Light track synchronous warning method and device, electronic equipment and storage medium

Abstract

The invention relates to a light track synchronous warning method and device, electronic equipment and a storage medium, and belongs to the technical field of safety monitoring, and the method comprises the steps: recognizing and obtaining a current real-time hooking route, and carrying out the consistency verification of the current real-time hooking route and a real-time projection parameter of a projection optical module, and obtaining a warning consistency parameter; performing hook-line feature prediction in the next preset time period to obtain a predicted hook-line route and predicted hook-line features; obtaining predicted visual identification parameters, and performing visual data identification configuration in a next preset time period; and optimal projection parameters are obtained, and light track synchronous warning control of the projection optical module is carried out in the next preset time period. The technical problems that in the prior art, due to the fact that the projection light track cannot be synchronized with the hook running path in real time, the light track warning area lags behind the actual hook running path, and the warning precision is insufficient are solved, and the technical effects that the projection light track and the hook running path are synchronized in real time, and accurate light track warning is achieved are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of safety monitoring, in particular to a light rail synchronous warning method and device, an electronic device and a storage medium. BACKGROUND

[0002] In industrial hoisting operations, in order to ensure operation safety, the running path of the hook needs to be warned in real time to remind surrounding personnel to avoid and identify dangerous areas. In the prior art, a projection optical module is often used to project a light rail on the ground to achieve the warning function, and a light spot or light line is used to mark the expected running path of the hook.

[0003] However, the existing light rail synchronous warning usually uses preset fixed projection parameters, or only performs simple tracking projection according to the current position of the hook. When the hook changes the path, adjusts the speed or changes the posture during actual operation, the projection optical module cannot respond to these dynamic changes in time, resulting in a significant time lag between the projected light rail and the actual hook running path. This lag makes the projected light rail warning area unable to accurately reflect the real running state of the hook, especially in the operation scene where the hook runs at a high speed or the path frequently changes, the deviation between the warning area and the actual dangerous area is more significant, thereby reducing the accuracy and reliability of the warning and affecting the operation safety protection effect. SUMMARY

[0004] The purpose of the present application is to provide a light rail synchronous warning method and device, an electronic device and a storage medium, so that the projected light rail and the hook running path are real-time synchronized, and accurate light rail warning is achieved.

[0005] In a first aspect of the present application, a light rail synchronous warning method is provided, comprising: collecting a visual data sequence in the running process of a hook through a visual module arranged in space, identifying a current real-time hook running path according to a visual recognition parameter determined in a previous preset time period, performing consistency verification with a real-time projection parameter of a projection optical module, and obtaining a warning consistency parameter; predicting a hook running feature in a next preset time period according to the visual data sequence, and obtaining a predicted hook running path and a predicted hook running feature; configuring a visual recognition parameter in the next preset time period according to the predicted hook running feature and the warning consistency parameter, and obtaining a predicted visual recognition parameter to configure visual data recognition in the next preset time period; configuring a projection parameter in the next preset time period according to the predicted hook running path, and obtaining an optimal projection parameter to control the light rail synchronous warning of the projection optical module in the next preset time period.

[0006] In a second aspect, the application provides a light track synchronous warning device, a visual data acquisition module is configured to acquire a visual data sequence of a hook during operation by a visual module arranged in a space, identify a current real-time hook route according to a visual recognition parameter determined in a previous preset time period, and verify consistency of the real-time projection parameter of a projection optical module, to obtain a warning consistency parameter; a hook route feature prediction module is configured to predict a hook route feature in a next preset time period according to the visual data sequence, to obtain a predicted hook route and a predicted hook route feature; a visual parameter configuration module is configured to configure a visual recognition parameter in the next preset time period according to the predicted hook route feature and the warning consistency parameter, to obtain a predicted visual recognition parameter, and to configure visual data recognition in the next preset time period; and a projection parameter configuration module is configured to configure a projection parameter in the next preset time period according to the predicted hook route, to obtain an optimal projection parameter, and to control the light track synchronous warning of the projection optical module in the next preset time period.

[0007] In a third aspect, the application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the light track synchronous warning method described above when running the computer program.

[0008] In a fourth aspect, the application provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the light track synchronous warning method described above when executed by a processor.

[0009] The light track synchronous warning method and device, the electronic device, and the storage medium provided by the application have the following advantages: The visual module arranged in the space is used to collect a visual data sequence in the running process of the hook, and the current real-time hook running route is identified according to the visual recognition parameter determined in the last preset time period, and the real-time projection parameter of the projection optical module is verified for consistency, and the warning consistency parameter is obtained, and a real-time evaluation mechanism of the projection warning effect is established, which provides an accuracy basis for subsequent parameter adjustment. According to the visual data sequence, the hook running feature in the next preset time period is predicted, the predicted hook running route and the predicted hook running feature are obtained, the change from passive tracking to active prediction is realized, the running trend of the hook is grasped in advance, and a prediction basis is provided for the advanced configuration of the projection parameter. According to the predicted hook running feature and the warning consistency parameter, the visual recognition parameter configuration in the next preset time period is performed, the predicted visual recognition parameter is obtained, and the visual data recognition configuration in the next preset time period is performed, so as to dynamically adjust the recognition accuracy according to the predicted running feature, and realize the adaptive balance of the recognition efficiency and the accuracy. According to the predicted hook running route, the projection parameter configuration in the next preset time period is performed, the optimal projection parameter is obtained, the light track synchronous warning control of the projection optical module in the next preset time period is performed, the predictive projection of the projection light track is realized, the lag problem in the traditional technology is eliminated, and the real-time synchronization of the light track warning area and the actual hook running route is ensured.

[0010] Through the above technical solution, the technical problem that the projection light track cannot be synchronized with the hook running path in real time is solved, the accurate synchronization of the projection light track and the hook running path is realized, and the technical effect of precise light track warning is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0012] Figure 1 The flowchart of the light track synchronous warning method provided by an embodiment of the present application is shown in the figure. Figure 2 The structural block diagram of the light track synchronous warning device provided by an embodiment of the present application is shown in the figure. Figure 3 The schematic block diagram of the electronic device provided by an embodiment of the present application is shown in the figure.

[0013] In the drawings, the reference signs are as follows: Track synchronization warning device 20, visual data acquisition module 21, hook running feature prediction module 22, visual parameter configuration module 23, projection parameter configuration module 24, electronic device 300, processor 301, input device 302, output device 303, memory 304, communication bus 305. DETAILED DESCRIPTION

[0014] In the following description, for the purpose of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0015] In order to make the purpose, technical scheme and advantages of the present application clearer, specific embodiments will be described below with reference to the accompanying drawings.

[0016] Please refer to Figure 1 , Figure 1 The flowchart of the track synchronization warning method provided by an embodiment of the present application includes: S1, through the visual module arranged in the space, the visual data sequence in the hook running process is collected, the real-time hook running route is identified according to the visual recognition parameter decided in the last preset time period, and the consistency verification is performed with the real-time projection parameter of the projection optical module, and the warning consistency parameter is obtained.

[0017] Specifically, first, the visual module is arranged at a predetermined position in the hook operation space, and the image sequence in the hook running process is continuously collected in the current preset time period to form the visual data sequence. Then, the visual recognition parameter decided in the last preset time period is called, and the visual recognition parameter is adaptively configured based on the early hook running feature analysis and warning consistency feedback. Based on the visual recognition parameter, the visual data sequence is intelligently analyzed, the actual running track of the hook in the current preset time period is identified and extracted, and the real-time hook running route is obtained.

[0018] Subsequently, the real-time projection parameter of the projection optical module in the current preset time period is obtained, including the spatial positioning information such as the projection coverage area coordinates. By spatially matching and analyzing the real-time hook running route and the real-time projection parameter, the degree of coincidence of the hook running route and the projection coverage area is calculated, and the synchronization accuracy of the projection track and the actual hook running path is evaluated. Then, based on the matching analysis result, the warning consistency parameter is generated, which is used to quantify the accuracy of the current track synchronization warning, and provides a feedback basis for subsequent visual recognition parameter optimization and projection parameter adjustment.

[0019] The current real-time hook route is obtained by recognizing and consistency verification, the warning consistency parameter is obtained, the hook real-time path monitoring and the verification effect of the light rail warning are realized. The warning consistency parameter is used to evaluate the accuracy of the light rail synchronous warning in the current preset time period, and provides feedback basis for the visual recognition parameter configuration of the next preset time period.

[0020] S2, according to the visual data sequence, the hook running feature in the next preset time period is predicted, the predicted hook running route and the predicted hook running feature are obtained.

[0021] Specifically, first, the preconfigured hook running prediction network is called for prediction processing. The hook running prediction network adopts a double branch structure design, including a hook running route prediction branch and a hook running feature prediction branch. The hook running route prediction branch is specially used for predicting the spatial running trajectory of the hook, and the hook running feature prediction branch is used for predicting the motion characteristic parameters of the hook. Both of the two prediction branches take the sample visual data sequence set as the training data, take the sample predicted hook running route set and the sample predicted hook running feature set as the output labels respectively, and are trained by supervised learning until the network configuration is completed.

[0022] Then, the visual data sequence collected in the current preset time period is input as input, and is input into the hook running route prediction branch and the hook running feature prediction branch for parallel prediction processing. The hook running route prediction branch outputs the expected running trajectory of the hook in the next preset time period based on the time sequence continuity and trajectory regularity of the hook motion, and obtains the predicted hook running route. The predicted hook running route contains the expected spatial coordinate information of the hook at each time in the next time period.

[0023] Subsequently, the hook running feature prediction branch outputs the predicted hook running feature based on the motion mode and speed change trend in the visual data sequence. The predicted hook running feature includes the predicted hook running speed and other key motion parameters, reflecting the motion intensity and dynamic characteristics of the hook in the next preset time period.

[0024] By obtaining the predicted hook running route and the predicted hook running feature, the forward-looking perception of the hook running state is realized. The predicted hook running route provides the target trajectory basis for the projection parameter optimization configuration of the next preset time period, and the predicted hook running feature provides the motion characteristic reference for the adaptive adjustment of the visual recognition parameter.

[0025] S3, according to the predicted hook running feature and the warning consistency parameter, the visual recognition parameter configuration of the next preset time period is performed, the predicted visual recognition parameter is obtained, and the visual data recognition configuration in the next preset time period is performed.

[0026] Specifically, first, a preset maximum hooking feature is obtained as a reference benchmark. The maximum hooking feature represents a limit state parameter of the hook operation considered in the design, which is used to standardize the relative strength of the predicted hooking feature. Then, the ratio of the predicted hooking feature to the maximum hooking feature is calculated to obtain a first predicted visual recognition parameter. The first predicted visual recognition parameter reflects the proportion of the expected hook operation strength relative to the design limit in the next preset time period. The larger the ratio, the more complex the expected operation state, and higher visual recognition resource configuration is required.

[0027] Subsequently, a second predicted visual recognition parameter is calculated according to the warning consistency parameter obtained in the current preset time period. The specific calculation method is: second predicted visual recognition parameter = 1 - warning consistency parameter. The calculation principle is that when the warning consistency parameter is low, it indicates that the current visual recognition accuracy is insufficient, and the recognition resource investment needs to be increased in the next period; when the warning consistency parameter is high, it indicates that the current recognition effect is good, and the resource configuration can be appropriately reduced in the next period to optimize the calculation efficiency.

[0028] Next, the final predicted visual recognition parameter is calculated by weighted fusion according to the first predicted visual recognition parameter and the second predicted visual recognition parameter. The predicted visual recognition parameter considers the expected operation complexity in the next period and the performance feedback of the current track synchronous warning, achieving dynamic balance configuration of recognition accuracy and calculation resources.

[0029] By obtaining the predicted visual recognition parameter, the corresponding visual data recognition configuration is performed in the next preset time period, ensuring that the visual recognition can adapt to the expected operation state in the next period, providing a guarantee for accurately obtaining the real-time hooking route.

[0030] S4, according to the predicted hooking route, the projection parameter configuration of the next preset time period is performed to obtain the optimal projection parameter, and the track synchronous warning control of the projection optical module is performed in the next preset time period.

[0031] Specifically, first, the projection parameter of the projection optical module is randomly configured to obtain a first projection parameter. The first projection parameter is a projection coverage area coordinate configuration randomly generated in a preset parameter range, which is used as the initial solution of the optimization algorithm. The purpose of random configuration is to provide a search starting point for subsequent iterative optimization and avoid the optimization process falling into local optimum.

[0032] Then, the intersection area ratio of the first projection parameter and the predicted hooking route is obtained to obtain a first projection fitness. Specifically, the predicted hooking route is projected onto the ground plane to form a hooking route projection area; then the intersection area of the projection coverage area corresponding to the first projection parameter and the hooking route projection area is calculated; and then the projection fitness is obtained by dividing the intersection area by the total area of the hooking route projection area. The greater the value of the projection fitness, the higher the coincidence degree of the projection light track and the predicted hooking route, and the more accurate the expected warning accuracy.

[0033] Subsequently, iterative optimization of the projection parameter is continued, and new projection parameter configurations are continuously generated and the corresponding projection fitness is calculated. Through multiple iteration optimization processes, the projection fitness of different projection parameter configurations is compared to obtain the optimal projection parameter with the maximum projection fitness. The optimal projection parameter ensures that the projection light track can cover the predicted hooking route to the greatest extent, and the best path warning effect is achieved.

[0034] By obtaining the optimal projection parameter, the light track synchronous warning control of the projection optical module is performed in the next preset time period. The projection optical module projects a warning light track at the corresponding position on the ground according to the projection coverage area coordinates in the optimal projection parameter, forming a visual dangerous area identifier. The projected warning light track is consistent with the projection area of the predicted hooking route on the ground. When the hook runs according to the predicted route, the light track on the ground can indicate the running range of the hook synchronously, and the operator can identify the hook running path in advance and avoid in time by observing the ground light track, thereby achieving accurate path warning.

[0035] Further, the visual data sequence in the hook running process is collected by the visual module arranged in the space, and the current real-time hooking route is identified according to the visual recognition parameter decided in the last preset time period, including: S11, collecting the visual data sequence in the hook running process in the current preset time period by the visual module arranged in the space; S12, obtaining a preconfigured visual data recognition network set; S13, calculating the data recognition network quantity according to the visual recognition parameter decided in the last preset time period and the total number of visual data recognition networks in the visual data recognition network set, wherein the visual recognition parameter includes a visual recognition resource coefficient; S14, randomly selecting a visual data recognition network of the data recognition network quantity to determine a plurality of visual data recognition networks, inputting the visual data sequence into the plurality of visual data recognition networks to obtain a plurality of network hooking routes; each network hooking route includes hooking coordinates at a plurality of time points; S15, for each time point, calculating the mean value of the hooking coordinates at the time point in the plurality of network hooking routes to obtain a real-time hooking route.

[0036] In a preferred embodiment, first, the image frames in the running process of the hook are continuously collected in the current preset time period by the visual module arranged in the space to form a visual data sequence. The visual data sequence contains the time sequence information of the position change and motion trajectory of the hook in the current preset time period. Then, a preconfigured visual data recognition network set is obtained. The visual data recognition network set is composed of multiple convolutional neural networks with the same structure but different training data, and is configured by performing supervised training on each network after performing multiple-fold division on a sample visual data sequence set and a sample hook running route set.

[0037] Subsequently, according to the visual recognition parameters decided in the last preset time period, the total number of visual data recognition networks in the visual data recognition network set is combined to calculate the number of data recognition networks required in the current preset time period. The visual recognition parameters include a visual recognition resource coefficient, which reflects the balanced configuration requirement of recognition accuracy and computing resources. The calculation formula of the number of data recognition networks is: number of data recognition networks = visual recognition resource coefficient x total number of visual data recognition networks. Through this calculation method, dynamic allocation of recognition resources is realized. When the expected running complexity is high or the warning consistency parameter is low, the number of called networks is increased to improve the recognition accuracy, and vice versa.

[0038] Next, the visual data recognition network corresponding to the calculated number of data recognition networks is randomly selected from the visual data recognition network set, so as to determine multiple visual data recognition networks. Then, the visual data sequence is input into these selected multiple visual data recognition networks for parallel recognition processing. Each network independently outputs a corresponding network hook running route, and each network hook running route contains hook running coordinate information at multiple time points. Then, the mean value of the hook running coordinates of the multiple network hook running routes at each time point is calculated to obtain a real-time hook running route.

[0039] Through the multi-network integrated recognition mechanism, the stability and accuracy of the hook running route recognition are effectively improved, and reliable path data basis is provided for subsequent consistency verification and parameter optimization.

[0040] Further, the configuration step of the visual data recognition network set comprises: S121, according to the running data of the hook, a sample visual data sequence set of multiple sample preset time periods is collected, and a hook running route under different sample visual data sequences is collected to obtain a sample hook running route set, wherein each sample hook running route includes hook running coordinates at multiple time points; S122, based on a convolutional neural network, a plurality of visual data recognition networks with the same structure are constructed; S123, divide the sample visual data sequence set and the sample hook running route set into multiple folds, respectively supervise and train multiple visual data recognition networks, and obtain a visual data recognition network set after convergence and configuration on the central control platform.

[0041] In a preferred embodiment, first, a sample visual data sequence set of multiple sample preset time periods is collected according to historical operation data of the hook. Each sample preset time period corresponds to a complete set of hook operation image sequences, forming a sample visual data sequence, and the sample visual data sequence set covers different running scenes, speed states and path characteristics. At the same time, the hook running routes corresponding to each sample visual data sequence are collected to obtain a sample hook running route set. Each sample hook running route includes hook running coordinates at multiple time points, recording the actual running trajectory of the hook in the corresponding time period, and providing a standard label for network training.

[0042] Then, multiple visual data recognition networks with the same structure are constructed based on convolutional neural networks. These networks use the same network layer number, convolution kernel configuration and activation function structure parameters to ensure that each network has the same learning ability and processing complexity, laying a consistent foundation for subsequent integrated recognition.

[0043] Subsequently, the sample visual data sequence set and the sample hook running route set are divided into multiple folds. The multiple fold division means that all sample data is divided into multiple different training data subsets according to a preset ratio, and each subset contains part of the sample visual data sequence and its corresponding sample hook running route. Different training data subsets are used to supervise and train multiple visual data recognition networks, and the network parameters are optimized through a back propagation algorithm, so that each network can learn the mapping relationship from the visual data sequence to the hook running route. When the loss function of each network converges to a stable state, the training is completed, and a set of visual data recognition networks with different training data backgrounds is obtained.

[0044] By configuring the set of trained visual data recognition networks on the central control platform, diversified network resources are provided for subsequent real-time hook running route recognition, which can improve the recognition accuracy and robustness through integrated learning mechanism.

[0045] Further, the real-time projection parameters of the projection optical module are verified for consistency to obtain warning consistency parameters, including: S15, obtaining real-time projection parameters of the projection optical module in a current preset time period, wherein the projection parameters include projection coverage area coordinates; S16, calculating an intersection area ratio of the real-time projection parameters and the real-time hook running route to obtain a warning consistency parameter.

[0046] In a preferred embodiment, first, the real-time projection parameters of the projection optical module in the current preset time period are obtained. The projection optical module projects a warning light track on the ground according to the projection parameters configured in the last preset time period. The real-time projection parameters record the spatial distribution information of the current projection light track, wherein the projection parameters include projection coverage area coordinates, which define the specific coverage range of the projection light track in the ground coordinate system, including boundary coordinates, shape parameters and other spatial positioning information of the coverage area.

[0047] Then, the intersection area ratio of the real-time projection parameters and the real-time hooking route is calculated to obtain a warning consistency parameter. Specifically, first, the real-time hooking route is projected onto the ground coordinate system to form a hooking route ground coverage area; then the intersection area of the projection coverage area and the hooking route ground coverage area is calculated; then the total area of the hooking route ground coverage area is calculated; and then the intersection area is divided by the total area of the hooking route ground coverage area to obtain the warning consistency parameter. The value range of the warning consistency parameter is 0-1, and the closer the value is to 1, the higher the coincidence degree of the current projection light track and the actual hooking route, and the more accurate the light track synchronous warning accuracy.

[0048] By obtaining the warning consistency parameter, the quantitative evaluation of the light track synchronous warning effect in the current preset time period is realized, and the performance feedback basis for subsequent optimization of the visual recognition parameters in the next preset time period is provided.

[0049] Further, the hooking feature in the next preset time period is predicted according to the visual data sequence to obtain a predicted hooking route and a predicted hooking feature, including: S21, calling a pre-configured hooking prediction network, wherein the hooking prediction network includes a hooking route prediction branch and a hooking feature prediction branch, the training data are all sample visual data sequence sets, the output data are sample predicted hooking route sets and sample predicted hooking feature sets respectively, and the supervised training is performed until the configuration is completed. S22, inputting the visual data sequence into the pre-configured hooking prediction network to output the predicted hooking route and the predicted hooking feature, wherein the predicted hooking feature includes a predicted hooking speed.

[0050] In a preferred embodiment, first, a pre-configured hooking prediction network is called for prediction processing. The hooking prediction network adopts a double-branch structure design, including a hooking route prediction branch and a hooking feature prediction branch, which respectively predict a predicted hooking route and a predicted hooking feature from the same visual data sequence. In the network configuration stage, the hooking route prediction branch and the hooking feature prediction branch both take a sample visual data sequence set as the training input data, and take a sample predicted hooking route set and a sample predicted hooking feature set as the training output labels respectively. The hooking route prediction branch learns the mapping relationship from the visual data sequence to the spatial trajectory, and the hooking feature prediction branch learns the mapping relationship from the visual data sequence to the motion parameter. The two branches are trained synchronously in a supervised learning manner, and when the network loss function converges to a stable state, the configuration of the hooking prediction network is completed.

[0051] Then, the visual data sequence collected in the current preset time period is input into the pre-configured hooking prediction network to predict the running state in the next preset time period. The hooking route prediction branch analyzes the timing continuity and trajectory regularity of the hook movement based on the input visual data sequence, outputs the expected running trajectory of the hook in the next preset time period, and obtains the predicted hooking route. At the same time, the hooking feature prediction branch analyzes the motion mode and dynamic change trend of the hook based on the same visual data sequence, and outputs the predicted hooking feature. The predicted hooking feature includes the predicted hooking speed and other key motion parameters, reflecting the motion intensity and dynamic characteristics of the hook in the next preset time period.

[0052] Through the double-branch parallel prediction mechanism, the predicted hooking route and the predicted hooking feature are obtained simultaneously, providing a prediction basis for the subsequent adaptive configuration of the visual recognition parameter and the optimization of the projection parameter, and realizing the forward-looking control.

[0053] Further, according to the predicted hooking feature and the warning consistency parameter, the visual recognition parameter configuration in the next preset time period is performed to obtain the predicted visual recognition parameter, and the visual data recognition configuration in the next preset time period is performed, including: S31, obtaining a maximum hooking feature; S32, calculating the ratio of the predicted hooking feature and the maximum hooking feature to obtain a first predicted visual recognition parameter; S33, calculating a second predicted visual recognition parameter according to the warning consistency parameter; S34, calculating the predicted visual recognition parameter according to the first predicted visual recognition parameter and the second predicted visual recognition parameter, and performing the visual data recognition configuration in the next preset time period.

[0054] In a preferred embodiment, first, a preset maximum hook travel feature is obtained. The maximum hook travel feature corresponds to the parameter type of the predicted hook travel feature, specifically the maximum hook travel speed, which is a predefined upper limit value of the speed according to the performance of the hook device and the safety requirements of the operation, and is used as a reference parameter for standardized calculation. Then, the ratio of the predicted hook travel feature and the maximum hook travel feature is calculated to obtain the first predicted visual recognition parameter. Since the predicted hook travel feature includes the predicted hook travel speed, the calculation is the ratio of the predicted hook travel speed and the maximum hook travel speed, i.e. the first predicted visual recognition parameter = predicted hook travel speed / maximum hook travel speed. This ratio reflects the proportion of the expected hook running speed relative to the design limit in the next preset time period. When the predicted hook travel speed is close to the maximum hook travel speed, the ratio is close to 1, indicating that the expected running speed is high and the visual recognition is difficult, requiring more recognition resources; when the predicted hook travel speed is low, the ratio is small, indicating that the recognition is relatively easy, and the resource configuration can be appropriately reduced to optimize the calculation efficiency.

[0055] Subsequently, the second predicted visual recognition parameter is calculated according to the warning consistency parameter obtained in the current preset time period. The calculation formula of the second predicted visual recognition parameter is: second predicted visual recognition parameter = 1 - warning consistency parameter. When the warning consistency parameter is low, it indicates that the current visual recognition accuracy is insufficient, resulting in poor light rail synchronous warning effect, at this time the value of the second predicted visual recognition parameter is large, indicating that the visual recognition resource needs to be increased in the next period to improve the recognition accuracy; when the warning consistency parameter is high, it indicates that the current recognition effect is good, at this time the value of the second predicted visual recognition parameter is small, indicating that the resource configuration can be appropriately reduced in the next period to optimize the calculation efficiency.

[0056] Next, the final predicted visual recognition parameter is calculated by weighted fusion according to the first predicted visual recognition parameter and the second predicted visual recognition parameter. The predicted visual recognition parameter considers the expected running complexity in the next period and the performance feedback of the current light rail synchronous warning, achieving dynamic balance configuration of recognition accuracy and calculation resources. The specific calculation formula is: predicted visual recognition parameter = a x first predicted visual recognition parameter + b x second predicted visual recognition parameter; where a and b are weight coefficients, satisfying a + b = 1. The weight coefficients can be determined according to actual operation experience, for example, a = 0.6, b = 0.4, or a = 0.7, b = 0.3, etc.

[0057] Afterwards, according to the obtained predicted visual recognition parameter, corresponding visual data recognition configuration is carried out in the next preset time period. Specifically, the predicted visual recognition parameter will serve as the visual recognition resource coefficient of the next preset time period, and according to the product of the coefficient and the total number of visual data recognition networks in the visual data recognition network set, the number of recognition networks selected from the visual data recognition network set is determined, realizing dynamic allocation of recognition resources and ensuring that visual recognition can adapt to the expected running state in the next preset time period.

[0058] Further, according to the predicted hooking route, projection parameter configuration of the next preset time period is carried out to obtain optimal projection parameters, and optical track synchronous warning control of the projection optical module in the next preset time period is carried out, including: S41, random projection parameter configuration of the projection optical module is carried out to obtain first projection parameters; S42, the intersection area ratio of the first projection parameters and the predicted hooking route is processed to obtain a projection fitness; S43, iterative optimization of the projection parameters is continued to obtain optimal projection parameters with the maximum projection fitness, and optical track synchronous warning control of the projection optical module in the next preset time period is carried out.

[0059] In a preferred embodiment, first, random projection parameter configuration of the projection optical module is carried out to obtain first projection parameters. The first projection parameters are projection coverage area coordinate configurations randomly generated in a preset parameter range, serving as initial solutions of the optimization algorithm. The purpose of random configuration is to provide a search starting point for subsequent iterative optimization and avoid the optimization process falling into a local optimal solution.

[0060] Then, the intersection area ratio of the first projection parameters and the predicted hooking route is processed to obtain a first projection fitness. Specifically, first, the predicted hooking route is projected to the ground coordinate system to form a coverage area of the predicted hooking route on the ground; then the intersection area of the projection coverage area corresponding to the first projection parameters and the coverage area of the predicted hooking route on the ground is calculated; then the total area of the coverage area of the predicted hooking route on the ground is calculated; and then the intersection area is divided by the total area of the coverage area of the predicted hooking route on the ground to obtain the projection fitness corresponding to the first projection parameters. The greater the projection fitness value, the higher the coincidence degree of the projection track and the predicted hooking route, and the more accurate the expected warning accuracy.

[0061] Subsequently, the iterative optimization process of the projection parameters is continued. By using heuristic optimization methods such as genetic algorithm and particle swarm algorithm, new projection parameter configurations are constantly generated and the corresponding projection fitness is calculated. After multiple iteration optimization processes, the projection fitness of different projection parameter configurations is compared, and the optimal projection parameter with the maximum projection fitness is obtained. The optimal projection parameter ensures that the projection light track can cover the predicted hook travel route to the greatest extent, and realizes the best path warning effect.

[0062] By obtaining the optimal projection parameters, parameter configuration basis is provided for the light track synchronous warning control of the projection optical module in the next preset time period. When the next preset time period starts, the projection optical module projects a warning light track at the corresponding position on the ground according to the projection coverage area coordinates in the optimal projection parameters, and realizes the synchronous warning with the actual running path of the hook.

[0063] Corresponding to the light track synchronous warning method of the above embodiment, Figure 2 A structural block diagram of a light track synchronous warning device is provided for an embodiment of the present application. Only parts related to the embodiments of the present application are shown for ease of description. For reference Figure 2 The light track synchronous warning device 20 comprises: A visual data acquisition module 21 is configured to acquire a visual data sequence in the running process of a hook by a visual module arranged in space, identify the current real-time hook travel route according to the visual recognition parameters determined in the last preset time period, and perform consistency verification with the real-time projection parameters of the projection optical module to obtain warning consistency parameters; A hook travel feature prediction module 22 is configured to predict the hook travel feature in the next preset time period according to the visual data sequence to obtain a predicted hook travel route and a predicted hook travel feature; A visual parameter configuration module 23 is configured to configure visual recognition parameters in the next preset time period according to the predicted hook travel feature and the warning consistency parameters, and to obtain predicted visual recognition parameters for visual data recognition configuration in the next preset time period; A projection parameter configuration module 24 is configured to configure projection parameters in the next preset time period according to the predicted hook travel route, and to obtain optimal projection parameters for light track synchronous warning control of the projection optical module in the next preset time period.

[0064] Further, the visual data acquisition module 21 is further configured to: acquire a visual data sequence in the running process of a hook in the current preset time period by a visual module arranged in space; obtain a preconfigured visual data recognition network set; According to the visual recognition parameter of the last preset time period decision, and in combination with the total number of the visual data recognition network in the visual data recognition network set, the number of data recognition networks is calculated and obtained, wherein the visual recognition parameter includes a visual recognition resource coefficient; A visual data recognition network of the number of data recognition networks is randomly selected, a plurality of visual data recognition networks are determined, the visual data sequence is input into the plurality of visual data recognition networks, and a plurality of network hooking routes are obtained; each network hooking route includes hooking coordinates at a plurality of time points; For each time point, the mean value of the hooking coordinates at the time point in the plurality of network hooking routes is calculated to obtain a real-time hooking route.

[0065] Further, the configuration step of the visual data recognition network set includes: According to the running data of the hook, a plurality of sample visual data sequence sets of a sample preset time period are collected, and hooking routes under different sample visual data sequences are collected to obtain a sample hooking route set, wherein each sample hooking route includes hooking coordinates at a plurality of time points; Based on a convolutional neural network, a plurality of visual data recognition networks with the same structure are constructed; The sample visual data sequence set and the sample hooking route set are divided by multiple folds, and the plurality of visual data recognition networks are supervised trained, and the visual data recognition network set is obtained after convergence and configured in the central control platform.

[0066] Further, the visual data acquisition module 21 is further used for: Obtaining real-time projection parameters of the projection optical module in the current preset time period, wherein the projection parameters include projection coverage area coordinates; Calculating the intersection area proportion of the real-time projection parameters and the real-time hooking route to obtain an alarm consistency parameter.

[0067] Further, the hooking feature prediction module 22 is further used for: Calling a pre-configured hooking prediction network, wherein the hooking prediction network includes a hooking route prediction branch and a hooking feature prediction branch, the training data are all the sample visual data sequence set, the output data are respectively a sample predicted hooking route set and a sample predicted hooking feature set, and the supervised training is performed until the configuration is completed after convergence; The visual data sequence is input into the pre-configured hooking prediction network, and a predicted hooking route and a predicted hooking feature are output and obtained, wherein the predicted hooking feature includes a predicted hooking speed.

[0068] Further, the visual parameter configuration module 23 is further used for: Obtaining the maximum hooking feature; Calculate the ratio of the predicted hook line feature to the maximum hook line feature to obtain the first predicted visual recognition parameter; The second predictive visual recognition parameter is calculated based on the warning consistency parameter. Based on the first and second predicted visual recognition parameters, the predicted visual recognition parameters are calculated, and the visual data recognition configuration is performed in the next preset time period.

[0069] Furthermore, the projection parameter configuration module 24 is also used for: Randomly configure the projection parameters of the projection optical module to obtain the first projection parameters; The proportion of the intersection area between the first projection parameters and the predicted hook route is obtained to obtain the projection fitness. Continue iteratively optimizing the projection parameters to obtain the optimal projection parameters with the greatest projection adaptability, and perform optical track synchronization warning control of the projection optical module in the next preset time period.

[0070] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of each module / unit in the above-described device embodiments, for example... Figure 2 The functions of the visual data acquisition module 21 and the hook feature prediction module 22 are shown.

[0071] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0072] The input device 302 can include a touchpad, a fingerprint collection sensor (for collecting fingerprint information and direction information of a fingerprint of a user), a microphone, etc., and the output device 303 can include a display (LCD, etc.), a speaker, etc.

[0073] The memory 304 can include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 can also include a non-volatile random access memory. For example, the memory 304 can also store device type information.

[0074] In a specific implementation, the processor 301, the input device 302, and the output device 303 described in the embodiments of the present application can execute the implementation manners described in the first and second embodiments of the light track synchronous warning method provided by the embodiments of the present application, and can also execute the implementation manners of the electronic device described in the embodiments of the present application, which will not be described here.

[0075] In another embodiment of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. The computer program includes program instructions, and the program instructions are executed by a processor to implement all or part of the processes of the above-mentioned embodiments. The computer program can also be used to instruct related hardware to complete the processes. The computer program can be stored in a computer readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in a source code form, an object code form, an executable file, or some intermediate form, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0076] The computer readable storage medium can be an internal storage unit of the electronic device of any of the preceding embodiments, such as a hard disk or a memory of the electronic device. The computer readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, or the like. Further, the computer readable storage medium can include both the internal storage unit and the external storage device of the electronic device. The computer readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer readable storage medium can also be used to temporarily store data that has been output or will be output.

[0077] Those skilled in the art can understand that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the foregoing description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0078] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic device and the units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0079] In several embodiments provided in the present application, it should be understood that the disclosed electronic device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces or units, and can also be electrical, mechanical or other forms of connection.

[0080] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0081] In addition, each of the function units in each of the embodiments of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0082] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of track-synchronous alerting, characterized in that The method comprises: Through the visual module arranged in the space, a visual data sequence in the running process of the lifting hook is collected, a current real-time hook running route is recognized according to the visual recognition parameter decided in the last preset time period, consistency verification is performed on the real-time projection parameter of the projection optical module, and a warning consistency parameter is obtained; According to the visual data sequence, a hook running feature in the next preset time period is predicted, a predicted hook running route and a predicted hook running feature are obtained; According to the predicted hook running feature and the warning consistency parameter, the visual recognition parameter configuration in the next preset time period is performed, a predicted visual recognition parameter is obtained, and the visual data recognition configuration in the next preset time period is performed; According to the predicted hook running route, the projection parameter configuration in the next preset time period is performed, an optimal projection parameter is obtained, and the light track synchronous warning control of the projection optical module in the next preset time period is performed.

2. The method of claim 1, wherein the alerting is performed in synchronization with the optical track. Through the visual module arranged in the space, a visual data sequence in the running process of the lifting hook is collected, a current real-time hook running route is recognized according to the visual recognition parameter decided in the last preset time period, consistency verification is performed on the real-time projection parameter of the projection optical module, and a warning consistency parameter is obtained; Through the visual module arranged in the space, a visual data sequence in the running process of the lifting hook is collected, a current real-time hook running route is recognized according to the visual recognition parameter decided in the last preset time period, consistency verification is performed on the real-time projection parameter of the projection optical module, and a warning consistency parameter is obtained; A preconfigured visual data recognition network set is obtained; According to the visual recognition parameter decided in the last preset time period, the total number of visual data recognition networks in the visual data recognition network set is combined to calculate a data recognition network number, wherein the visual recognition parameter comprises a visual recognition resource coefficient; A visual data recognition network of the data recognition network number is randomly selected to determine a plurality of visual data recognition networks, the visual data sequence is input into the plurality of visual data recognition networks, and a plurality of network hook running routes are obtained; each network hook running route comprises hook running coordinates at a plurality of time points; For each time point, the mean value of the hook running coordinates at the time point in the plurality of network hook running routes is calculated to obtain a real-time hook running route.

3. The method of claim 2, wherein the alerting is performed in synchronization with the playing of the optical track. The configuration step of the visual data recognition network set comprises: According to the running data of the lifting hook, a sample visual data sequence set of a plurality of sample preset time periods is collected, and a hook running route under different sample visual data sequences is collected to obtain a sample hook running route set, wherein each sample hook running route comprises hook running coordinates at a plurality of time points; Based on a convolutional neural network, a plurality of visual data recognition networks with the same structure are constructed; The sample visual data sequence set and the sample hook running route set are divided into multiple folds, the plurality of visual data recognition networks are supervised and trained respectively, and the visual data recognition network set is obtained after convergence and is configured in the central control platform.

4. The method of claim 1, wherein the alerting is performed in synchronization with the optical track. Consistency verification is performed on the real-time projection parameter of the projection optical module to obtain a warning consistency parameter, comprising: The real-time projection parameter of the projection optical module in the current preset time period is obtained, wherein the projection parameter comprises a projection coverage area coordinate; The intersection area proportion of the real-time projection parameter and the real-time hook running route is calculated to obtain a warning consistency parameter.

5. The method of claim 1, wherein the alerting is performed in synchronization with the playing of the optical track. According to the visual data sequence, a hook running feature in the next preset time period is predicted, a predicted hook running route and a predicted hook running feature are obtained, comprising: calling a pre-configured hooking prediction network, wherein the hooking prediction network comprises a hooking route prediction branch and a hooking feature prediction branch, the training data are all sample visual data sequence sets, the output data are sample predicted hooking route sets and sample predicted hooking feature sets respectively, and supervised training is performed until convergence is completed; inputting the visual data sequence into the pre-configured hooking prediction network to output a predicted hooking route and a predicted hooking feature, wherein the predicted hooking feature comprises a predicted hooking speed.

6. The method of claim 1, wherein the alerting is performed in synchronization with the playing of the optical track. According to the predicted hooking feature and the warning consistency parameter, the visual recognition parameter configuration in the next preset time period is performed to obtain a predicted visual recognition parameter, and the visual data recognition configuration in the next preset time period is performed, including: obtaining a maximum hooking feature; calculating the ratio of the predicted hooking feature and the maximum hooking feature to obtain a first predicted visual recognition parameter; obtaining a second predicted visual recognition parameter according to the warning consistency parameter; calculating the predicted visual recognition parameter according to the first predicted visual recognition parameter and the second predicted visual recognition parameter, and performing the visual data recognition configuration in the next preset time period.

7. The method of claim 1, wherein the alerting is performed in synchronization with the playing of the optical track. According to the predicted hooking route, the projection parameter configuration in the next preset time period is performed to obtain an optimal projection parameter, and the optical track synchronous warning control of the projection optical module in the next preset time period is performed, including: performing random configuration of the projection parameter of the projection optical module to obtain a first projection parameter; processing the intersection area ratio of the first projection parameter and the predicted hooking route to obtain a projection fitness; continuing to perform iterative optimization of the projection parameter to obtain an optimal projection parameter with the maximum projection fitness, and performing the optical track synchronous warning control of the projection optical module in the next preset time period.

8. A track-synchronous warning device, characterized in that The device for implementing the optical track synchronous warning method according to any one of claims 1 to 7 comprises: a visual data acquisition module configured to acquire visual data sequences in the hook running process through a visual module arranged in space, recognize a current real-time hooking route according to a visual recognition parameter determined in a previous preset time period, and verify the consistency of the real-time projection parameter of the projection optical module to obtain a warning consistency parameter; a hooking feature prediction module configured to predict a hooking feature in the next preset time period according to the visual data sequence to obtain a predicted hooking route and a predicted hooking feature; a visual parameter configuration module configured to configure a visual recognition parameter in the next preset time period according to the predicted hooking feature and the warning consistency parameter to obtain a predicted visual recognition parameter, and perform visual data recognition configuration in the next preset time period; a projection parameter configuration module configured to configure a projection parameter in the next preset time period according to the predicted hooking route to obtain an optimal projection parameter, and perform optical track synchronous warning control of the projection optical module in the next preset time period.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the optical track synchronous warning method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the steps of the optical track synchronous warning method according to any one of claims 1 to 7.

Images