A test method and device for a train intelligent detection system based on image recognition

Abstract

The embodiments of the present application provide a test method and device for a train intelligent detection system based on image recognition, which relate to the field of rail transit technology. The test method for a train intelligent detection system based on image recognition includes: simulating multiple obstacles and multiple operating scenarios on rail transit; based on multiple obstacles and multiple operating scenarios, performing simulation tests on a train intelligent detection system based on image recognition to obtain simulation test results; collecting the driving log of a train intelligent detection system based on image recognition during the operation of the train on an actual line, as well as the driving video corresponding to the driving log; based on the driving log and the driving video, analyzing and obtaining the on-site test results of a train intelligent detection system based on image recognition; based on the simulation test results and the on-site test results, obtaining the performance parameters of a train intelligent detection system based on image recognition. Through the above settings, the accuracy of the system performance test can be improved.

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a testing method and device for a train intelligent detection system based on image recognition. Background Art

[0002] The intelligent detection system for urban rail transit trains uses artificial intelligence and sensors for image recognition to detect information about obstacles ahead of the line, provide the driver with the corresponding obstacle type, distance, and recommended speed of the train, and give advance notice of the line ahead, thereby guiding the driver to drive and take protective measures. The inventors have found that in the prior art, the obstacles and scenes in the actual lines of urban rail transit are relatively simple, and the measured data cannot accurately judge the system performance, resulting in the problem of low accuracy of system performance testing. Summary of the invention

[0003] In view of this, the purpose of the present application is to provide a testing method and device for a train intelligent detection system based on image recognition to improve the problems existing in the prior art.

[0004] To achieve the above purpose, the embodiment of the present application adopts the following technical solution:

[0005] A testing method for a train intelligent detection system based on image recognition, comprising:

[0006] Simulating multiple obstacles and multiple operation scenarios on rail transit, wherein the multiple obstacles include multiple types of obstacles;

[0007] Based on the multiple obstacles and multiple operating scenarios, a simulation test is performed on the train intelligent detection system based on image recognition to obtain a simulation test result;

[0008] Collecting the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log;

[0009] Based on the driving log and the driving video, analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition;

[0010] Based on the simulation test results and the field test results, performance parameters of the train intelligent detection system based on image recognition are obtained.

[0011] In a preferred embodiment of the present application, the step of performing a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles and multiple operating scenarios to obtain the simulation test results includes:

[0012] Performing a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles to obtain a first test result;

[0013] Performing a simulation test on the train intelligent detection system based on image recognition based on the multiple operation scenarios to obtain a second test result;

[0014] A simulation test result is calculated based on the first test result and the second test result.

[0015] In a preferred embodiment of the present application, the step of calculating the simulation test result according to the first test result and the second test result includes:

[0016] Obtaining first weights corresponding to the plurality of obstacles and second weights corresponding to the plurality of operating scenarios;

[0017] The sum of the product of the first test result and the first weight and the product of the second test result and the second weight is taken as the simulation test result.

[0018] In a preferred embodiment of the present application, the step of analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition based on the driving log and the driving video includes:

[0019] Converting the driving log into a first information file in a preset format, and converting the driving video into a second information file in the preset format, wherein both the first information file and the second information file include an obstacle detection type and a time period in which the obstacle is detected;

[0020] The obstacle detection type and the time period in the first information file and the second information file are compared to obtain a field test result.

[0021] In a preferred embodiment of the present application, the field test result includes a missed recognition rate;

[0022] The step of comparing the obstacle detection type and the time period in the first information file and the second information file to obtain the field test result includes:

[0023] Obtaining an obstacle recognition result in a first information file corresponding to a first time period, wherein the first time period is a time period in which an obstacle is recognized in the second information file;

[0024] Determining whether the obstacle recognition result includes an obstacle;

[0025] If the obstacle recognition result does not include an obstacle, a missed recognition rate is calculated according to the number of frames in the first time period and the total number of frames of the driving video.

[0026] In a preferred embodiment of the present application, the field test result includes a false recognition rate;

[0027] The step of comparing the obstacle detection type and the time period in the first information file and the second information file to obtain the field test result includes:

[0028] Obtaining an obstacle recognition result of the first information file and a video recognition result of the second information file corresponding to the second time period, wherein the obstacle recognition result includes the first obstacle, and the video recognition result includes the second obstacle;

[0029] Determining whether the type of the first obstacle is the same as the type of the second obstacle;

[0030] If the type of the first obstacle is different from the type of the second obstacle, the misrecognition rate is calculated according to the number of frames in the second time period and the total number of frames of the driving video.

[0031] In a preferred embodiment of the present application, the testing method further includes:

[0032] Simulate various fault scenarios on rail transit, and based on the various fault scenarios, perform simulation tests on the train intelligent detection system based on image recognition to obtain fault test results;

[0033] Based on the simulation test results, the field test results and the fault test results, performance parameters of the train intelligent detection system based on image recognition are obtained.

[0034] The embodiment of the present application also provides a test device for a train intelligent detection system based on image recognition, comprising:

[0035] A simulation module, used to simulate multiple obstacles and multiple operation scenarios on rail transit, wherein the multiple obstacles include multiple types of obstacles;

[0036] A simulation test module, used to perform a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles and multiple operation scenarios, and obtain a simulation test result;

[0037] A collection module, used to collect the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log;

[0038] An analysis module, used for analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition based on the driving log and the driving video;

[0039] A parameter acquisition module is used to obtain the performance parameters of the train intelligent detection system based on image recognition based on the simulation test results and the field test results.

[0040] An embodiment of the present application further provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned test method when executing the program.

[0041] An embodiment of the present application further provides a storage medium, wherein the storage medium includes a computer program, and when the computer program is executed, the electronic device where the storage medium is located is controlled to execute the above-mentioned test method.

[0042] The embodiments of the present application provide a testing method and device for a train intelligent detection system based on image recognition. A simulation test is performed based on multiple obstacles and multiple operating scenarios to obtain simulation test results, a field test result is obtained based on driving logs and driving video analysis, and performance parameters of the system are obtained based on the simulation test results and the field test results. Thus, the final performance parameters of the system are obtained by comprehensively considering the simulation test results and the field test results of simulating multiple obstacles and multiple scenarios, thereby avoiding the problem of low accuracy of system performance testing in the prior art due to the fact that the actual line obstacles and scenarios are relatively simple and the measured data cannot accurately judge the system performance, thereby effectively improving the accuracy of the system performance testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 The structure block diagram of the train intelligent detection system based on image recognition provided by the embodiment of the present application is shown;

[0045] Figure 2 A structural block diagram of an electronic device provided in an embodiment of the present application is shown;

[0046] Figure 3 A flow chart of a test method for a train intelligent detection system based on image recognition provided in an embodiment of the present application;

[0047] Figure 4A structural block diagram of a test device for a train intelligent detection system based on image recognition provided in an embodiment of the present application.

[0048] Icons: 10-train intelligent detection system based on image recognition; 100-electronic device; 110-first memory; 120-first processor; 130-communication module; 200-train; 400-test device for train intelligent detection system based on image recognition; 410-simulation module; 420-simulation test module; 430-acquisition module; 440-analysis module; 450-parameter acquisition module. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0050] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0051] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0052] Figure 1 This is a structural block diagram of the train intelligent detection system 10 based on image recognition provided in the embodiment of the present application, which provides a possible implementation method of the train intelligent detection system, see Figure 1 The train intelligent detection system may include an electronic device 100 and a train 200 which are communicatively connected.

[0053] The specific type of the electronic device 100 is not limited and can be set according to actual application requirements. For example, in an alternative example, the electronic device 100 can be an onboard computer, which is a part of the train 200 .

[0054] For another alternative example, the electronic device 100 may be a server device, which is a different device from the train 200 .

[0055] As for train 200, it should be noted that train 200 can be an urban rail transit train. Urban rail transit trains can include long-focus and short-focus cameras, millimeter-wave radars and laser radars. Obstacles can be identified using image data collected by the long-focus and short-focus cameras. Because the distance between the two cameras is certain, based on the binocular distance-fixed principle, with the assistance of millimeter-wave radars and laser radars, combined with algorithms, it can be determined whether the object ahead is an obstacle and the exact distance of the obstacle ahead can be identified.

[0056] Regarding the electronic device 100 , it should be noted that the electronic device 100 is installed with simulation software or can be communicatively connected with the simulation software, so that a simulation test can be performed on the operation of the train 200 .

[0057] Please refer to Figure 2 , is a block diagram of an electronic device 100 provided in an embodiment of the present application. The electronic device 100 in this embodiment may be a server, a processing device, a processing platform, etc. capable of data interaction and processing. The electronic device 100 includes a first memory 110, a first processor 120, and a communication module 130. The first memory 110, the first processor 120, and the communication module 130 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0058] The first memory 110 is used to store programs or data. The first memory 110 can be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc.

[0059] The first processor 120 is used to read / write data or programs stored in the first memory 110 and execute corresponding functions. The communication module 130 is used to establish a communication connection between the electronic device 100 and other communication terminals through a network and to send and receive data through the network.

[0060] It should be understood that Figure 2 The structure shown is only a schematic diagram of the structure of the electronic device 100. The electronic device 100 may also include Figure 2More or fewer components as shown, or with Figure 2 Different configurations are shown. Figure 2 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0061] Figure 3 A flowchart of a test method for a train intelligent detection system based on image recognition provided by an embodiment of the present application is shown. The method can be applied to Figure 2 The electronic device 100 shown comprises Figure 2 It should be understood that in other embodiments, the order of some steps in the vehicle coupling control method of this embodiment can be interchanged according to actual needs, or some steps can be omitted or deleted. Figure 3 The process of the vehicle coupling control method shown is described in detail.

[0062] Step S310, simulating multiple obstacles and multiple operation scenarios on rail transit.

[0063] The multiple obstacles include obstacles of multiple types.

[0064] Step S320, based on multiple obstacles and multiple operating scenarios, a simulation test is performed on the train intelligent detection system based on image recognition to obtain a simulation test result.

[0065] Step S330, collecting the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log.

[0066] Step S340: Based on the driving log and driving video, analyze and obtain the field test results of the train intelligent detection system based on image recognition.

[0067] Step S350, based on the simulation test results and the field test results, the performance parameters of the train intelligent detection system based on image recognition are obtained.

[0068] The above method performs simulation tests based on multiple obstacles and multiple operating scenarios to obtain simulation test results, obtains field test results based on driving logs and driving video analysis, and obtains system performance parameters based on the simulation test results and the field test results, so as to obtain the final performance parameters of the system by comprehensively considering the simulation test results and the field test results of simulating multiple obstacles and multiple scenarios. On the one hand, simulation tests are performed on multiple obstacles and multiple operating scenarios, which increases the scope of obstacle and scene simulation tests and improves the accuracy of obstacle tests. On the other hand, performance parameters are obtained by comprehensively considering the simulation test results and the field test results, which improves the accuracy of performance parameter calculation, avoids the problem of low accuracy of system performance testing caused by the fact that the actual line obstacles and scenarios in the prior art are relatively simple and the measured data cannot accurately judge the system performance, and effectively improves the accuracy of system performance testing.

[0069] Before step S310, it should be noted that the testing method of the train intelligent detection system based on image recognition provided in the embodiment of the present application may also include the step of integrated testing.

[0070] In detail, the electronic device 100 can verify the correctness of the functional implementation of the train intelligent detection system 10 by using simulation software to simulate different types of obstacles for laboratory testing. This method mainly includes two stages of testing: integration testing and confirmation testing. In principle, the integration test must be completed before the confirmation test can be carried out. The integration test stage mainly tests whether the interface protocols between the various subsystems within the train intelligent detection system 10 and with the external signal system are correct. The confirmation test stage mainly uses black box testing, that is, verifying the overall system function, without considering the internal structure and characteristics of the system, and only checking whether the program function is used normally according to the requirements of the specification, and whether the program can properly receive input data and generate correct output information.

[0071] The specific content of the confirmation test is not limited and can be set according to actual application requirements. For example, in an alternative example, the confirmation test can include an obstacle test and a scenario test, that is, the simulation test based on multiple obstacles and multiple operating scenarios mentioned in step S320.

[0072] For step S310, it should be noted that the type of obstacle is not limited and can be set according to actual application requirements. For example, in an alternative example, the obstacle may include, but is not limited to, a train, a car stop, a pedestrian, a signal, a switch, a small animal, a small obstacle, etc.

[0073] Furthermore, the types of operation scenarios are not limited and can be set according to actual application requirements. For example, in an alternative example, the operation scenarios may include, but are not limited to, scenarios with vehicles at different positions of turnouts, scenarios of chasing vehicles on ramps, scenarios of chasing vehicles on curves, scenarios with vehicles in bad weather (e.g., rainy days, snowy days, foggy days, etc.), scenarios with multiple vehicles at close intervals, etc.

[0074] For step S320, the specific method of performing the simulation test is not limited and can be set according to actual application requirements. For example, in an alternative example, step S320 may include the following sub-steps:

[0075] A simulation test is performed on a train intelligent detection system based on image recognition based on multiple obstacles to obtain a first test result; a simulation test is performed on a train intelligent detection system based on image recognition based on multiple operating scenarios to obtain a second test result; and a simulation test result is calculated based on the first test result and the second test result.

[0076] The above method calculates the final simulation test result based on the first test result obtained by testing based on multiple obstacles and the second test result obtained by testing based on multiple operating scenarios, thereby comprehensively considering the simulation results obtained by simulating based on obstacle data and operating scenarios respectively, avoiding the problem of low accuracy of simulation test results caused by simulating based on obstacle data alone or simulating based on operating scenarios alone, thereby improving the accuracy of simulation testing.

[0077] It should be noted that the obstacle test mainly simulates the obstacles that may appear on the rail transit line through simulation software, and then checks whether the train intelligent detection system 10 can detect the obstacle within the detection range required by the system and issue an early warning. This type of test is used to ensure the correctness of the obstacle detection function of the train intelligent detection system 10. Among them, the simulation software mainly includes functions such as vision simulation, sensor simulation, train speed simulation, and simulated train.

[0078] The scenario test mainly simulates the detection function of the train intelligent detection system 10 for train obstacles in all scenarios that may occur during rail transit operation. By running the train intelligent detection system 10 in different scenarios, it is verified whether the system's detection function for the train 200 in different scenarios is correct, and whether there will be missed and false alarms.

[0079] The specific method of calculating the simulation test result according to the first test result and the second test result is not limited and can be set according to actual application requirements. For example, in an alternative example, the above steps may include the following sub-steps:

[0080] Obtain first weights corresponding to the multiple obstacles and second weights corresponding to the multiple operating scenarios; and take the sum of the product of the first test result and the first weight and the product of the second test result and the second weight as the simulation test result.

[0081] The above method performs weighted summation on the simulation test results based on obstacles and the simulation test results based on the operation scenarios, thereby calculating the simulation test results according to the importance of the obstacles and the operation scenarios. Obstacles or operation scenarios with high importance can account for a high proportion in the simulation test results, thereby avoiding the problem of low accuracy of the simulation test results caused by directly adding the first test result corresponding to the obstacle and the second test result corresponding to the operation scenario to calculate the simulation test results, thereby further improving the accuracy of the simulation test.

[0082] It should be noted that the first test result, the second test result and the simulation test result can be expressed in numbers. The simulation test result can characterize the functional completeness of the train intelligent detection system 10. When the simulation test result is 100, it means that all functions of the train intelligent detection system 10 are intact. When the simulation test result is 80-100, it means that most functions of the train intelligent detection system 10 are intact. When the simulation test result is 60-80, it means that the basic functions of the train intelligent detection system 10 are intact. When the simulation test result is 0-60, it means that only a few functions of the train intelligent detection system 10 are intact.

[0083] For example, in an alternative example, the first test result may be 70 points, the first weight may be 0.5, the second test result may be 80 points, the second weight may be 0.5, and the calculated simulation test result may be 75 points, indicating that the basic functions of the train intelligent detection system 10 are intact.

[0084] Furthermore, in order to reduce the amount of data processing while ensuring the basic functions of the system are intact, after calculating the first test result and the second test result, it can be determined whether the first test result or the second test result is greater than a preset score (for example, the preset score can be 60 points). If the first test result is not greater than the preset score, the second test result is used as a simulation test result. If the second test result is not greater than the preset score, the first test result is used as a simulation test result.

[0085] For example, in an alternative example, the first test result may be 50 points, and the second test result may be 80 points. The first test result is no greater than 60 points, so the second test result is directly used as the simulation test result, and the test results no greater than 60 points are eliminated to ensure that the basic functions of the system are intact. There is no need to perform weighted sum calculations on the simulation test results, thereby reducing the amount of data processing.

[0086] Regarding step S330, it should be noted that the electronic device 100 can verify the performance of the intelligent detection system by collecting actual line operation conditions on site for data analysis, and can verify the performance of the system by running the system for a long time on site and performing a large amount of data analysis.

[0087] In detail, after the train 200 runs on site, there will be a corresponding driving log. The log and the long- and short-focus videos of the driving can be uploaded and saved to a designated storage location. The electronic device 100 can download the driving log and driving video generated by the train intelligent detection system 10 during the train operation from the designated storage location.

[0088] Regarding step S340, it should be noted that the specific steps of analyzing and obtaining the field test results are not limited and can be set according to actual application requirements. For example, in an alternative example, step S340 may include the following sub-steps:

[0089] The driving log is converted into a first information file in a preset format, and the driving video is converted into a second information file in a preset format, wherein both the first information file and the second information file include an obstacle detection type and a time period when the obstacle is detected; the obstacle detection type and time period in the first information file and the second information file are compared to obtain a field test result.

[0090] The above method converts the driving log and the driving video into a first information file and a second information file in a preset format, and compares the first information file and the second information file to obtain the field test result, thereby converting the driving log and the driving video into information files in the same preset format for comparison, avoiding the problem of low efficiency of field testing caused by the inability to compare and analyze the driving log and the driving video due to different formats, and requiring staff to perform comparative analysis, thereby improving the efficiency of field testing.

[0091] In detail, the preset format can be a table, and the title of the table can include, but is not limited to, the operation date (the date of the analyzed data), the start time of the main line (the time when the train 200 turns on the lights when going up / the time when the train 200 stops at the TSW turnaround when going down), the end time of the main line (the time when the train 200 reaches the TSW turnaround when going up / the time when the train 200 turns off the lights when going down), the operation mode (there are three operation modes: LE, LN, and LH, and the LN mode is generally written), the number of correctly identified targets (the summary of the number of correctly identified train 200 frames), the number of incorrectly identified targets (the summary of the number of incorrectly identified frames), and the number of missed identified targets (the number of missed identified train 200 frames). Summary), correct identification time period (number of frames for correctly identifying train 200), correct identification quantity (number of frames for correctly identifying train 200), correct identification distance (number of frames for correctly identifying the distance between the rear vehicle and the front vehicle), incorrect identification time period (number of frames for incorrectly identifying obstacles), incorrect identification quantity (number of frames for incorrectly identifying obstacles), incorrect identification distance (distance for incorrectly identifying obstacles), missed identification time period (time period for missing identification of train 200), missed identification quantity (number of frames for missing identification of train 200), distance jump (distance continuously jumps during correct identification), mainline frame number (number of frames used for train 200 to run one lap, that is, the number of frames within the start time and end time range), etc.

[0092] The driving log can be stored in a folder named tids_log, and the driving video can be stored in the logo_video folder. The Myeclipse analysis tool can be used to convert the tids_log file and the logo_video file into two tables. The config configuration file in the Myeclipse tool can be changed to the corresponding train number, running time and storage path, and the tids_log file and the logo_video file can be converted into a table containing information such as the running date, the main line start time, the main line end time, and the running mode.

[0093] In detail, the tids_log file can be converted into a table containing obstacle detection type and distance information. The table can contain three columns: time period, type and distance. Time refers to the time period of the current analysis, type refers to the obstacle type (0 means there is no train in the driving direction or it is beyond the recognition range; 1 means a train is recognized; 2 means a small obstacle is recognized; 4 means a car stop is recognized; 5 means a pedestrian is recognized), and distance refers to the distance of the recognized obstacle. Each row in the table represents a frame, and one second includes 4-5 frames.

[0094] It should be noted that, depending on the specific type of the field test result, the specific steps of comparing the obstacle detection type and time period in the first information file and the second information file to obtain the field test result are also different. For example, in an alternative example, the field test result may include a missed recognition rate, and the step of comparing the first information file and the second information file may include the following sub-steps:

[0095] Obtain an obstacle recognition result in a first information file corresponding to a first time period, wherein the first time period is a time period in which an obstacle is recognized in a second information file; determine whether the obstacle recognition result includes an obstacle; if the obstacle recognition result does not include an obstacle, calculate a missed recognition rate based on the number of frames in the first time period and the total number of frames in the driving video.

[0096] On the one hand, the above method determines whether there is a missed recognition by searching for the obstacle in the first information file corresponding to the time period in which the obstacle is identified in the second information file. This method can quickly determine the missed recognition and avoids the problem of low efficiency of on-site testing caused by comparing all time periods to determine whether there is a missed recognition. On the other hand, the missed recognition rate is calculated by the proportion of the number of frames in the missed recognition time period, and each frame image with a missed recognition can be included in the calculation range, avoiding the problem of low accuracy in calculating the missed recognition rate by calculating the missed recognition rate by time period and calculating the frame images without missed recognition.

[0097] In detail, during the time period when an obstacle is identified in the second information file (corresponding to the driving video), if the obstacle is not identified in the first information file (corresponding to the driving log), it means that a missed identification has occurred, and the missed identification rate can be calculated based on the number of frames in the missed identification time period.

[0098] For example, in an alternative example, the time periods during which obstacles are identified in the second information file include 1:00-2:00, 10:00-10:10, and 13:00-13:05, the first information file identifies obstacles at 10:00-10:10 and 13:00-13:05, and no obstacles are identified at 1:00-2:00, the time period of the driving video is 0:00-24:00, the number of frames per second in the driving video is the same, and the missed recognition rate can be calculated to be 4.17%.

[0099] For example, in an alternative example, the field test result may include a false recognition rate, and the step of comparing the first information file with the second information file may include the following sub-steps:

[0100] Obtain an obstacle recognition result of a first information file and a video recognition result of a second information file corresponding to a second time period, wherein the obstacle recognition result includes the first obstacle and the video recognition result includes the second obstacle; determine whether the type of the first obstacle is the same as the type of the second obstacle; if the type of the first obstacle is different from the type of the second obstacle, calculate a false recognition rate according to the number of frames in the second time period and the total number of frames of the driving video.

[0101] On the one hand, the above method determines whether there is a misidentification by comparing the type of obstacle identified in the first information file with the type of obstacle identified in the second information file. It can quickly determine the misidentification and avoid the problem of low efficiency of on-site testing caused by comparing all time periods to determine whether there is a misidentification. On the other hand, the misidentification rate is calculated by the proportion of the number of frames in the misidentification time period, and each frame image with a misidentification situation can be included in the calculation range, avoiding the misidentification rate calculated by time period, and calculating the frame images without misidentification, resulting in low accuracy of the misidentification rate calculation.

[0102] In detail, in the time period when obstacles are identified in both the first information file (corresponding to the driving log) and the second information file (corresponding to the driving video), if the first obstacle contained in the first information file is of a different type from the second obstacle contained in the second information file, it means that misidentification has occurred, and the misidentification rate can be calculated based on the number of frames in the misidentification time period.

[0103] For example, in an alternative example, in the time period 10:00-10:10, the type of the first obstacle contained in the first information file is a car stop, and the type of the second obstacle contained in the second information file is a pedestrian. The types of obstacles are different, the time period of the driving video is 10:00-11:00, and the number of frames per second in the driving video is the same. The misrecognition rate can be calculated to be 16.67%.

[0104] It should be noted that, in the embodiment of the present application, the missed recognition rate and the false recognition rate can be calculated by the following formula: missed recognition rate = number of frames of missed reporting duration / number of frames of actual car chasing duration * 100%, false recognition rate = number of frames of false reporting duration / number of frames of no-car running duration * 100%, and generally the false recognition rate and the missed recognition rate retain 4 significant digits.

[0105] Regarding step S350 , it should be noted that after obtaining the simulation test results and the field test results, the simulation test results and the field test results can be compiled into a data analysis report, and the data analysis report includes the performance parameters of the train intelligent detection system 10 .

[0106] In detail, after obtaining the on-site test results, the misidentification screenshots, time periods, corresponding logo_video names and misidentification distances that occur can be recorded together, and the missed identification screenshots, time periods, corresponding logo_video names that occur can be recorded together. Finally, all the misidentification results, missed identification results and simulation test results of this train 200 are summarized, and the train number, oTIDS software version number, operating time period, total number of operating laps, single lap kilometres and total kilometres of the train 200 for this data analysis are indicated to obtain a data analysis report.

[0107] Among them, the performance parameters of the system may include the final false recognition rate and the final missed recognition rate, the simulation test results may include the simulated false recognition rate and the simulated missed recognition rate, the field test results may include the field false recognition rate and the field missed recognition rate, the final false recognition rate can be calculated based on the simulated false recognition rate and the field false recognition rate, and the final missed recognition rate can be calculated based on the simulated missed recognition rate and the field missed recognition rate.

[0108] For example, in an alternative example, the simulation misrecognition rate and the on-site misrecognition rate may be directly multiplied to obtain the final misrecognition rate, and the simulation missed recognition rate and the on-site missed recognition rate may be directly multiplied to obtain the final missed recognition rate.

[0109] For another example, in another alternative example, the first simulation weight corresponding to the simulation false recognition rate, the first on-site weight corresponding to the on-site false recognition rate, the second simulation weight corresponding to the simulation missed recognition rate, and the second on-site weight corresponding to the on-site missed recognition rate can be obtained, and the product of the simulation false recognition rate and the first simulation weight and the product of the on-site false recognition rate and the first on-site weight are added together as the final false recognition rate, and the product of the simulation missed recognition rate and the second simulation weight and the product of the on-site missed recognition rate and the second on-site weight are added together as the final missed recognition rate. It should be noted that in an alternative example, after obtaining the simulation test results and the on-site test results through the above steps S310-S340, the test method of the train intelligent detection system 10 based on image recognition provided in the embodiment of the present application can also include a fault test step, which can specifically include the following steps:

[0110] Simulate various fault scenarios on rail transit, and based on various fault scenarios, conduct simulation tests on the train intelligent detection system based on image recognition to obtain fault test results; based on the simulation test results, field test results and fault test results, obtain the performance parameters of the train intelligent detection system based on image recognition.

[0111] The above method obtains performance parameters based on simulation test results, field test results and fault test results, incorporates fault test results into the data source of performance parameter analysis, expands the scope of performance parameter analysis, and thus obtains more comprehensive and accurate performance parameters, avoiding the problem of too few data sources resulting in low accuracy of performance parameters.

[0112] In detail, the fault test is mainly to test the reaction of the system when a certain sensor fails, verify the impact of a certain part of the system failure on the detection result and the corresponding alarm signal generated to ensure the integrity of the system. The steps of the fault test may include: first, start the train intelligent detection system 10, all system components are connected intact, and the train intelligent detection system 10 can operate normally, then create a certain part of the fault, such as disconnecting the camera from the host of the train intelligent detection system 10, and check whether the display interface of the train intelligent detection system 10 displays a prompt of camera failure.

[0113] After obtaining the fault test results, the simulation test results, field test results and fault test results can be organized into a data analysis report with reference to the description of the above step S350. The data analysis report includes the performance parameters of the train intelligent detection system 10.

[0114] Combination Figure 4 , the embodiment of the present application also provides a test device 400 for a train intelligent detection system 10 based on image recognition, and the functions implemented by the test device 400 correspond to the steps performed by the above method. The test device 400 can be understood as a processor of the above electronic device 100, or as a component that is independent of the above electronic device 100 or the processor and implements the functions of the present application under the control of the electronic device 100. Among them, the test device 400 may include a simulation module 410, a simulation test module 420, a collection module 430, an analysis module 440 and a parameter acquisition module 450.

[0115] The simulation module 410 is used to simulate multiple obstacles and multiple operation scenarios on rail transit, wherein the multiple obstacles include multiple types of obstacles. In the embodiment of the present application, the simulation module 410 can be used to execute Figure 3 As shown in step S310, the relevant contents of the simulation module 410 can refer to the above description of step S310.

[0116] The simulation test module 420 is used to perform a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles and multiple operation scenarios to obtain a simulation test result. Figure 3As shown in step S320, the relevant contents about the simulation test module 420 can refer to the above description of step S320.

[0117] The acquisition module 430 is used to collect the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log. Figure 3 As shown in step S330, the relevant contents of the acquisition module 430 can refer to the above description of step S330.

[0118] The analysis module 440 is used to analyze and obtain the field test results of the train intelligent detection system based on image recognition based on the driving log and the driving video. In the embodiment of the present application, the analysis module 440 can be used to perform Figure 3 As shown in step S340, the relevant contents of the analysis module 440 can refer to the above description of step S340.

[0119] The parameter acquisition module 450 is used to obtain the performance parameters of the train intelligent detection system based on image recognition based on the simulation test results and the field test results. Figure 3 As shown in step S350, the relevant contents of the parameter acquisition module 450 can refer to the above description of step S350.

[0120] In addition, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned testing method are executed.

[0121] The computer program product of the data acquisition method provided in the embodiment of the present application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the steps of the above-mentioned test method. For details, please refer to the above-mentioned method embodiment, which will not be repeated here.

[0122] In summary, a testing method and device for a train intelligent detection system based on image recognition provided in an embodiment of the present application are adopted, simulation tests are performed based on multiple obstacles and multiple operating scenarios to obtain simulation test results, field test results are obtained based on driving logs and driving video analysis, and system performance parameters are obtained based on the simulation test results and the field test results. Thus, by comprehensively considering the simulation test results and the field test results of simulating multiple obstacles and multiple scenarios, the final performance parameters of the system are obtained, thereby avoiding the problem of low accuracy of system performance testing in the prior art due to the fact that the actual line obstacles and scenarios are relatively simple and the measured data cannot accurately judge the system performance, thereby effectively improving the accuracy of the system performance test.

[0123] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0124] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0125] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0126] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1A function specified in one or more boxes.

[0127] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0128] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0129] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A test method for a train intelligent detection system based on image recognition, characterized in that: include: Simulating multiple obstacles and multiple operation scenarios on rail transit, wherein the multiple obstacles include multiple types of obstacles; Performing a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles to obtain a first test result, performing a simulation test on the train intelligent detection system based on image recognition based on the multiple operation scenarios to obtain a second test result, obtaining first weights corresponding to the multiple obstacles and second weights corresponding to the multiple operation scenarios, and taking the product of the first test result and the first weight and the product of the second test result and the second weight as the simulation test result; Collecting the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log; Based on the driving log and the driving video, analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition; Based on the simulation test results and the field test results, performance parameters of the train intelligent detection system based on image recognition are obtained.

2. The testing method according to claim 1, characterized in that: The step of analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition based on the driving log and the driving video includes: Converting the driving log into a first information file in a preset format, and converting the driving video into a second information file in the preset format, wherein both the first information file and the second information file include an obstacle detection type and a time period in which the obstacle is detected; The obstacle detection type and the time period in the first information file and the second information file are compared to obtain a field test result.

3. The testing method according to claim 2, characterized in that: The field test results include missed recognition rate; The step of comparing the obstacle detection type and the time period in the first information file and the second information file to obtain the field test result includes: Obtaining an obstacle recognition result in a first information file corresponding to a first time period, wherein the first time period is a time period in which an obstacle is recognized in the second information file; Determining whether the obstacle recognition result includes an obstacle; If the obstacle recognition result does not include an obstacle, a missed recognition rate is calculated according to the number of frames in the first time period and the total number of frames of the driving video.

4. The testing method according to claim 2, characterized in that: The field test results include false recognition rate; The step of comparing the obstacle detection type and the time period in the first information file and the second information file to obtain the field test result includes: Obtaining an obstacle recognition result of the first information file and a video recognition result of the second information file corresponding to the second time period, wherein the obstacle recognition result includes the first obstacle, and the video recognition result includes the second obstacle; Determining whether the type of the first obstacle is the same as the type of the second obstacle; If the type of the first obstacle is different from the type of the second obstacle, the misrecognition rate is calculated according to the number of frames in the second time period and the total number of frames of the driving video.

5. The testing method according to claim 1, characterized in that: The test method also includes: Simulate various fault scenarios on rail transit, and based on the various fault scenarios, perform simulation tests on the train intelligent detection system based on image recognition to obtain fault test results; Based on the simulation test results, the field test results and the fault test results, performance parameters of the train intelligent detection system based on image recognition are obtained.

6. A test device for a train intelligent detection system based on image recognition, characterized in that: include: A simulation module, used to simulate multiple obstacles and multiple operation scenarios on rail transit, wherein the multiple obstacles include multiple types of obstacles; a simulation test module, configured to perform a simulation test on the train intelligent detection system based on image recognition based on the multiple obstacles to obtain a first test result, perform a simulation test on the train intelligent detection system based on image recognition based on the multiple operation scenarios to obtain a second test result, obtain first weights corresponding to the multiple obstacles and second weights corresponding to the multiple operation scenarios, and take a sum of a product of the first test result and the first weight and a sum of a product of the second test result and the second weight as a simulation test result; A collection module, used to collect the driving log of the train intelligent detection system based on image recognition during the operation of the train on the actual line, and the driving video corresponding to the driving log; An analysis module, used for analyzing and obtaining the on-site test results of the train intelligent detection system based on image recognition based on the driving log and the driving video; A parameter acquisition module is used to obtain the performance parameters of the train intelligent detection system based on image recognition based on the simulation test results and the field test results.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the test method according to any one of claims 1 to 5 when executing the program.

8. A storage medium, characterized in that: The storage medium includes a computer program, and when the computer program is executed, the electronic device where the storage medium is located is controlled to execute the test method according to any one of claims 1 to 5.

Images