Target segmentation method and system for feature intensity training
By obtaining the basic information and video data sets of the target lane, feature extraction and image segmentation are performed, the problems of low target segmentation efficiency and waste of human resources in the prior art are solved, and the effect of rapid analysis of lane abnormalities is achieved.
Patent Information
- Application Number
- CN202411260756.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-11-09
- Filing Date
- 2024-09-10
- Publication Date
- 2024-12-27
AI Technical Summary
The existing target segmentation method is slower, which causes manpower and material resources to be consumed when analyzing lane abnormalities.
By determining the target lane, obtaining basic information and video data sets, performing feature extraction and image segmentation, analyzing abnormal situations, improving target segmentation efficiency and saving human resources.
It achieves rapid analysis of lane anomalies, improves target segmentation efficiency, and saves human resources.
Smart Images
Figure CN119206574A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of target segmentation, and in particular to a target segmentation method and system for feature intensity training. Background Art
[0002] The characteristics of features are information content, distinctiveness and independence. The idea behind it is to represent complex things through an abstract and simplified mathematical concept. Each thing has different characteristics. The so-called image segmentation refers to dividing the image into several non-intersecting areas based on grayscale, color, spatial texture, geometric shape and other features, so that these features show consistency or similarity in the same area, and show obvious differences between different areas. Simply put, in an image, separate the target from the background. Since the current target segmentation method is inefficient and the abnormal query and judgment of the lane consumes manpower and material resources, a target segmentation method for feature intensity training is proposed.
[0003] In summary, the existing technology has the problems of low efficiency in target segmentation and waste of manpower and material resources in analyzing lane anomalies. Summary of the invention
[0004] Based on this, it is necessary to provide a target segmentation method and system for feature intensity training that can improve the efficiency of target segmentation and save human resources in order to address the above technical problems.
[0005] In a first aspect, the present application provides a target segmentation method for feature intensity training, the method comprising: determining a target lane, connecting to a traffic management system to obtain basic information of the target lane, and obtaining a basic information set; determining a target monitoring port of the target lane based on the basic information, and obtaining a corresponding video data set in real time; inputting the video data set into a feature extraction module for feature extraction to obtain a target lane feature set; performing image segmentation on the video data set based on the target lane feature set to obtain a feature image set of the target lane; and analyzing the feature image set to obtain abnormal conditions of the target lane.
[0006] In the second aspect, the present application provides a target segmentation system for feature intensity training, the system comprising: a basic information set acquisition module, the basic information set acquisition module is used to determine the target lane, connect to the traffic management system to obtain basic information of the target lane, and obtain a basic information set; a video data set acquisition module, the video data set acquisition module is used to determine the target monitoring port of the target lane according to the basic information, and obtain the corresponding video data set in real time; a target lane feature set acquisition module, the target lane feature set acquisition module is used to input the video data set into a feature extraction module for feature extraction, and obtain a target lane feature set; a feature image set acquisition module, the feature image set acquisition module is used to perform image segmentation on the video data set according to the target lane feature set, and obtain a feature image set of the target lane; an abnormal situation acquisition module, the abnormal situation acquisition module is used to analyze the feature image set to obtain the abnormal situation of the target lane.
[0007] The above one or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:
[0008] First, determine the target lane, connect to the traffic management system to obtain the basic information of the target lane, and obtain the basic information set; secondly, determine the target monitoring port of the target lane based on the basic information, and obtain the corresponding video data set in real time; then input the video data set into the feature extraction module for feature extraction to obtain the target lane feature set; then perform image segmentation on the video data set based on the target lane feature set to obtain the feature image set of the target lane; finally, analyze the feature image set to obtain the abnormal situation of the target lane. This application solves the technical problems in the prior art that the speed of target segmentation is slow and the analysis of lane abnormalities consumes manpower and material resources, and achieves the technical effect of improving the efficiency of target segmentation and saving human resources.
[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0011] Figure 1 is a flow chart of a target segmentation method for feature strength training in one embodiment;
[0012] Figure 2 A schematic diagram of a process for screening feature intensity coefficients of a target segmentation method for feature intensity training in one embodiment;
[0013] Figure 3 1 is a block diagram of a target segmentation system for feature strength training in one embodiment.
[0014] Explanation of reference numerals: basic information set acquisition module 11 , video data set acquisition module 12 , target lane feature set acquisition module 13 , feature image set acquisition module 14 , abnormal situation acquisition module 15 . DETAILED DESCRIPTION
[0015] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0016] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0017] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.
[0018] like Figure 1 As shown, the present application provides a target segmentation method for feature strength training, the method comprising:
[0019] Determine the target lane, connect to the traffic management system to obtain basic information of the target lane, and obtain a basic information set;
[0020] A feature refers to the characteristics of a thing that are different from other things. The feature intensity in this application refers to the area ratio of the features in the target lane that are different from other parts; target segmentation refers to dividing an image into several non-overlapping areas based on features such as grayscale, color, spatial texture, and geometric shape, so that these features show consistency or similarity in the same area, but show obvious differences between different areas. In this application, by providing a target segmentation method for feature intensity training, the technical problem of the target segmentation method in the prior art consuming manpower and material resources during processing is solved, thereby achieving the technical effect of saving human resources.
[0021] The target lane refers to a lane selected by the staff as a target for research in the infrastructure for various trackless vehicles and pedestrians, which is recorded as the target lane; the traffic management system refers to a system that provides the traffic management department and the driver with the functions of real-time guidance, control and emergency response to emergencies of road traffic flow through advanced monitoring, control and information processing subsystems. The monitoring and information processing subsystems are mainly used in this application; the basic information of the target lane refers to the length, width and height of the target lane measured after the completion of the target lane, the infrastructure of the target lane such as flower beds, etc. The basic information is sorted and analyzed to obtain the basic information set of the target lane. By confirming the target lane, connecting to the traffic management system to query the basic information of the target lane, and integrating the basic information of the target lane to obtain the basic information set of the target lane, data support is provided for subsequent abnormal comparison of the target lane.
[0022] Determine the target lane target monitoring port according to the basic information, and obtain the corresponding video data set in real time;
[0023] The target monitoring port refers to a plurality of cameras arranged at different positions on the target lane, and the target monitoring camera refers to a monitoring camera arranged perpendicular to the ground, which can clearly obtain a real-time top view of the target lane; real-time refers to doing something at the same time during the occurrence and development of an event, and in this application, it refers to monitoring the target road at the same time when it is necessary to monitor the target road; the video data set refers to the collection obtained after the video of the target lane taken by the target monitoring port is integrated. Through the basic information of the target lane, all the monitoring ports of the target lane are queried, and according to the set conditions, the monitoring cameras that are not arranged perpendicular to the ground are selected to screen the monitoring ports, and all the monitoring ports arranged perpendicular to the ground are obtained, recorded as the target monitoring ports, and then the target monitoring ports are opened synchronously to obtain the video data set corresponding to each target monitoring port. The target monitoring port of the target lane is determined according to the basic information and the set conditions, and the target monitoring port is opened synchronously to obtain the video data set corresponding to each target monitoring port in real time. By obtaining the video data set, it contributes to the subsequent acquisition of the feature set of the target lane.
[0024] Inputting the video data set into a feature extraction module to extract features and obtain a target lane feature set;
[0025] Feature extraction starts with an initial set of measurement data and establishes derived values (features) that are intended to provide information and are non-redundant, thereby facilitating subsequent learning and generalization steps, and in some cases, leading to better interpretability. Feature extraction is related to dimensionality reduction, which refers to reducing the dimensions of data features by retaining some relatively important features and removing some redundant features. A feature extraction module refers to a system that performs feature extraction operations on the video data set, obtains the features of the target lane in the video data set according to the feature extraction module, and integrates the features to obtain a target lane feature set. The video data set is input into a feature extraction module, feature extraction is performed, all features of the target lane are output, the features are analyzed and sorted, and the target lane feature set is obtained. By obtaining the target lane feature set, data support is provided for subsequent image segmentation.
[0026] Set feature points;
[0027] Constructing a feature extraction module according to the feature points;
[0028] The video data set is input into a feature extraction module, and all features contained in the video data set are output to obtain the target lane feature set.
[0029] In this application, feature points refer to the characteristics of the target lane that are different from other places, such as carriageways, sidewalks, streets, flower beds, etc., which are all different feature points. The feature point setting here refers to finding all the feature points of the lane according to the traffic management system, building a feature extraction module according to the feature points, inputting the video data set into the feature extraction module, traversing the feature points, screening out all the features of the video data set, and outputting all the features contained in the video data set to obtain the target lane feature set. By setting feature points, a feature extraction module is established according to the feature points, and then the target lane feature set is obtained according to the feature extraction module. By obtaining the target lane feature set, the subsequent image segmentation is paved.
[0030] like Figure 2 As shown, set the feature intensity threshold;
[0031] Performing feature strength weight allocation according to the target lane feature set to obtain a feature strength coefficient;
[0032] The feature intensity coefficient is compared with the feature intensity threshold, and the feature intensity coefficient that does not meet the feature intensity threshold is screened out.
[0033] The feature intensity threshold refers to the data set by the staff based on expert experience and feature area ratio, which is used to determine whether the feature of the target lane can be excluded or divided with other features at the same time; weight refers to the importance of a factor or indicator relative to a certain thing, which is different from the general proportion. It reflects not only the percentage of a factor or indicator, but emphasizes the relative importance of the factor or indicator, tending to contribution or importance. Usually, the weight can be judged and calculated by dividing multiple levels of indicators. In this application, the weight is allocated according to the feature intensity of the target lane feature set. The feature intensity refers to the size of the feature area in the target lane feature set to the total area. According to the traffic management system, the feature intensity coefficient is obtained, where the feature intensity coefficient refers to the feature weight in the target lane feature set; the feature intensity coefficient is compared with the feature intensity threshold, and the feature intensity coefficient that does not meet the feature intensity threshold is screened out, that is, when the feature area in the target lane feature set accounts for too small a total area, the feature corresponding to the feature intensity coefficient can be selected. By screening the feature intensity coefficient that does not meet the feature intensity threshold, support is provided for subsequent image segmentation.
[0034] Performing image segmentation on the video data set according to the target lane feature set to obtain a feature image set of the target lane;
[0035] Image segmentation refers to the technology and process of dividing an image into regions with corresponding characteristics and extracting the target of interest. The characteristics can be grayscale, color, texture, etc. The target can correspond to a single region or multiple regions. The video data set is segmented according to the target lane feature set to obtain multiple feature images of the target lanes, and the feature images are classified and sorted according to the features of the feature images to obtain the feature image set. The feature image set of the target lane is obtained through the target lane feature set, which contributes to the subsequent acquisition of abnormal conditions of the target lane.
[0036] Collecting a historical video data set, and using the historical video data set and the target lane feature set as a sample data set;
[0037] Establishing a target segmentation model, inputting the historical video data set into the target segmentation model, performing image segmentation according to the target lane feature set, and obtaining the feature image set;
[0038] The target segmentation model is trained and verified by using the sample data set to obtain the target segmentation model.
[0039] The process of constructing the target segmentation model is as follows: construct the network structure of the target segmentation model, which includes multiple simple units that simulate human brain neurons. The target segmentation model can form parameters such as weights and thresholds for the connection between simple units during supervised training. After training, the target segmentation model can perform complex nonlinear logical operations based on the input data and output the data after the operation. The input data of the target segmentation model is the historical video data set, and the output data is the feature image set. Obtain the historical video data set and construct a target lane feature set, wherein the historical video data set includes the historical video data set and the target lane feature set, wherein the sample data set refers to the video data set of the target monitoring port in the past time period; further annotate the sample data set and divide it according to a certain ratio to obtain a training data set and a verification data set. Train and verify the target segmentation model through the sample data set to obtain the target segmentation model.
[0040] Obtaining a preset data division ratio, and dividing the sample data set into a training data set and a verification data set according to the preset data division ratio;
[0041] Performing supervised training on the target segmentation model using the training data set, and when the model output result tends to converge, verifying the output result of the target segmentation model using the verification data set;
[0042] A preset model verification accuracy index is obtained, and when the accuracy of the output result of the target segmentation model meets the preset model verification accuracy index, the target segmentation model is obtained.
[0043] The convergence function tends to infinity (including infinitesimal or infinite), and the function always approaches a certain value, which is called the convergence of the function; construct a target segmentation model, which is a neural network model that can be continuously iterated and optimized in machine learning, and is obtained through supervised training of the training data set. According to the preset data division rule, the sample data set is divided into a sample training set and a sample verification set. The preset data division ratio can be customized by technicians in this field based on actual conditions, for example: 85%, 15%. The target segmentation model is supervised and trained through the sample training set. When the model output result tends to the convergence state, the output result accuracy of the target segmentation model is verified through the sample verification set to obtain a preset verification accuracy index. The preset verification accuracy index can be customized by technicians in this field based on actual conditions, for example: 95%. When the accuracy of the output result of the target segmentation model is greater than or equal to the preset verification accuracy index, the target segmentation model is obtained. By constructing the target segmentation model, the accuracy of the feature image set after segmentation can be improved.
[0044] The feature image set is analyzed to obtain abnormal conditions of the target lane.
[0045] According to the feature image set, the basic information set and the video data set of the target lane are compared, and the abnormal processing information library is traversed to find the image and processing solution corresponding to the feature image set, so as to judge the abnormal situation of the target lane and find the abnormal situation of the target lane accurately and labor-savingly.
[0046] Collecting historical image information of the target lane, and obtaining a corresponding solution according to the historical image information;
[0047] Establishing an exception handling information library according to the corresponding solution;
[0048] Traverse the feature image set through the exception handling information library to determine whether the feature image set exists;
[0049] If so, it is determined that the target lane is abnormal;
[0050] If not present, it is transmitted to the traffic management system for manual comparison.
[0051] Historical image information refers to images that have appeared in the target lane in the past. The historical image information is used to determine whether the state of the target lane at that time is normal or abnormal. The corresponding solution for the normal state of the target lane is unified to maintain the status quo, and the abnormal state corresponds to the historical solution one by one. According to the corresponding solution, an abnormality processing information library is established, and the feature image set is queried and compared in the abnormality processing information library to determine whether the feature image set exists. If it exists, it is determined that the target lane is abnormal, and the target lane is processed according to the abnormality processing information library. If it does not exist, it is transmitted to the traffic management system for manual comparison. Because the example images collected in the abnormality processing information library are limited, it is necessary to manually confirm whether the situation of the target lane is no abnormality or an unrecorded abnormality. By traversing the feature image set through the abnormality processing information library and determining whether the feature image set exists, it contributes to saving human resources.
[0052] Transmitting the feature image set to the traffic management system for manual processing to obtain a processing result;
[0053] If the processing result is that the target lane has no abnormality, the feature image set has no abnormality;
[0054] If the processing result is that an abnormality occurs in the target lane, the video data set corresponding to the feature image set and the processing result are added to the abnormality processing information library.
[0055] Manual processing refers to the handling, arrangement, and processing of the feature image set by the staff, and the processing result of the feature image set is obtained based on manual judgment. If the processing result is that there is no abnormality in the target lane, the feature image set is considered to have no abnormality; if the processing result is that there is an abnormality in the target lane, it is manually identified and processed according to the abnormality. After the processing is completed, it is integrated according to the processing solution corresponding to the feature image set, and added to the abnormality processing information library to update the abnormality processing information library. The present application solves the technical problems in the prior art that the speed of target segmentation is slow and the analysis of lane abnormalities consumes manpower and material resources, and achieves the technical effect of improving the efficiency of target segmentation and saving human resources.
[0056] like Figure 3 As shown, the embodiment of the present application also provides a target segmentation system for feature strength training, the system comprising:
[0057] A basic information set acquisition module 11, the basic information set acquisition module 11 is used to determine the target lane, connect to the traffic management system to obtain basic information of the target lane, and obtain a basic information set;
[0058] A video data set acquisition module 12, the video data set acquisition module 12 is used to determine the target lane target monitoring port according to the basic information, and acquire the corresponding video data set in real time;
[0059] A target lane feature set acquisition module 13, wherein the target lane feature set acquisition module 13 is used to input the video data set into a feature extraction module for feature extraction to acquire a target lane feature set;
[0060] A feature image set acquisition module 14, wherein the feature image set acquisition module 14 is used to perform image segmentation on the video data set according to the target lane feature set to obtain the feature image set of the target lane;
[0061] Abnormal situation acquisition module 15, the abnormal situation acquisition module 15 is used to analyze the feature image set to obtain the abnormal situation of the target lane.
[0062] Furthermore, the embodiment of the present application also includes:
[0063] A feature point setting module, wherein the feature point setting module is used to set feature points;
[0064] A feature extraction module, wherein the feature extraction module is used to construct a feature extraction module according to the feature points;
[0065] The target lane feature set acquisition module is used to input the video data set into the feature extraction module, output all features contained in the video data set, and obtain the target lane feature set.
[0066] Furthermore, the embodiment of the present application also includes:
[0067] A feature strength threshold setting module, wherein the feature strength threshold setting module is used to set a feature strength threshold;
[0068] A feature strength coefficient acquisition module, the feature strength coefficient acquisition module is used to perform feature strength weight allocation according to the target lane feature set to obtain a feature strength coefficient;
[0069] A feature intensity coefficient screening module is used to compare the feature intensity coefficient with the feature intensity threshold and screen out the feature intensity coefficient that does not meet the feature intensity threshold.
[0070] Furthermore, the embodiment of the present application also includes:
[0071] A historical video data set acquisition module, wherein the historical video data set acquisition module is used to acquire a historical video data set, and use the historical video data set and the target lane feature set as a sample data set;
[0072] A target segmentation model building module, wherein the target segmentation model building module is used to build a target segmentation model, input the historical video data set into the target segmentation model, perform image segmentation according to the target lane feature set, and obtain the feature image set;
[0073] The target segmentation model training and verification module is used to train and verify the target segmentation model through the sample data set to obtain the target segmentation model.
[0074] Furthermore, the embodiment of the present application also includes:
[0075] A preset data partition ratio acquisition module, the preset data partition ratio acquisition module is used to obtain a preset data partition ratio, and divide the sample data set into a training data set and a verification data set according to the preset data partition ratio;
[0076] A target segmentation model output result verification module, wherein the target segmentation model output result verification module is used to perform supervised training on the target segmentation model through the training data set, and when the model output result tends to converge, the output result of the target segmentation model is verified through the verification data set;
[0077] The target segmentation model acquisition module is used to obtain a preset model verification accuracy index. When the accuracy of the output result of the target segmentation model meets the preset model verification accuracy index, the target segmentation model is obtained.
[0078] Furthermore, the embodiment of the present application also includes:
[0079] A historical image information acquisition module, the historical image information acquisition module is used to acquire historical image information of the target lane and obtain a corresponding solution according to the historical image information;
[0080] An exception handling information base establishment module, the exception handling information base establishment module is used to establish an exception handling information base according to the corresponding solution;
[0081] A feature image set judgment module, the feature image set judgment module is used to traverse the feature image set through the exception processing information library to determine whether the feature image set exists;
[0082] A target lane abnormality determination module, the target lane abnormality determination module is used to determine if the target lane is abnormal, if any;
[0083] A manual comparison module is used to transmit the information to the traffic management system for manual comparison if it does not exist.
[0084] Furthermore, the embodiment of the present application also includes:
[0085] A feature image set transmission module, the feature image set transmission module is used to transmit the feature image set to the traffic management system for manual processing to obtain a processing result;
[0086] A feature image set determination module, wherein if the processing result shows that the target lane has no abnormality, the feature image set determination module is used to determine that there is no abnormality in the feature image set;
[0087] The corresponding processing result adding module is used to add the video data set corresponding to the feature image set and the processing result to the abnormality processing information library if the processing result is that the target lane is abnormal.
[0088] For the specific embodiment of the target segmentation system for feature strength training, please refer to the embodiment of the target segmentation method for feature strength training mentioned above, which will not be repeated here. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0089] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0090] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.
Claims
1. A target segmentation method for feature strength training, characterized in that: The method comprises: Determine the target lane, connect to the traffic management system to obtain basic information of the target lane, and obtain a basic information set; Determine the target lane target monitoring port according to the basic information, and obtain the corresponding video data set in real time; Inputting the video data set into a feature extraction module to extract features and obtain a target lane feature set; Performing image segmentation on the video data set according to the target lane feature set to obtain a feature image set of the target lane; The feature image set is analyzed to obtain abnormal conditions of the target lane.
2. The method according to claim 1, characterized in that The video data set is input into a feature extraction module for feature extraction to obtain a target lane feature set. The method further includes: Set feature points; Constructing a feature extraction module according to the feature points; The video data set is input into a feature extraction module, and all features contained in the video data set are output to obtain the target lane feature set.
3. The method according to claim 2, characterized in that The method further comprises: Set feature intensity threshold; Performing feature strength weight allocation according to the target lane feature set to obtain a feature strength coefficient; The feature intensity coefficient is compared with the feature intensity threshold, and the feature intensity coefficient that does not meet the feature intensity threshold is screened out.
4. The method according to claim 1, characterized in that Performing image segmentation on the video data set according to the target lane feature set to obtain a feature image set of the target lane, the method further includes: Collecting a historical video data set, and using the historical video data set and the target lane feature set as a sample data set; Establishing a target segmentation model, inputting the historical video data set into the target segmentation model, performing image segmentation according to the target lane feature set, and obtaining the feature image set; The target segmentation model is trained and verified by using the sample data set to obtain the target segmentation model.
5. The method according to claim 4, characterized in that The target segmentation model is trained and verified by the sample data set to obtain the target segmentation model, and the method further includes: Obtaining a preset data division ratio, and dividing the sample data set into a training data set and a verification data set according to the preset data division ratio; Performing supervised training on the target segmentation model using the training data set, and when the model output result tends to converge, verifying the output result of the target segmentation model using the verification data set; A preset model verification accuracy index is obtained, and when the accuracy of the output result of the target segmentation model meets the preset model verification accuracy index, the target segmentation model is obtained.
6. The method according to claim 1, characterized in that Analyzing the feature image set to obtain an abnormal condition of the target lane, the method further includes: Collecting historical image information of the target lane, and obtaining a corresponding solution according to the historical image information; Establishing an exception handling information library according to the corresponding solution; Traverse the feature image set through the exception handling information library to determine whether the feature image set exists; If so, it is determined that the target lane is abnormal; If not present, it is transmitted to the traffic management system for manual comparison.
7. The method according to claim 6, characterized in that The method further comprises: Transmitting the feature image set to the traffic management system for manual processing to obtain a processing result; If the processing result is that the target lane has no abnormality, the feature image set has no abnormality; If the processing result is that an abnormality occurs in the target lane, the video data set corresponding to the feature image set and the processing result are added to the abnormality processing information library.
8. An object segmentation system for feature strength training, characterized in that: The system comprises: A basic information set acquisition module, the basic information set acquisition module is used to determine the target lane, connect to the traffic management system to obtain basic information of the target lane, and obtain a basic information set; A video data set acquisition module, the video data set acquisition module is used to determine the target lane target monitoring port according to the basic information, and acquire the corresponding video data set in real time; A target lane feature set acquisition module, the target lane feature set acquisition module is used to input the video data set into a feature extraction module for feature extraction, and acquire a target lane feature set; A feature image set acquisition module, the feature image set acquisition module is used to perform image segmentation on the video data set according to the target lane feature set to obtain the feature image set of the target lane; An abnormal situation acquisition module is used to analyze the feature image set to obtain the abnormal situation of the target lane.