Video recharge test method and system
By synchronous processing and multimodal feature fusion based on video frame number changes in video recharge test, the problem of high processor requirements in the prior art is solved, which reduces system costs and improves vehicle safety.
Patent Information
- Application Number
- CN202411458674.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-18
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing video refilling test method requires the physical quantity and video data of the actual vehicle data to be packaged into original messages in real time, which leads to extremely high processor requirements and increases system costs.
Test messages and videos are received through the CAN transceiver module and the HDMI to MIPI module, and synchronous processing is performed based on the changes in the video frame number, discarding the video frames corresponding to the repeated frame number, and using a convolutional neural network to extract image features, vehicle speed, yaw angle, The brake and steering wheel angles are multi-modal fusion, and the state fusion is performed through the gated circulation unit and the full connection layer to determine the fault.
Reduces processor load, avoids the demand for high-performance processors, reduces the cost of video refilling tests, and improves vehicle safety.
Smart Images

Figure CN119052391A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle function testing, and in particular to a video replay testing method and system. Background Art
[0002] With the rapid development of modern society and technology, as well as people's increasing requirements for automobile driving safety, Advanced Driver Assistance Systems (ADAS) have become standard for commercial vehicles. ADAS is a technology system installed on vehicles that aims to improve driving safety and convenience by providing automated and enhanced driver assistance functions. The system uses a variety of sensors, cameras, radars, lidars and other devices to collect information about the vehicle and its surroundings, and then analyzes this data through complex algorithms and computer processing technology to assist drivers in achieving a safer and more efficient driving experience.
[0003] In order to ensure the effectiveness of the lane departure warning function, multiple long-mileage road tests are required on vehicles equipped with the product during the research and development process. During the road tests, the product will expose several problems, which the testers will record and feedback to the developers, who will then fix the problems and upgrade the system.
[0004] After the problem is located and repaired, it is usually necessary to verify in the same scenario to confirm whether the problem has been fixed. At the same time, it is necessary to verify whether the repair of the problem will cause new problems. For this purpose, video replay is usually used to send videos in various actual scenarios to the device under test for testing.
[0005] The existing video replay test method requires that the physical quantity of the real vehicle data and the video data be packaged into original messages in real time and sent to the DUT, which will then process them. This process places extremely high demands on the processor, and the use of high-performance processors and their accessories will increase the cost of the entire system. Summary of the invention
[0006] The main purpose of the present invention is to provide a video replay test method and system to solve the problem that the prior art needs to package the physical quantity of real vehicle data and video data into original messages in real time and send them all to the device under test, which is then processed by the device under test process. This process has extremely high requirements on the processor, and the use of high-performance processors and their accessories will lead to an increase in the cost of the entire system.
[0007] In order to achieve the above object, according to one aspect of the present invention, a video replay test method is provided, comprising:
[0008] S1: Receive the test message fed back from the control end through the CAN transceiver module;
[0009] S2: Receives the test video fed back from the control end through the HDMI to MIPI module;
[0010] S3: Based on the change of the video frame number, synchronously process the received test message and the test video;
[0011] S4: Perform functional testing based on the synchronized test message and test video.
[0012] Furthermore, the test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
[0013] Furthermore, the S3 specifically includes:
[0014] S301: Obtaining a current frame number from the test message;
[0015] S302: Determine whether the current frame number is the same as the previous frame number; if so, discard the video frame corresponding to the current frame number; otherwise, proceed to the next step;
[0016] S303: Calculate the frame number difference between the current frame number and the previous frame number;
[0017] S304: Reading a corresponding number of video frames according to the frame number difference;
[0018] S305: Repeat S301 to S305 until all test messages and test videos are received.
[0019] Furthermore, the S4 specifically includes:
[0020] S401: extracting image features of each video frame in the test video through a convolutional neural network;
[0021] S402: Perform multimodal fusion on the vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video to form a multimodal feature sequence;
[0022] S403: Inputting the multimodal feature sequence into a first gated recurrent unit to obtain a first hidden state at each moment;
[0023] S404: inputting the first hidden state into a second gated recurrent unit to obtain a second hidden state at each moment;
[0024] S405: In a fully connected layer, the first hidden state and the second hidden state are merged to obtain a comprehensive hidden state;
[0025] S406: Determine whether a fault occurs and the type of fault according to the comprehensive hidden state.
[0026] Furthermore, the video replay test method further includes:
[0027] S5: Analyze the functional test results of the device under test to determine whether the functional test results between the device under test and the reference part are consistent; if so, complete the test; otherwise, perform manual re-inspection.
[0028] According to one aspect of the present invention, a video replay test system is provided, comprising:
[0029] The control end feeds back the test message to the device under test through the CAN transceiver module, and feeds back the test video to the device under test through the HDMI to MIPI module;
[0030] The device under test synchronizes the received test message and test video based on the change of the video frame number, and performs a functional test according to the synchronized test message and test video.
[0031] Furthermore, the test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
[0032] Furthermore, the device under test is specifically used for:
[0033] Obtaining a current frame number from the test message;
[0034] Determine whether the current frame number is the same as the previous frame number; if so, discard the video frame corresponding to the current frame number; otherwise, proceed to the next step;
[0035] Calculating a frame number difference between the current frame number and the previous frame number;
[0036] Reading a corresponding number of video frames according to the frame number difference;
[0037] Repeat the above steps until all test messages and test videos are received.
[0038] Furthermore, the device under test is specifically used for:
[0039] Extracting image features of each video frame in the test video through a convolutional neural network;
[0040] Perform multimodal fusion on the vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video to form a multimodal feature sequence;
[0041] Inputting the multimodal feature sequence into a first gated recurrent unit to obtain a first hidden state at each moment;
[0042] Inputting the first hidden state into a second gated recurrent unit to obtain a second hidden state at each moment;
[0043] In the fully connected layer, the first hidden state and the second hidden state are fused to obtain a comprehensive hidden state;
[0044] According to the comprehensive hidden state, it is determined whether a fault occurs and the type of fault.
[0045] Furthermore, the device under test is specifically used for:
[0046] Analyze the functional test results of the device under test to determine whether the functional test results between the device under test and the reference part are consistent; if so, complete the test; otherwise, perform manual re-inspection.
[0047] By applying the technical solution of the present invention, when the device under test receives the test message and the test video, it can synchronously process the received test message and the test video based on the change of the video frame number, and discard the repeated video frames. There is no need to receive all the physical quantities of the real vehicle data and the video data, which reduces the processor load and does not require a high-performance processor and its accessories, thereby reducing the cost of the video re-injection test, which is conducive to the application of the video re-injection test and improves vehicle safety.
[0048] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:
[0050] Figure 1 A schematic diagram of a process flow of a video replay test method provided by an embodiment of the present invention is shown;
[0051] Figure 2 A schematic structural diagram of a video replay test system provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION
[0052] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments 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.
[0053] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0054] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so as to describe the embodiments of the present invention described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0055] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0056] Reference Manual Attached Figure 1 , showing a flow chart of a video replay test method provided by an embodiment of the present invention.
[0057] Reference Manual Attached Figure 2 , showing a schematic structural diagram of a video replay test system provided by an embodiment of the present invention.
[0058] An embodiment of the present invention provides a video replay test method, which is applied to a device under test and includes:
[0059] S1: Receive the test message fed back from the control end through the CAN transceiver module.
[0060] Among them, the CAN transceiver module is a hardware component used for Controller Area Network (CAN) communication. CAN is a multi-master communication protocol widely used in automotive electronic control systems, industrial automation and other embedded systems for data transmission between different nodes (devices).
[0061] Optionally, the test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
[0062] Optionally, the control terminal may be a personal computer PC.
[0063] Furthermore, the CAN transceiver module is also responsible for receiving the result message output by the device under test and parsing it into physical quantities related to the function for further evaluation.
[0064] S2: Receives the test video fed back from the control end through the HDMI to MIPI module.
[0065] Among them, the HDMI to MIPI module is a hardware module used to convert HDMI (High Definition Multimedia Interface) signals to MIPI (Mobile Industry Processor Interface) signals. HDMI and MIPI are two different digital video signal standards.
[0066] S3: Based on the change of the video frame number, the received test message and the test video are synchronously processed.
[0067] In a possible implementation, S3 specifically includes:
[0068] S301: Obtain the current frame number from the test message.
[0069] S302: Determine whether the current frame number is the same as the previous frame number. If so, discard the video frame corresponding to the current frame number. Otherwise, proceed to the next step.
[0070] It should be noted that by discarding the video frames corresponding to the repeated frame numbers, the same visual information can be avoided from being processed, thereby reducing data redundancy. This can reduce the demand for data storage and the complexity of processing, and there is no need to receive all the physical quantities of real vehicle data and video data, reduce the processor load, and do not require high-performance processors and their accessories, reduce the cost of video re-injection testing, facilitate the application of video re-injection testing, and improve vehicle safety.
[0071] S303: Calculate the frame number difference between the current frame number and the previous frame number.
[0072] S304: Read a corresponding number of video frames according to the frame number difference.
[0073] S305: Repeat S301 to S305 until all test messages and test videos are received.
[0074] In the present invention, the test message and the test video are ensured to correspond precisely in time, so that each set of video frames and message data can be accurately matched. This is very important for functional testing and fault diagnosis, because data and video at specific time points need to be analyzed.
[0075] S4: Perform functional testing based on the synchronized test message and test video.
[0076] In a possible implementation, S4 specifically includes:
[0077] S401: Extracting image features of each video frame in the test video through a convolutional neural network.
[0078] Among them, convolutional neural network (CNN) has powerful capabilities in image feature extraction and can automatically learn and extract discriminative features from test videos.
[0079] Specifically, how to extract image features of each video frame in a test video through a convolutional neural network is already a very mature prior art. Reference may be made to relevant literature on convolutional neural networks. To avoid repetition, the present invention will not go into details.
[0080] S402: Perform multimodal fusion on the vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video to form a multimodal feature sequence:
[0081]
[0082] Among them, X represents the multimodal feature sequence, x t represents the feature vector at time t, x represents the feature vector, a1 represents the vehicle speed, β1 represents the fusion coefficient of the vehicle speed, a2 represents the yaw angle, β2 represents the fusion coefficient of the yaw angle, a3 represents braking, β3 represents the fusion coefficient of braking, a4 represents the steering wheel angle, β4 represents the fusion coefficient of the steering wheel angle, a5 represents the image features of the test video, and β5 represents the fusion coefficient of the image features of the test video.
[0083] Among them, technical personnel in this field can set the fusion coefficient β1 of vehicle speed, the fusion coefficient β2 of yaw angle, the fusion coefficient β3 of brake, the fusion coefficient β4 of steering wheel angle and the fusion coefficient β5 of image features of test video according to actual conditions, and the present invention does not limit this.
[0084] It should be noted that different data modalities can provide complementary information. For example, test videos can provide visual information about the vehicle's surroundings, while data such as speed and braking provide precise numerical information about the vehicle's dynamics. Through multimodal fusion, this information can be effectively combined to improve the overall understanding of the vehicle's status.
[0085] S403: Input the multimodal feature sequence into the first gated recurrent unit to obtain the first hidden state at each moment.
[0086] Among them, the gated recurrent unit (GRU) is good at processing time series data and can capture temporal dependencies.
[0087] Specifically, how to extract the hidden state of time series data through a gated recurrent unit is a very mature prior art, and reference may be made to the literature on the gated recurrent unit. To avoid repetition, the present invention will not go into details.
[0088] S404: Input the first hidden state into the second gated recurrent unit to obtain the second hidden state at each moment.
[0089] It should be noted that through the processing of two layers of gated recurrent units, deeper feature extraction of time series data can be performed. The first layer of hidden states may capture lower-level temporal dependencies and patterns, while the second layer of hidden states can further abstract and summarize information to form higher-level feature representations.
[0090] S405: In the fully connected layer, the first hidden state and the second hidden state are fused to obtain a comprehensive hidden state:
[0091]
[0092] Among them, H t represents the comprehensive hidden state, h 1t represents the first hidden state, h 2t represents the second hidden state, and λ represents the fusion coefficient of the first hidden state.
[0093] It should be noted that by weighted fusion of two layers of hidden states, redundant information that may be contained in a single layer of hidden states can be avoided, making the comprehensive hidden state more representative and concentrated. This fusion helps to extract more accurate features and reduce the impact of noise.
[0094] S406: Determine whether a fault occurs and the type of fault based on the comprehensive hidden state:
[0095]
[0096] Among them, P represents the probability value of each fault type, P i represents the probability value of the i-th fault type, Softmax( ) represents the Softmax activation function, W h represents the weight matrix between the fully connected layer and the prediction layer, b h represents the prediction layer bias term, i=1,2,…,n, and n represents the total number of fault types.
[0097] In the present invention, through multimodal fusion and deep learning technology, more comprehensive, accurate and efficient fault detection and classification can be provided, thereby improving the safety and reliability of the vehicle system.
[0098] In a possible implementation manner, the video replay test method further includes:
[0099] S5: Analyze the functional test results of the tested device and determine whether the functional test results of the tested device and the reference device are consistent. If yes, complete the test. Otherwise, perform manual re-inspection.
[0100] In the actual testing process, the fault diagnosis and retesting process starts with the processing of real vehicle test data. In the real vehicle test, the vehicle will be equipped with the system under test and a benchmark system. This benchmark system is usually a system recognized by the industry for excellent performance. It obtains data on the activation time of the ADAS function of the tested device and the benchmark device (including time, message, and video).
[0101] Afterwards, check the alarm status of the device under test and the reference device at around the same time to determine whether the functional test results between the device under test and the reference device are consistent.
[0102] For example, if the device under test alarms, but the reference device does not, the device under test is judged to be a suspected false alarm. If the device under test does not alarm, but the reference device alarms, the device under test is judged to be a suspected missed alarm; if the reference device and the device under test both alarm around the same time, no manual re-judgment is required.
[0103] Next, we will conduct a manual re-inspection to determine whether the problem of the tested part exists objectively based on the performance of the tested part and the benchmark part as well as the actual driving scenario. If there is a problem with the tested part, we will classify it according to the function type and the problem manifestation, including but not limited to: LDW function false alarm, missed alarm; AEB function false alarm, missed alarm; ACC speed adjustment abnormality, etc. Then we will capture the data (messages, videos) 10 seconds before and after the problem occurs.
[0104] Finally, the captured problem data is placed on the video re-injection platform to re-test the DUT. Because the re-injected message also contains the functional performance of the benchmark during the actual vehicle test, the processing method for the video re-injection test results can follow the actual vehicle test data processing method until the software problem of the DUT is solved or it is confirmed that the hardware performance boundary of the DUT has been reached.
[0105] It should be noted that through manual re-inspection, it is possible to further confirm whether the problem of the tested part objectively exists in combination with the actual driving scenario. This step of human intervention can avoid the misjudgment that may occur when relying on automated testing and improve the accuracy of the test.
[0106] From the above description, it can be seen that in the video replay test method provided by the present invention, when the device under test receives the test message and the test video, it can synchronously process the received test message and the test video based on the change of the video frame number, and discard the repeated video frames. There is no need to receive all the physical quantities of the real vehicle data and the video data, which reduces the processor load and does not require a high-performance processor and its accessories, thereby reducing the cost of the video replay test, which is conducive to the application of the video replay test and improves vehicle safety.
[0107] The embodiment of the present invention provides a video replay test system, including: a control terminal 201, a CAN transceiver module 202, an HDMI to MIPI module 203 and a device under test 204.
[0108] The control end 201 feeds back the test message to the device under test 204 through the CAN transceiver module 202 , and feeds back the test video to the device under test 204 through the HDMI to MIPI module 203 .
[0109] The device under test 204 synchronizes the received test message and test video based on the change of the video frame number, and performs a functional test according to the synchronized test message and test video.
[0110] In a possible implementation, the test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
[0111] In a possible implementation manner, the device under test 204 is specifically used for:
[0112] Get the current frame number from the test message.
[0113] Determine whether the current frame number is the same as the previous frame number. If so, discard the video frame corresponding to the current frame number. Otherwise, proceed to the next step.
[0114] Calculates the frame number difference between the current frame number and the previous frame number.
[0115] According to the frame number difference, read the corresponding number of video frames.
[0116] Repeat the above steps until all test messages and test videos are received.
[0117] In a possible implementation manner, the device under test 204 is specifically used for:
[0118] The image features of each video frame in the test video are extracted through the convolutional neural network.
[0119] The vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video are multimodally fused to form a multimodal feature sequence.
[0120] The multimodal feature sequence is input into the first gated recurrent unit to obtain the first hidden state at each moment.
[0121] The first hidden state is input into the second gated recurrent unit to obtain the second hidden state at each moment.
[0122] In the fully connected layer, the first hidden state and the second hidden state are fused to obtain a comprehensive hidden state.
[0123] According to the comprehensive hidden status, determine whether a fault occurs and the type of fault.
[0124] In a possible implementation manner, the device under test 204 is specifically used for:
[0125] Analyze the functional test results of the device under test 204 to determine whether the functional test results of the device under test 204 and the reference device are consistent. If yes, complete the test. Otherwise, perform manual re-inspection.
[0126] The video replay test system provided in the embodiment of the present invention can implement each step in the above-mentioned video replay test method and can obtain the same technical effect. To avoid repetition, the present invention will not go into details.
[0127] From the above description, it can be seen that in the video replay test system provided by the present invention, when the device under test receives the test message and the test video, it can synchronously process the received test message and the test video based on the change of the video frame number, and discard the repeated video frames. There is no need to receive all the physical quantities of the real vehicle data and the video data, which reduces the processor load and does not require a high-performance processor and its accessories, thereby reducing the cost of the video replay test, which is conducive to the application of the video replay test and improves vehicle safety.
[0128] Unless otherwise specifically stated, the relative arrangement of the parts and steps described in these embodiments, numerical expressions and numerical values do not limit the scope of the present invention. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, method and equipment known to ordinary technicians in the relevant field may not be discussed in detail, but in appropriate cases, the technology, method and equipment should be regarded as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.
[0129] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used here to describe the spatial positional relationship between a device or feature and other devices or features as shown in the figure. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figure. For example, if the device in the accompanying drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" can include both "above" and "below". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used here are interpreted accordingly.
[0130] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "lateral, vertical, perpendicular, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the devices or elements referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of the present invention. The directional words "inside and outside" refer to the inside and outside relative to the contours of each component itself.
[0131] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A video replay test method, characterized in that: Applicable to the DUT, including: S1: Receive the test message fed back from the control end through the CAN transceiver module; S2: Receives the test video fed back from the control end through the HDMI to MIPI module; S3: Based on the change of the video frame number, synchronously process the received test message and the test video; S4: Perform functional testing based on the synchronized test message and test video.
2. The video replay test method according to claim 1, characterized in that: The test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
3. The video replay test method according to claim 1, characterized in that: The S3 specifically includes: S301: Obtaining a current frame number from the test message; S302: Determine whether the current frame number is the same as the previous frame number; if so, discard the video frame corresponding to the current frame number; otherwise, proceed to the next step; S303: Calculate the frame number difference between the current frame number and the previous frame number; S304: Reading a corresponding number of video frames according to the frame number difference; S305: Repeat S301 to S305 until all test messages and test videos are received.
4. The video replay test method according to claim 1, characterized in that: The S4 specifically includes: S401: extracting image features of each video frame in the test video through a convolutional neural network; S402: Perform multimodal fusion on the vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video to form a multimodal feature sequence; S403: Inputting the multimodal feature sequence into a first gated recurrent unit to obtain a first hidden state at each moment; S404: inputting the first hidden state into a second gated recurrent unit to obtain a second hidden state at each moment; S405: In a fully connected layer, the first hidden state and the second hidden state are merged to obtain a comprehensive hidden state; S406: Determine whether a fault occurs and the type of fault according to the comprehensive hidden state.
5. The video replay test method according to claim 1, characterized in that: Also includes: S5: Analyze the functional test results of the device under test to determine whether the functional test results of the device under test and the reference device are consistent; If yes, complete the test; Otherwise, manual re-inspection is performed.
6. A video replay test system, characterized in that: include: Control terminal, CAN transceiver module, HDMI to MIPI module and DUT; The control end feeds back the test message to the device under test through the CAN transceiver module, and feeds back the test video to the device under test through the HDMI to MIPI module; The device under test synchronizes the received test message and test video based on the change of the video frame number, and performs a functional test according to the synchronized test message and test video.
7. The video replay test system according to claim 6, characterized in that: The test message includes: vehicle speed, yaw angle, brake, and steering wheel angle.
8. The video replay test system according to claim 6, characterized in that: The device under test is specifically used for: Obtaining a current frame number from the test message; Determine whether the current frame number is the same as the previous frame number; if so, discard the video frame corresponding to the current frame number; otherwise, proceed to the next step; Calculating a frame number difference between the current frame number and the previous frame number; Reading a corresponding number of video frames according to the frame number difference; Repeat the above steps until all test messages and test videos are received.
9. The video replay test system according to claim 6, characterized in that: The device under test is specifically used for: Extracting image features of each video frame in the test video through a convolutional neural network; Perform multimodal fusion on the vehicle speed, yaw angle, brake, steering wheel angle in the test message and the image features of the test video to form a multimodal feature sequence; Inputting the multimodal feature sequence into a first gated recurrent unit to obtain a first hidden state at each moment; Inputting the first hidden state into a second gated recurrent unit to obtain a second hidden state at each moment; In the fully connected layer, the first hidden state and the second hidden state are fused to obtain a comprehensive hidden state; According to the comprehensive hidden state, it is determined whether a fault occurs and the type of fault.
10. The video replay test system according to claim 6, characterized in that: The device under test is specifically used for: Analyze the functional test results of the device under test to determine whether the functional test results between the device under test and the reference part are consistent; if so, complete the test; otherwise, perform manual re-inspection.
Citation Information
Patent Citations
Unmanned driving vehicle cognitive competence testing system and method
CN103335853A
Synchronous media transmitting and playing method based on data sensation
CN107835424A
Multi-task encryption network traffic classification method based on cross-modal feature fusion
CN114358118A
Testing method and system based on AVM video recharge simulation technology
CN115914515A
Marine safety information intelligent decision support method based on multi-modal fusion network
CN118230108A
