A DMS real scene automatic detection system and method based on local area network

Abstract

The present invention provides a local area network-based DMS real-scene automatic detection system and method. The system includes: a PC host computer for acquiring and outputting first video data; a DMS test device for acquiring the first video data; the DMS test device is configured with a DMS detection algorithm, which runs the DMS detection algorithm to detect the acquired first video data, marks the detection results, generates second video data, and transmits it back to the PC host computer; after receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs the detection results. The present invention enables the new version of the DMS system to detect different drivers in multiple scenarios, automatically completes the re-injection of a large number of test scenarios, can quickly locate algorithm problems, and automatically compiles test results, making it convenient, fast, and intuitive.

Description

Technical Field

[0001] The present invention relates to the field of vehicle-mounted vision technology, and in particular to a DMS real scene automatic detection system and method based on a local area network. Background Art

[0002] A DMS (Driver Monitoring System) is a driver fatigue warning system. Its initial application was to monitor the driver's driving behavior throughout the entire driving process to prevent driver fatigue, distraction, and other behaviors that endanger the driving process. With the advancement of computer vision technology, DMS systems now cover almost all in-vehicle vision solutions. Existing methods for evaluating DMS system data reliability mainly include real-vehicle testing and indoor simulated speed testing.

[0003] Actual vehicle testing: The device is installed on a real vehicle at the actual installation angle. A person sitting in the driver's seat performs corresponding actions, triggering the corresponding alarm and visually recording the detection effect of the DMS. Relying solely on this method consumes a lot of manpower and material resources, and the test verification cycle is long. Sometimes the number of testers is limited, and it is impossible to ensure that each version has multiple people testing in multiple scenarios. It is also difficult to reproduce some trigger scenarios to ensure consistency. This method can intuitively see the performance of the DMS, but it is not feasible to rely on this method to improve the DMS detection effect in a short period of time.

[0004] Indoor opening simulation speed test: Open the simulated speed indoors, perform corresponding actions to trigger the alarm, and count the DMS algorithm detection results. This method can maintain the same environment, but it deviates from the actual situation in the car and the test scenario is relatively simple. Summary of the Invention

[0005] In view of this, an embodiment of the present application provides a method for automatic detection based on DMS real scenarios. This method can realize the detection of different drivers in multiple scenarios by the new version of the DMS system, and can automatically complete the re-injection of a large number of test scenarios. The test environment of each version can be guaranteed to be completely consistent, and algorithm problems can be quickly located. The test results can also be automatically counted, which is convenient, fast and intuitive, so as to achieve the purpose of improving the detection performance of the DMS system.

[0006] The present application provides the following technical solution: a DMS real scene automatic detection system based on a local area network, comprising:

[0007] A PC host computer, the PC host computer is used to obtain the first video data and output it;

[0008] a DMS testing device, the DMS testing device being communicatively connected to the PC host computer and configured to obtain the first video data; the DMS testing device being configured with a DMS detection algorithm, detecting the obtained first video data by running the DMS detection algorithm, marking the detection results, generating second video data, and transmitting the generated second video data back to the PC host computer;

[0009] After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result.

[0010] According to an implementation manner of the embodiment of the present application, the first video data includes camera raw video data and DMS parameters.

[0011] According to an implementation manner of the embodiment of the present application, the second video data includes closed-eye video data, yawning video data, distracted video data, head-down video data, phone call video data, smoking video data, and occlusion video data.

[0012] According to an implementation manner of the embodiment of the present application, marking the detection result specifically includes: marking the second video data as a sample and marking the alarm point true value.

[0013] According to one implementation of the embodiment of the present application, the PC host computer includes a connection module, a setting module, a statistics module and a version module;

[0014] The connection module is used to trigger the connection button between the DMS test equipment and the PC host computer;

[0015] The setting module includes the IP address of the DMS test device, the save path after the file is re-injected, the video playback frame rate, the video rotation angle, the file save format, and the DMS parameter filling;

[0016] The statistics module is used for time selection, refilling file lists, and deleting / querying / exporting files;

[0017] The version module is used to visually display the current DMS version.

[0018] According to an implementation of the embodiment of the present application, after the PC host computer receives the first video data, it can cut and copy the first video data.

[0019] According to an implementation of an embodiment of the present application, the PC host computer compares the marked points of the second video data with the true value of the first video data, outputs the detection result, and classifies the re-injection result data at the same time, automatically generates a folder for the re-injection result data, and automatically classifies all re-injection videos according to the detection result.

[0020] According to one implementation of an embodiment of the present application, the recharge result data is classified into positive alarms and false alarms, and the false alarms include false alarms of closing eyes, false alarms of yawning, false alarms of distraction, false alarms of lowering the head, false alarms of making phone calls, false alarms of smoking, and false alarms of occlusion.

[0021] According to an implementation manner of the embodiment of the present application, the DMS test device is communicatively connected to the PC host computer via an IPC to network port cable.

[0022] The present application also provides a method for automatically detecting a DMS real scene based on a local area network, using the above-mentioned automatic detection system for DMS real scene based on a local area network. The method includes:

[0023] Obtaining the first video data through the PC host computer and outputting it;

[0024] The DMS test device receives the first video data, runs a DMS detection algorithm built into the DMS test device, detects the acquired first video data, marks the detection result, generates second video data, and transmits it back to the PC host computer;

[0025] After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result.

[0026] Compared with the prior art, the at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects:

[0027] (1) The embodiment of the present invention verifies the effect of the new algorithm modification based on the data uploaded by the existing real alarm and quickly outputs the algorithm version test results.

[0028] (2) The environment is simple to set up and supports the re-injection of DMS video data collected indoors and on real vehicles. It is highly operational and automatically verifies the re-injection algorithm for all videos in a folder. The re-injection results are automatically classified, and false alarms are also automatically classified, saving manpower and material resources. The test verification cycle is short, and the scene coverage is complete and consistent with the actual situation.

[0029] (3) The embodiment of the present invention can control a single variable and count the accuracy of each alarm of multiple DMS algorithm versions. The test environment of each version can be guaranteed to be completely consistent, and algorithm problems can be quickly located. The test results can be automatically counted, which is convenient, fast and intuitive. In addition, the video of true alarm and false alarm can be automatically screened out from multiple videos, effectively improving the detection accuracy.

[0030] (4) The embodiment of the present invention can cut or copy the replayed video according to user needs, thereby meeting the processing requirements of different users for the original video. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0032] Figure 1 It is a DMS real scene automatic detection system based on a local area network according to an embodiment of the present invention;

[0033] Figure 2 The present invention relates to a method for automatically detecting a DMS real scene based on a local area network. DETAILED DESCRIPTION

[0034] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0035] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of this application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments, and the technical solutions of the present invention will be clearly and completely described. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0036] like Figure 1 As shown, an embodiment of the present invention provides a DMS real scene automatic detection system based on a local area network, including:

[0037] A PC host computer, the PC host computer is used to obtain the first video data and output it;

[0038] a DMS testing device, the DMS testing device being communicatively connected to the PC host computer and configured to obtain the first video data; the DMS testing device being configured with a DMS detection algorithm, detecting the obtained first video data by running the DMS detection algorithm, marking the detection results, generating second video data, and transmitting the generated second video data back to the PC host computer;

[0039] The first video data includes the camera's raw video data and DMS parameters (eye_detect, eye_dist, curve_down, and down_detect). After the DMS test device acquires the first video data, it runs the DMS detection algorithm and detects each frame in the video data. It detects all detection types implemented by the algorithm, including closed eyes, yawning, distracted, head-down, phone calls, smoking, and occlusion. The detection results are then labeled with samples and alarm point true values, and transmitted to the PC host computer as the second video data. In this embodiment of the present application, the first video data only needs to be in MP4 format.

[0040] After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result.

[0041] The DMS test device is the carrier of the re-injection program and the new version of the DMS detection algorithm. After the PC host computer sets the path of the re-injection file, it is transmitted to the DMS test device. The DMS test device parses the video data through the path of the re-injection file and uses the new DMS detection algorithm version on the DMS test device to detect the video detection points of the re-injection file. The results of the latest algorithm detection are transmitted to the PC host computer. The PC host computer compares the results of the latest algorithm detection with the original algorithm detection results, that is, compares the marked second video data with the first video data. The PC host computer generates a corresponding alarm folder to facilitate final statistics.

[0042] In this embodiment of the present invention, the host computer runs the host software and opens the video directory to be replayed, such as C: / Closed Eyes. This loads a list of all videos in the current directory. Click the 'Start' button to begin replaying. Before replaying each new video, the host computer resets the DMS detection algorithm parameters eye_detect, eye_dist, curve_down, and down_detect in the DMS test device to zero, then begins sending replayed video frames.

[0043] When the DMS detection algorithm detects an alarm, it transmits the alarm information back to the PC host computer, which receives and displays it, and automatically archives the video file.

[0044] In the specific implementation of the embodiment of the present invention, the PC host computer mainly includes a connection module, a setting module, a statistics module, and a version module;

[0045] The connection module: a button for triggering the connection between the DMS test equipment and the PC host computer;

[0046] The setting module includes the IP address of the DMS test device, the save path after the file is re-injected, the video playback frame rate, the video rotation angle, the file saving method (copy, paste), and the DMS parameter filling;

[0047] The statistics module is used for time selection, file list refilling, file deletion / query / export, and the number of files displayed per page;

[0048] The version module is used to visually display the currently injected DMS version.

[0049] The DMS testing device in this embodiment primarily serves as a platform for the new DMS detection algorithm. Each frame of video data transmitted from the PC is fed back to the new DMS detection algorithm, which then performs a test on each frame. If an alarm is detected within the frame, the alarm result is transmitted back to the PC.

[0050] The DMS test equipment in the embodiment of the present invention is implemented as a device with IPC departmental standards.

[0051] In a preferred embodiment of the present invention, after the PC host computer sets the IP address of the connection port in the connection configuration, it is connected to the DMS test equipment via Ethernet, and the DMS video path is fed back through the host computer interface.

[0052] Specifically, when the PC host computer sets the second video data file path that can be re-injected to the DMS test device, the first video data directory to be re-injected is written into the mount path parameter of the DMS test device. When the PC host computer views the re-injection result data and status of the DMS test device in real time, after setting the IP address of the DMS test device in the connection configuration, it can connect to the designated re-injection device via Ethernet, and then set the file path for the re-injected DMS test device through the PC host computer interface to view the re-injection result and status of the DMS test device in real time.

[0053] In a specific implementation, marking the detection result specifically includes: the PC host computer marking the second video data as a sample and an alarm point true value. The first video data is a test case file, which is used as a reference when verifying the video replay performance of the DMS test equipment.

[0054] According to a preferred implementation of the embodiment of the present application, after the PC host computer receives the first video data, it can cut and copy the first video data to meet the processing requirements of different users for the original video data.

[0055] Furthermore, a specific implementation of the present invention also includes a PC that outputs the test results for each video in real time and automatically generates a folder for the video replay results, automatically categorizing all replay videos. After selecting a folder, the PC automatically replays the data files, eliminating the need for human oversight; users only need to view the results after completion.

[0056] Specifically, the PC compares the DMS test device's recharge result data (i.e., the second video data) with the first video data, outputs a detection result, and simultaneously categorizes the recharge result data, automatically generating a folder for the recharge result data. All recharge videos are automatically categorized based on the detection result. The recharge result data is categorized into positive and false positives, with false positives including those for eyes closed, yawning, distraction, looking down, making a phone call, smoking, and occlusion, all detection types implemented by the DMS algorithm.

[0057] Preferably, the DMS test device in the embodiment of the present application is communicatively connected to the PC host computer via an IPC to network port cable.

[0058] Next, the DMS real scene automatic detection system of the embodiment of the present application is further explained by taking the replay of the closed-eye video as an example.

[0059] The PC host computer and the DMS test equipment are connected via Ethernet. When the closed-eye video starts to be replayed, the PC host computer resets the detection algorithm in the DMS test equipment while the PC host computer is connected to the DMS test equipment, setting the algorithm parameters to 0. The PC host computer reads each frame of the MP4 video and transmits the data of each frame to the DMS test equipment end. The algorithm on the DMS test equipment end recognizes the first detection result in the video and transmits this result to the PC host computer via Ethernet. The PC host computer compares the video result with the original video annotation points, and then the PC host computer automatically performs classification.

[0060] The embodiment of the present application adopts the DMS client re-injection method to re-inject video data through the client according to the real scene, so as to realize the detection of different drivers in multiple scenarios in the new version. It can automatically complete the re-injection of thousands of test scenes. The test environment of each version can be guaranteed to be completely consistent, and algorithm problems can be quickly located. The test results of the algorithm version can be quickly output, and the test results can be automatically counted, which is convenient, fast and intuitive.

[0061] like Figure 2 As shown, the embodiment of the present application further provides a method for automatically detecting a DMS real scene based on a local area network, using the above-mentioned DMS real scene automatic detection system based on a local area network, the method comprising:

[0062] S101, obtaining first video data through a PC host computer and outputting it;

[0063] S102: The DMS test device receives the first video data, runs a DMS detection algorithm built into the DMS test device, detects the acquired first video data, marks the detection result, generates second video data, and transmits it back to the PC host computer;

[0064] S103: After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result.

[0065] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A DMS real scene automatic detection system based on local area network, characterized in that: include: A PC host computer, the PC host computer is used to obtain the first video data and output it; The PC host computer includes a connection module, a setting module, a statistics module and a version module; The connection module is used to trigger the connection button between the DMS test equipment and the PC host computer; The setting module includes the IP address of the DMS test device, the save path after the file is re-injected, the video playback frame rate, the video rotation angle, the file save format, and the DMS parameter filling; The statistics module is used for time selection, refilling file lists, and deleting / querying / exporting files; The version module is used to visually display the current DMS version; A DMS testing device, wherein the DMS testing device is communicatively connected to the PC host computer via an IPC to network port cable, and is used to obtain the first video data. The DMS testing device is configured with a DMS detection algorithm, and detects the obtained first video data by running the DMS detection algorithm, marks the detection results, generates second video data, and transmits the generated second video data back to the PC host computer; After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result; Among them, the first video data includes camera original video data and DMS parameters; the second video data includes closed-eye video data, yawning video data, distracted video data, head-down video data, phone call video data, smoking video data, and occlusion video data.

2. The DMS real scene automatic detection system based on local area network according to claim 1 is characterized in that: Marking the detection result specifically includes: marking the second video data with samples and alarm point true values.

3. The DMS real scene automatic detection system based on local area network according to claim 1 is characterized in that: After receiving the first video data, the PC host computer can cut and copy the first video data.

4. The DMS real scene automatic detection system based on local area network according to claim 1 is characterized in that: It also includes that the PC host computer compares the marked points of the second video data with the true value of the first video data, outputs the detection result, and classifies the re-injection result data at the same time, automatically generates a folder for the re-injection result data, and automatically classifies all re-injection videos according to the detection result.

5. The LAN-based DMS real scene automatic detection system according to claim 4 is characterized in that: The recharge result data is classified into positive alarms and false alarms, wherein the false alarms include false alarms of closing eyes, false alarms of yawning, false alarms of distraction, false alarms of lowering the head, false alarms of making phone calls, false alarms of smoking, and false alarms of occlusion.

6. A DMS real scene automatic detection method based on local area network, characterized in that: Using the local area network-based DMS real scene automatic detection system according to any one of claims 1 to 5, the method includes: Obtaining the first video data through the PC host computer and outputting it; The DMS test device receives the first video data, runs a DMS detection algorithm built into the DMS test device, detects the acquired first video data, marks the detection result, generates second video data, and transmits it back to the PC host computer; After receiving the second video data, the PC host computer compares the marked points of the second video data with the true values ​​of the first video data and outputs a detection result.

Images