A target extraction method based on cat's eye effect
By collecting and analyzing stable information and processing performance information in the target extraction method, and establishing a time-based analysis model, the problem of lack of real-time self-test function in the existing technology is solved, the timeliness and reliability of the target extraction method is improved, and the Rapid response to emergencies and stability of monitoring systems.
Patent Information
- Application Number
- CN202410296017.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-15
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-03-15
AI Technical Summary
The existing target extraction method based on the cat-eye effect lacks real-time self-test function in road traffic monitoring scenarios, especially in emergency situations, the reliability and adaptability of target vehicle tracking is low, and it is difficult for managers to judge the timeliness of the method. and lack of flexibility in the maintenance inspection cycle.
By collecting the extraction of stable information and processing performance information during the target extraction process, establishing a time-saving analysis model, generating a comprehensive performance index of extracted time-saving performance, comparing and analyzing with the preset threshold, filtering abnormal signals, calculating standard deviations for risk assessment and early warning processing , realize real-time monitoring and evaluation.
Improve the timeliness and reliability of the target extraction method, quickly adapt to emergencies, reduce false detection, missed detection and tracking interruptions, improve the stability and reliability of the monitoring system, and reduce negative consequences caused by failed or delayed target extraction Influence.
Smart Images

Figure CN118135234B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of target extraction, and more specifically, to a target extraction method based on cat's eye effect. Background Art
[0002] The cat's eye effect refers to the phenomenon that when light shines on certain objects, the objects can focus the light and form obvious light spots. This phenomenon is common in nature, such as cat's eyes and insect's compound eyes. The target extraction method based on the cat's eye effect utilizes this principle. By simulating the structure of cat's eyes, the target object is highlighted from the background, and the target extraction technology such as image processing and computer vision is used to realize the recognition and locking of the required target.
[0003] The existing target extraction method based on the cat's eye effect lacks the real-time self-checking function of the method in the monitoring scenario of road traffic, especially for some emergency situations such as the tracking scenario of the target vehicle in the accident, the adaptability to the reliability changes of the target vehicle tracking is low. It is difficult for managers to judge the timeliness status of the target extraction method based on the cat's eye effect during operation through the apparent phenomenon, and there is a lack of flexibility in adjusting the maintenance and inspection cycle of the method.
[0004] In order to solve the above defects, a technical solution is now provided. Summary of the invention
[0005] In order to overcome the above defects of the prior art, an embodiment of the present invention provides a target extraction method based on the cat's eye effect to solve the problems raised in the above background technology.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] A target extraction method based on cat's eye effect comprises the following steps:
[0008] Step S1, collecting extraction stability information and extraction processing performance information when a target extraction method based on the cat's eye effect is used to extract a target;
[0009] Step S2, normalizing the extraction stability information and the extraction processing performance information, establishing a timeliness analysis model, and generating a comprehensive performance index of extraction timeliness performance;
[0010] Step S3, comparing and analyzing the comprehensive performance index of extraction timeliness performance with the comprehensive performance index threshold of extraction timeliness performance, and screening out data whose comprehensive performance index of extraction timeliness performance is greater than or equal to the comprehensive performance index threshold of extraction timeliness performance and marking them as abnormal signals;
[0011] Step S4, establish a data set for several consecutive extraction timeliness performance comprehensive performance indices after the abnormal signal is generated, merge and calculate the standard deviation of several extraction timeliness performance comprehensive performance indices in the data set, and perform risk assessment and early warning processing based on the comparison result of the extraction timeliness performance comprehensive performance index standard deviation and the extraction timeliness performance comprehensive performance index standard deviation threshold.
[0012] In a preferred embodiment, in step S1, the extraction stability information and the extraction processing performance information when the target extraction method based on the cat's eye effect is used to extract the target are collected;
[0013] The extracted stability information includes the target extraction response time floating coefficient and the feature extraction efficiency coefficient. The extracted processing performance information includes the local load comprehensive performance value. After collection, the target extraction response time floating coefficient, the feature extraction efficiency coefficient, and the local load comprehensive performance value are calibrated as KA, QF, and FZ, respectively.
[0014] In a preferred embodiment, the acquisition logic of the target extraction response time floating coefficient KA is as follows:
[0015] Obtain the optimal response time range for target vehicle extraction using a target extraction method based on the cat's eye effect, and calibrate the optimal time range as K1-K2;
[0016] When the target extraction method based on the cat's eye effect is used to extract the target vehicle, several target extraction response durations within T time are obtained, and the target extraction response duration is calibrated as R y ,y represents the number of different target extraction response durations, and y={1,2,3…n}, where n is a positive integer;
[0017] The target extraction response duration greater than the maximum value K2 of the optimal duration range K1-K2 is calibrated as R e , e is the target extraction response duration number greater than the maximum value K2 of the optimal duration range K1-K2, and e={1,2,3…h}, where h is a positive integer;
[0018] The expression for calculating the target extraction response time floating coefficient is: in, is the mean of the optimal response time range, that is
[0019] In a preferred embodiment, the acquisition logic of the feature extraction efficiency coefficient QF is as follows:
[0020] Obtain all feature points extracted from the image data when the target vehicle is extracted using the target extraction method based on the cat's eye effect, and number the extracted feature points as q, q = {1, 2, 3...w}, where w is a positive integer;
[0021] When the target extraction method based on the cat's eye effect is used to extract the target vehicle, the selected feature points are obtained after screening and selecting the feature points extracted from the image data containing the target vehicle, and the selected feature points are numbered as f, f = {1, 2, 3 ... g}, where g is a positive integer;
[0022] The expression for calculating the feature extraction efficiency coefficient is:
[0023] In a preferred embodiment, the logic for obtaining the local load comprehensive performance value FZ is as follows:
[0024] Obtain the load data of the local hardware system when the target extraction method based on the cat's eye effect is used for target extraction, including CPU occupancy, memory occupancy, and the total number of running processes;
[0025] The expression of the comprehensive performance value of local load is as follows Wherein, CP represents CPU occupancy, BR represents memory occupancy, WH represents the total number of running processes, α represents the weight coefficient of CPU occupancy, β represents the weight coefficient of memory occupancy, and γ represents the weight coefficient of the total number of running processes.
[0026] In a preferred embodiment, a time efficiency analysis model is established based on the target extraction response time floating coefficient, the feature extraction efficiency coefficient, and the comprehensive performance value of the local load, and the comprehensive performance index of the extraction time efficiency performance is obtained by weighted summation calculation.
[0027] In a preferred embodiment, the comprehensive performance index of the extraction timeliness performance is compared and analyzed with the comprehensive performance index threshold of the extraction timeliness performance, and an abnormal signal and a stable signal are generated according to the comparative analysis results;
[0028] If the comprehensive performance index of the extraction time-efficiency performance is greater than or equal to the comprehensive performance index threshold of the extraction time-efficiency performance, an abnormal signal is generated;
[0029] If the comprehensive performance index of the extraction time-efficiency performance is less than the comprehensive performance index threshold of the extraction time-efficiency performance, a stable signal is generated.
[0030] In a preferred embodiment, when an abnormal signal is generated, the timeliness state of the target extraction method based on the cat's eye effect is evaluated, and a data set of a number of consecutive extraction timeliness comprehensive performance indexes after the abnormal signal is generated is established and calibrated as LA, then LA = {SX i}, where i = {1, 2, 3…m}, m is a positive integer;
[0031] Calculate the standard deviation of several extraction timeliness performance comprehensive performance indexes in the data set, and mark the standard deviation of the extraction timeliness performance comprehensive performance index as NB, and compare the standard deviation NB of the extraction timeliness performance comprehensive performance index with the preset extraction timeliness performance comprehensive performance index standard deviation threshold YZ, and perform risk warning classification according to the comparison results;
[0032] If the standard deviation of the comprehensive performance index of the extraction time efficiency performance is greater than or equal to the standard deviation threshold of the comprehensive performance index of the extraction time efficiency performance, a high risk level signal is generated;
[0033] If the standard deviation of the comprehensive performance index of extraction timeliness performance is less than the standard deviation threshold of the comprehensive performance index of extraction timeliness performance, a low risk level signal is generated.
[0034] Technical effects and advantages of the present invention:
[0035] 1. The present invention collects stability information and processing performance information during the target extraction process, establishes a time analysis model, monitors and evaluates the time performance of the target extraction method in real time, quickly adapts to emergency situations, better copes with the tracking of vehicles involved in accidents, and ensures the reliability and timeliness of target tracking. By comparing and analyzing the comprehensive performance index of the extraction time performance with a preset threshold, abnormal signals are automatically screened, and the abnormal signals are continuously monitored, and their standard deviations are calculated to evaluate the risk warning level, so that managers can promptly discover and respond to possible performance degradation or failures without relying on subjective judgment, thereby improving the stability and reliability of the entire monitoring system and reducing the negative impact caused by target extraction failure or delay. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;
[0037] Figure 1 This is a schematic diagram of the structure of Example 1 of the present invention. DETAILED DESCRIPTION
[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0039] Example 1
[0040] Figure 1 The present invention provides a target extraction method based on the cat's eye effect, comprising the following steps:
[0041] Step S1, collecting extraction stability information and extraction processing performance information when a target extraction method based on the cat's eye effect is used to extract a target;
[0042] Step S2, normalizing the extraction stability information and the extraction processing performance information, establishing a timeliness analysis model, and generating a comprehensive performance index of extraction timeliness performance;
[0043] Step S3, comparing and analyzing the comprehensive performance index of extraction timeliness performance with the comprehensive performance index threshold of extraction timeliness performance, and screening out data whose comprehensive performance index of extraction timeliness performance is greater than or equal to the comprehensive performance index threshold of extraction timeliness performance and marking them as abnormal signals;
[0044] Step S4, establish a data set for several consecutive extraction timeliness performance comprehensive performance indices after the abnormal signal is generated, merge and calculate the standard deviation of several extraction timeliness performance comprehensive performance indices in the data set, and perform risk assessment and early warning processing based on the comparison result of the extraction timeliness performance comprehensive performance index standard deviation and the extraction timeliness performance comprehensive performance index standard deviation threshold.
[0045] The extraction stability information of the target extraction method based on the cat's eye effect includes the target extraction response time floating coefficient and the feature extraction efficiency coefficient. After collection, the target extraction response time floating coefficient and the feature extraction efficiency coefficient are calibrated as KA and QF respectively.
[0046] When the target extraction method based on the cat's eye effect has poor extraction stability, it may have the following impacts on the tracking of the vehicle causing the accident:
[0047] False detection and missed detection: Target extraction methods with poor stability are prone to false detection and missed detection, that is, incorrectly identifying non-target areas as targets (false detection) or failing to correctly extract target areas (missed detection). This may lead to inaccurate tracking of the vehicle at fault or missing important information.
[0048] Tracking interruption: Target extraction methods with poor stability are easily affected by environmental factors, such as lighting changes, occlusion, etc., which may cause target tracking interruption or loss. If the vehicle that caused the accident suddenly disappears during the tracking process or is blocked by other objects, it may not be found and re-tracked in time.
[0049] Tracking drift: Target extraction methods with poor stability may cause tracking drift, that is, the position and trajectory of the target do not match the actual situation. This may cause other vehicles or objects to be mistaken as the vehicle causing the accident, or the tracked target does not match the actual vehicle causing the accident.
[0050] Performance degradation: Target extraction methods with poor stability may affect the performance and stability of the entire monitoring system, reduce the ability to adapt to complex scenarios and emergency situations, and thus affect the tracking and handling of the vehicle involved in the accident.
[0051] The acquisition logic of the target extraction response time floating coefficient KA is as follows:
[0052] Obtain the optimal response time range for target vehicle extraction using a target extraction method based on the cat's eye effect, and calibrate the optimal time range as K1-K2;
[0053] It should be noted that the performance test of the target extraction method based on the cat's eye effect is carried out, and the optimal response time range for target vehicle extraction based on the cat's eye effect is obtained under different test data. The optimal response time range is not specifically limited here and is adjusted and set according to actual conditions.
[0054] When the target extraction method based on the cat's eye effect is used to extract the target vehicle, several target extraction response durations within T time are obtained, and the target extraction response duration is calibrated as R y ,y represents the number of different target extraction response durations, and y={1,2,3…n}, where n is a positive integer;
[0055] It should be noted that when the target extraction method based on the cat's eye effect is used to extract the target vehicle, the response durations of several target extractions within the T time are obtained according to the monitoring log;
[0056] The target extraction response duration greater than the maximum value K2 of the optimal duration range K1-K2 is calibrated as R e , e is the target extraction response duration number greater than the maximum value K2 of the optimal duration range K1-K2, and e={1,2,3…h}, where h is a positive integer;
[0057] The expression for calculating the target extraction response time floating coefficient is: in, is the mean of the optimal response time range, that is
[0058] From the expression of the target extraction response time floating coefficient, it can be seen that when the target extraction method based on the cat's eye effect is used to extract the target vehicle, the larger the target extraction response time floating coefficient is, the worse the real-time performance of the target extraction method based on the cat's eye effect is when extracting the target vehicle, and the higher the error risk of potential hidden dangers and abnormal warnings when the target extraction method based on the cat's eye effect is used to extract the target vehicle. On the contrary, the better the real-time performance of the target extraction method based on the cat's eye effect is when extracting the target vehicle, the lower the error risk of potential hidden dangers and abnormal warnings when the target extraction method based on the cat's eye effect is used to extract the target vehicle.
[0059] When the target extraction method based on the cat's eye effect is used for target extraction, the feature extraction efficiency is too low, which may cause the following effects on the tracking of the vehicle causing the accident:
[0060] Delayed response: Low feature extraction efficiency will result in failure to identify the target in time, thus delaying the identification and tracking response time of the vehicle involved in the accident. In an emergency, delayed response may result in failure to take timely countermeasures, increasing the danger and consequences of the accident.
[0061] Target loss: Due to the low efficiency of feature extraction, target loss may occur frequently. When the target features cannot be extracted in time, it may cause interruption or error in target tracking, making it impossible to accurately track the location and movement of the vehicle at fault.
[0062] Unstable tracking: Inefficient feature extraction may lead to instability in target tracking. Inaccurate and jittery tracking of the target position may occur frequently, making the tracking result of the vehicle in the accident unstable and difficult to obtain accurate target position and trajectory information.
[0063] Misjudgment and false detection: Due to the low efficiency of feature extraction, misjudgment and false detection may occur. For example, a non-target area may be mistaken for a target, or a target may be missed in a target area, resulting in inaccurate identification and tracking of the vehicle causing the accident.
[0064] The logic for obtaining the feature extraction efficiency coefficient QF is as follows:
[0065] Obtain all feature points extracted from the image data when the target vehicle is extracted using the target extraction method based on the cat's eye effect, and number the extracted feature points as q, q = {1, 2, 3...w}, where w is a positive integer;
[0066] It should be noted that feature points are extracted from the image data containing the target vehicle. These feature points may be features such as the edge, color, texture, shape, etc. of the target vehicle. The specific feature extraction rules are determined according to actual needs and are not limited here.
[0067] When the target extraction method based on the cat's eye effect is used to extract the target vehicle, the selected feature points are obtained after screening and selecting the feature points extracted from the image data containing the target vehicle, and the selected feature points are numbered as f, f = {1, 2, 3 ... g}, where g is a positive integer;
[0068] It should be noted that the selection of feature points is based on the principle of ensuring that the final feature set can represent the information of the target vehicle to the greatest extent, which is set by professionals in this field according to actual conditions and will not be elaborated here;
[0069] The expression for calculating the feature extraction efficiency coefficient is:
[0070] The extraction processing performance information when the target extraction method based on the cat's eye effect is used for target extraction includes a local load comprehensive performance value. After collection, the local load comprehensive performance value is calibrated as FZ;
[0071] When the target extraction method based on the cat's eye effect is used for target extraction, the higher the local hardware load, the following effects may be caused on the tracking of the vehicle causing the accident:
[0072] Delayed response: Excessive local hardware load may result in insufficient processing power to respond and process the target vehicle’s image data in a timely manner. This can result in failure to detect the vehicle in question, delaying the handling and response to the emergency.
[0073] Frame loss: Excessive local hardware load may cause frame loss, that is, the inability to fully process each frame of image data. This may lead to missed detection or misjudgment of the vehicle causing the accident, thus affecting the accurate tracking and identification of the vehicle causing the accident.
[0074] Unstable tracking: Excessive local hardware load may cause unstable operation of the target extraction algorithm. Since image data cannot be processed in time, it may cause jitter and inaccuracy in target tracking, making the tracking effect of the vehicle at fault unstable.
[0075] Performance degradation: Excessive local hardware load may lead to overall performance degradation. It may not be able to meet the needs of real-time processing of target vehicle image data, thus affecting the monitoring of traffic scenes and the ability to handle emergency situations.
[0076] The logic for obtaining the local load comprehensive performance value FZ is as follows:
[0077] Obtain the load data of the local hardware system when the target extraction method based on the cat's eye effect is used for target extraction, including CPU occupancy, memory occupancy, and the total number of running processes;
[0078] It should be noted that the selection of performance monitoring tools can be set by technicians in this field, such as top, Prometheus, and Grafana, which can provide detailed information about memory usage;
[0079] The expression of the comprehensive performance value of local load is as follows: Where CP represents CPU occupancy, BR represents memory occupancy, WH represents the total number of running processes, α represents the weight coefficient of CPU occupancy, β represents the weight coefficient of memory occupancy, and γ represents the weight coefficient of the total number of running processes;
[0080] Among them, the weight coefficient α of CPU occupancy, the weight coefficient β of memory occupancy, and the weight coefficient γ of the total number of running processes are further determined by historical data, suggestions from experts in this field, and by collecting a large amount of historical data for software and related algorithm simulation.
[0081] According to the target extraction method based on the cat's eye effect, the extraction stability information and extraction processing performance information during target extraction are used to establish a time-effectiveness analysis model and generate a comprehensive performance index SX of extraction time-effectiveness performance. The calculation expression based on it is as follows: SX = a 1 *KA+a 2 *QF+a 3 *FZ, where a 1 、a 2 、a 3 、a 4 are the target extraction response time floating coefficient, feature extraction efficiency coefficient, and the preset proportional coefficient of the comprehensive performance value of the local load, respectively, and a 1 、a 3 Greater than 0, a 2 Less than 0;
[0082] It can be seen from the above calculation formula that the larger the target extraction response time floating coefficient, the smaller the feature extraction efficiency coefficient, and the larger the local load comprehensive performance value, the larger the extraction time efficiency performance comprehensive performance index, indicating that the real-time performance of the target extraction method based on the cat's eye effect is worse when the target is extracted, and the higher the risk of untimely tracking of the target vehicle. On the contrary, the smaller the target extraction response time floating coefficient, the larger the feature extraction efficiency coefficient, and the smaller the local load comprehensive performance value, the smaller the extraction time efficiency performance comprehensive performance index, indicating that the real-time performance of the target extraction method based on the cat's eye effect is better when the target is extracted, and the lower the risk of untimely tracking of the target vehicle.
[0083] Compare and analyze the comprehensive performance index of extraction timeliness performance with the comprehensive performance index threshold of extraction timeliness performance, and generate abnormal signals and stable signals according to the comparative analysis results;
[0084] If the comprehensive performance index of extraction time efficiency is greater than or equal to the comprehensive performance index threshold of extraction time efficiency, it indicates that the worse the real-time performance of the target extraction method based on the cat's eye effect is when extracting the target, the higher the risk of untimely tracking of the target vehicle is, and an abnormal signal is generated;
[0085] If the comprehensive performance index of extraction time efficiency is less than the comprehensive performance index threshold of extraction time efficiency, it indicates that the real-time performance of the target extraction method based on the cat's eye effect is higher when performing target extraction, the risk of untimely tracking of the target vehicle is lower, and a stable signal is generated;
[0086] When an abnormal signal is generated, the timeliness status of the target extraction method based on the cat's eye effect is evaluated, and a data set is established for the comprehensive performance index of the extraction timeliness performance after the abnormal signal is generated, and the data is combined and calibrated as LA, then LA={SX i}, where i = {1, 2, 3…m}, m is a positive integer;
[0087] Calculate the standard deviation of several extraction timeliness performance comprehensive performance indexes in the data set, and mark the standard deviation of the extraction timeliness performance comprehensive performance index as NB, and compare the standard deviation NB of the extraction timeliness performance comprehensive performance index with the preset extraction timeliness performance comprehensive performance index standard deviation threshold YZ, and perform risk warning classification according to the comparison results;
[0088] If the standard deviation of the comprehensive performance index of extraction timeliness performance is greater than or equal to the standard deviation threshold of the comprehensive performance index of extraction timeliness performance, it indicates that the real-time performance status of the target extraction method based on the cat's eye effect at the current stage is continuously decreasing, and a high-risk level signal is generated, prompting the staff to have serious timeliness risk hazards for the target extraction method based on the cat's eye effect at the current stage, and maintenance is required;
[0089] If the standard deviation of the comprehensive performance index of extraction time efficiency is less than the standard deviation threshold of the comprehensive performance index of extraction time efficiency, it indicates that the real-time performance status of the target extraction method based on the cat's eye effect at the current stage has not been significantly and continuously deteriorated, and a low-risk level signal is generated, and there is no need to prompt the staff to perform maintenance on the target extraction method based on the cat's eye effect at the current stage;
[0090] The present invention collects stability information and processing performance information during the target extraction process and establishes a time analysis model to monitor and evaluate the time performance of the target extraction method in real time, quickly adapt to emergency situations, better cope with the tracking of the vehicle causing the accident, and ensure the reliability and timeliness of target tracking. By comparing and analyzing the comprehensive performance index of the extraction time performance with a preset threshold, abnormal signals are automatically screened, and the abnormal signals are continuously monitored, and their standard deviations are calculated to evaluate the risk warning level, so that managers can promptly discover and respond to possible performance degradation or failures without relying on subjective judgment, thereby improving the stability and reliability of the entire monitoring system and reducing the negative impact caused by target extraction failure or delay.
[0091] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.
[0092] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.
[0093] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0094] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A target extraction method based on cat's eye effect, characterized in that: The steps include: Step S1, collecting extraction stability information and extraction processing performance information when a target extraction method based on the cat's eye effect is used to extract a target; In step S1, the extraction stability information and the extraction processing performance information when the target extraction method based on the cat's eye effect is used to extract the target are collected; The extracted stability information includes the target extraction response time floating coefficient and the feature extraction efficiency coefficient, and the extracted processing performance information includes the local load comprehensive performance value. After collection, the target extraction response time floating coefficient, the feature extraction efficiency coefficient, and the local load comprehensive performance value are calibrated as KA, QF, and FZ respectively; The acquisition logic of the target extraction response time floating coefficient KA is as follows: Obtain the optimal response time range for target vehicle extraction using a target extraction method based on the cat's eye effect, and calibrate the optimal time range as K1-K2; When the target extraction method based on the cat's eye effect is used to extract the target vehicle, several target extraction response durations within T time are obtained, and the target extraction response duration is calibrated as R y ,y represents the number of different target extraction response durations, and y={1,2,3…n}, where n is a positive integer; The target extraction response duration greater than the maximum value K2 of the optimal duration range K1-K2 is calibrated as R e , e is the target extraction response duration number greater than the maximum value K2 of the optimal duration range K1-K2, and e=1,2,3...h}, where h is a positive integer; The expression for calculating the target extraction response time floating coefficient is: in, is the mean of the optimal response time range, that is The logic for obtaining the feature extraction efficiency coefficient QF is as follows: Obtain all feature points extracted from the image data when the target vehicle is extracted using the target extraction method based on the cat's eye effect, and number the extracted feature points as q, q = {1, 2, 3...w}, where w is a positive integer; When the target extraction method based on the cat's eye effect is used to extract the target vehicle, the selected feature points are obtained after screening and selecting the feature points extracted from the image data containing the target vehicle, and the selected feature points are numbered as f, f = {1, 2, 3 ... g}, where g is a positive integer; The expression for calculating the feature extraction efficiency coefficient is: The logic for obtaining the local load comprehensive performance value FZ is as follows: Obtain the load data of the local hardware system when the target extraction method based on the cat's eye effect is used for target extraction, including CPU occupancy, memory occupancy, and the total number of running processes; The expression of the comprehensive performance value of local load is as follows Where CP represents CPU occupancy, BR represents memory occupancy, WH represents the total number of running processes, α represents the weight coefficient of CPU occupancy, β represents the weight coefficient of memory occupancy, and γ represents the weight coefficient of the total number of running processes; Step S2, normalizing the extraction stability information and the extraction processing performance information, establishing a timeliness analysis model, and generating a comprehensive performance index of extraction timeliness performance; According to the target extraction response time floating coefficient, feature extraction efficiency coefficient, and local load comprehensive performance value, a time efficiency analysis model is established, and the comprehensive performance index of extraction time efficiency performance is obtained by weighted summation calculation. The calculation expression based on it is as follows: SX = a1*KA+a2*QF+a3*FZ, where a1, a2, and a3 are preset proportional coefficients of the target extraction response time floating coefficient, feature extraction efficiency coefficient, and local load comprehensive performance value, respectively, and a1 and a3 are greater than 0, and a2 is less than 0; Step S3, comparing and analyzing the comprehensive performance index of extraction timeliness performance with the comprehensive performance index threshold of extraction timeliness performance, and screening out data whose comprehensive performance index of extraction timeliness performance is greater than or equal to the comprehensive performance index threshold of extraction timeliness performance and marking them as abnormal signals; Step S4, establish a data set for several consecutive extraction timeliness performance comprehensive performance indices after the abnormal signal is generated, merge and calculate the standard deviation of several extraction timeliness performance comprehensive performance indices in the data set, and perform risk assessment and early warning processing based on the comparison result of the extraction timeliness performance comprehensive performance index standard deviation and the extraction timeliness performance comprehensive performance index standard deviation threshold.
2. The target extraction method based on cat's eye effect according to claim 1, characterized in that: Compare and analyze the comprehensive performance index of extraction timeliness performance with the comprehensive performance index threshold of extraction timeliness performance, and generate abnormal signals and stable signals according to the comparative analysis results; If the comprehensive performance index of the extraction time-efficiency performance is greater than or equal to the comprehensive performance index threshold of the extraction time-efficiency performance, an abnormal signal is generated; If the comprehensive performance index of the extraction time-efficiency performance is less than the comprehensive performance index threshold of the extraction time-efficiency performance, a stable signal is generated.
3. The target extraction method based on cat's eye effect according to claim 2, characterized in that: When an abnormal signal is generated, the timeliness status of the target extraction method based on the cat's eye effect is evaluated, and a data set is established for the comprehensive performance index of the extraction timeliness performance after the abnormal signal is generated, and the data is combined and calibrated as LA, then LA={SX i }, where i = {1, 2, 3…m}, m is a positive integer; Calculate the standard deviation of several extraction timeliness performance comprehensive performance indexes in the data set, and mark the standard deviation of the extraction timeliness performance comprehensive performance index as NB, and compare the standard deviation NB of the extraction timeliness performance comprehensive performance index with the preset extraction timeliness performance comprehensive performance index standard deviation threshold YZ, and perform risk warning classification according to the comparison results; If the standard deviation of the comprehensive performance index of the extraction time efficiency performance is greater than or equal to the standard deviation threshold of the comprehensive performance index of the extraction time efficiency performance, a high risk level signal is generated; If the standard deviation of the comprehensive performance index of extraction timeliness performance is less than the standard deviation threshold of the comprehensive performance index of extraction timeliness performance, a low risk level signal is generated.
Citation Information
Patent Citations
Tunnel advanced geological forecast risk evaluation method and device
CN115640995A
Automatic weighting load balancing method and system, electronic equipment and storage medium
CN116016533A
Efficient data protection system based on hybrid encryption technology
CN117353924A
Water conservancy management system based on digital twinning
CN117455242A
Image recognition method and system based on computer network and storage medium
CN117523299A
