide the area according to the usage function of each location in the plant area to be tested to obtain the plant area division result; Benchmarking plant area determination module: configured to analyze the benchmarking plant area of the plant area to be tested according to the production product information of the plant area to be tested to obtain a benchmarking plant area data pool; The first inspection parameter analysis module is configured to obtain the first inspection parameter information of the plant to be tested by analyzing the historical safety problems of each benchmark plant and the plant to be tested in the benchmark plant data pool; The second inspection parameter analysis module is configured to determine the inspection time of each area of the plant to be tested according to the work and rest time of the plant to be tested, the plant area division result and the first inspection parameter information to obtain the second inspection parameter information of the plant to be tested; The third inspection parameter analysis module is configured to plan the factory inspection route based on the location of the factory equipment and the location of the factory personnel to obtain the third inspection parameter information of the factory to be tested; The plant inspection module is configured to detect the plant area to be tested based on the first inspection parameter information, the second inspection parameter information and the third inspection parameter information of the plant area to be tested, determine whether there is an abnormal situation, if there is an abnormal situation, determine the abnormal type to obtain abnormal type information, and locate the location where the abnormal situation occurs to obtain abnormal location information; The wireless communication module is configured to send the abnormality type information and the abnormality location information to a remote user terminal.
2. According to claim 1, a computer vision and VR integrated factory area intelligent safety inspection system is characterized in that: The specific implementation method of the plant area division module is as follows: The factory area to be tested is preliminarily divided into regions according to the use functions of each location in the factory area to be tested to obtain a first region division result, wherein the use functions of each location in the factory area to be tested refer to processing function, storage function, office management function and employee life function; The regions in the first region division result are further divided into multiple sub-regions to obtain the second region division result, wherein the region with processing function in the tested plant area is further divided into multiple sub-regions according to the scale of working equipment, the region with warehousing and logistics function in the tested plant area is further divided into multiple sub-regions according to the scale of stored items, the region with office management function in the tested plant area is further divided into multiple sub-regions according to the office management content, and the region with employee living function in the tested plant area is further divided into multiple sub-regions according to the number of employees and the living location of each employee; The first area division result and the second area division result are combined to form a plant area division result.
3. According to claim 2, a computer vision and VR integrated factory area intelligent safety inspection system is characterized in that: The specific implementation method of the benchmarking plant area determination module is as follows: Acquire a plant area database, wherein the plant area database includes multiple sample plant areas and sample production product information of each sample plant area; Obtaining information about the products produced in the plant to be tested, judging the characteristics of the products produced in the plant to be tested based on the information about the products produced in the plant to be tested, and then labeling the products produced in the plant to be tested to obtain product label information; Determine the characteristics of the sample production product information of each sample factory in the factory database, and then label the sample production products of each sample factory in the factory database to obtain sample label information; Determine priorities of multiple product labels of products produced in the tested plant area according to the product label information to obtain product label priority information, wherein the product label priority information includes a first product label priority and a second product label priority; Matching the first product label priority with the label information of each sample plant area sample in the plant area database, determining the label inclusion, and obtaining the first label similarity coefficient AY based on the label inclusion, wherein the higher the label inclusion, the higher the first similarity coefficient AY; Matching the second product label priority with the label information of each sample plant area sample in the plant area database, determining the label inclusion, and obtaining the second label similarity coefficient AE based on the label inclusion, wherein the higher the label inclusion, the higher the second similarity coefficient AE; According to the first similarity coefficient AY and the second similarity coefficient AE, based on the comparison similarity function AD=a1×AY+a2×AE, the similarity between each sample plant area in the plant area database and the plant area to be tested is analyzed to obtain the comparison similarity coefficient AD of each sample plant area, wherein a1 and a2 are proportional factors, and a1 and a2 are both greater than 0 and a1>a2; The comparison similarity coefficient AD of each sample plant area is compared with the preset comparison similarity threshold. If the comparison similarity coefficient AD of the sample plant area is greater than the preset comparison similarity threshold, it is marked as a benchmark plant area, and a benchmark plant area data pool is created, which includes multiple benchmark plant areas.
4. According to claim 3, a computer vision and VR integrated factory area intelligent safety inspection system is characterized in that: The specific execution method of the first inspection parameter analysis module is as follows: Acquire the historical safety situation of the plant to be tested to obtain first historical safety information, wherein the first historical safety information includes first historical safety problem type information of the plant to be tested and first historical safety problem occurrence frequency information BY of each type of historical safety problem; Acquire the historical safety situation of each benchmarking plant in the benchmarking plant data pool to obtain second historical safety information, wherein the second historical safety information includes second historical safety problem type information of each benchmarking plant in the benchmarking plant data pool and second historical safety problem occurrence frequency information BE of each type of historical safety problem; Determine the inspection item information of the plant area to be tested according to the first historical safety problem type information of the plant area to be tested and the second historical safety problem type information of each benchmark plant area in the benchmark plant area data pool; According to the first historical safety problem occurrence frequency information BY and the second historical safety problem occurrence frequency information BE, the inspection frequency coefficient BP of the factory inspection item to be tested is calculated based on the inspection frequency function BP=b1×BY+b2×BE, where b1 and b2 are proportional factors, b1 and b2 are both greater than 0 and b1>b2; Based on the inspection frequency coefficient BP of the inspection item in the plant area to be tested, the inspection frequency information of each inspection item in the plant area to be tested is obtained, wherein the larger the inspection frequency coefficient BP of the inspection item in the plant area to be tested is, the higher the inspection frequency information of each inspection item in the plant area to be tested is; The inspection item information of the plant area to be tested and the inspection frequency information of each inspection item of the plant area to be tested are combined to form the first inspection parameter information of the plant area to be tested.
5. According to claim 4, a factory intelligent safety inspection system integrating computer vision and VR is characterized in that: The specific execution method of the second inspection parameter analysis module is as follows: Obtain the work and rest time table of the factory to be tested to obtain the work and rest time information of the factory; According to the inspection item information in the first inspection parameter information and the plant area division result, the inspection items that need to be inspected in each area of the plant area to be tested are determined, and the regional inspection item information of each area of the plant area to be tested is obtained; Obtaining the production management scale information of each area of the plant to be tested, and according to the inspection frequency information of each inspection item of the plant to be tested in the first inspection parameter information, determining the detection frequency of the regional inspection item information of each area of the plant to be tested based on the production management scale information of each area of the plant to be tested to obtain the regional detection frequency information of the regional inspection items of each area of the plant to be tested; According to the factory work and rest time information of the factory to be tested, it is determined whether each area of the factory to be tested is in use. If it is in use, the inspection time of each area of the factory to be tested during all periods is determined based on the factory work and rest time information and the area detection frequency information to obtain the second inspection parameter information.
6. According to claim 5, a computer vision and VR integrated factory area intelligent safety inspection system is characterized in that: The specific implementation method of the third inspection parameter analysis module is as follows: A drone equipped with a camera flies around the plant to be tested, the camera photographs each area of the plant to be tested to obtain photographed image information, and a digital twin model of the plant to be tested is created based on the photographed image information; Obtain the location of the production equipment in the factory to obtain the location information of the production equipment in the factory; Perform real-time positioning operations on personnel in the plant to be tested to obtain personnel location information at each moment, and based on the personnel location information at each moment, count the number of personnel in each area of the plant to be tested in real time to obtain regional personnel number information CS. Based on VR fusion technology, display the personnel location information at each moment in each area of the digital twin model of the plant to be tested in real time to obtain regional personnel flow information CL; Determine whether there is a correlation between the regional inspection items of each area of the plant to be tested and the production equipment in the plant based on the regional inspection item information of the plant to be tested; If there is no association, the route for sequentially detecting each area of the tested plant is marked as the first plant inspection route. If there is an association, the priority of each area of the tested plant is determined based on the route planning function CG=c1×CS+c2×CL according to the regional personnel quantity information CS and the regional personnel flow information CL of each area of the tested plant, and the regional priority coefficient CG is obtained, where c1 and c2 are proportional factors and are both greater than 0. Based on the regional priority coefficient CG, the inspection routes of each area of the tested plant are planned to obtain the second plant inspection route, that is, the higher the regional priority coefficient CG, the higher the priority coefficient CG is. The first factory inspection route or the second factory inspection route is the third inspection parameter information.
7. The factory intelligent safety inspection system integrating computer vision and VR according to claim 6 is characterized in that: The specific implementation method of the plant inspection module is as follows: According to the inspection item information of the plant to be tested, the inspection frequency information of each inspection item of the plant to be tested, the second inspection parameter information and the third inspection parameter information, the drone is equipped with a camera, a sensor unit and a positioning unit to inspect each area of the plant to be tested, and whether there is an abnormality is determined. If there is an abnormality, the abnormality type is determined to obtain abnormality type information; Based on the positioning unit, a positioning operation is performed on the location where the abnormal situation exists to obtain abnormal location information.
8. The factory intelligent safety inspection system integrating computer vision and VR according to claim 7 is characterized in that: The specific implementation method of the wireless communication module is as follows: Acquire a wireless communication module and establish signal connection links between the wireless communication module and the camera, the sensor unit, and the positioning unit respectively; The wireless communication module obtains the abnormality type information and the abnormality location information and performs cache processing; The abnormality type information and the abnormality location information are sent to a remote user terminal based on a wireless communication module.