Automated Testing Method and Device for Automotive Lamp Control System

Abstract

The present application provides an automated testing method and device for an automotive lamp control system. The method includes: inputting a preset atmosphere simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs various atmosphere lights; receiving the atmosphere lights through an event camera at a preset position to generate an atmosphere event stream; inputting the atmosphere simulation instruction and the atmosphere event stream into a trained LSTM model to obtain test abnormal items; determining the abnormal automotive lamp control system according to the test abnormal items, and outputting a test abnormal report.

Description

Technical Field

[0001] The present application relates to the technical field of lamp testing, and particularly to an automated testing method and device for an automotive lamp control system. Background Art

[0002] Currently, as an important part of enhancing user experience and brand recognition, the functions and performance of automotive ambient lights are increasingly valued by automotive manufacturers. There are many technical problems in traditional testing methods for automotive lamp control systems. Existing testing methods usually rely on the actual vehicle development stage to conduct tests, resulting in a significant delay in the testing cycle. If problems are found, rework and modification are required, seriously affecting the development progress and cost control, and making it difficult to detect and correct early design defects in a timely manner. At the same time, traditional testing mostly relies on human eye observation to judge the display effect of ambient lights. This subjective evaluation method not only makes it difficult to form a unified standard due to individual differences in the human eye's perception of light, and cannot perform quantitative analysis, but also easily causes visual fatigue when observing the light source for a long time, affecting the judgment accuracy, and facing problems of low efficiency and high labor costs. Summary of the Invention

[0003] The present application provides an automated testing method and device for an automotive lamp control system, which is used to conduct ambient light testing on the automotive lamp control system in the early stage of development, improve testing efficiency and reduce testing costs.

[0004] In a first aspect, an embodiment of the present application provides an automated testing method for an automotive lamp control system, the method comprising:

[0005] Inputting a preset ambient simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs a variety of ambient lights;

[0006] Receiving the ambient light through an event camera at a preset position to generate an ambient event stream;

[0007] Inputting the ambient simulation instruction and the ambient event stream into a trained LSTM model to obtain test exception items;

[0008] Determining the abnormal automotive lamp control system according to the test exception items, and outputting a test exception report.

[0009] In a second aspect, an embodiment of the present application provides an automated testing device for an automotive lamp control system, the device comprising:

[0010] A test control module, configured to input a preset ambient simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs a variety of ambient lights;

[0011] An event generation module, configured to receive the ambient light through an event camera at a preset position and generate an ambient event stream;

[0012] An anomaly recognition module, configured to input the ambient simulation instruction and the ambient event stream into a trained LSTM model to obtain test anomaly items;

[0013] A result output module, configured to determine an abnormal automotive lamp control system according to the test anomaly items and output a test anomaly report.

[0014] In a third aspect, an embodiment of the present application provides a test device, which includes a memory and a processor;

[0015] The memory is used to store a computer program;

[0016] The processor is configured to execute the computer program and, when executing the computer program, implement the automated test method for an automotive lamp control system as described in any one of the embodiments of the present application.

[0017] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the automated test method for an automotive lamp control system as described in any one of the embodiments of the present application.

[0018] An embodiment of the present application provides an automated test method for an automotive lamp control system. The method includes: inputting a preset ambient simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs various ambient lights; receiving the ambient light through an event camera at a preset position to generate an ambient event stream; inputting the ambient simulation instruction and the ambient event stream into a trained LSTM model to obtain test anomaly items; determining an abnormal automotive lamp control system according to the test anomaly items and outputting a test anomaly report. Through the above method, an event camera is used to accurately capture the change of ambient light, an objective and unified test standard is established, visual fatigue is avoided, quantitative analysis of light characteristics is realized, anomaly recognition is combined with the LSTM model, full automation and standardization of testing are realized, consistency of test results is ensured, problems are accurately located by automatically generating an anomaly report, the test cost is reduced, the manpower input is reduced, and at the same time the test accuracy is improved, achieving the test goals of high efficiency, low cost and high quality. Description of the Drawings

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for description in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 Schematic flowchart of an automated test method for an automotive lamp control system provided by an embodiment of the present application;

[0021] Figure 2 Schematic block diagram of an automated test device for an automotive lamp control system provided by an embodiment of the present application. Detailed implementation manners

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] The flowchart shown in the accompanying drawings is only an example illustration, and does not necessarily include all the contents and operations / steps, nor does it necessarily need to be executed in the described order. For example, some operations / steps can also be decomposed, combined, or partially merged, so the actual execution order may change according to the actual situation.

[0024] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application and the appended claims, unless otherwise clearly specified in the context, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0025] It should be further understood that the term " / and / " used in the specification of the present application and the appended claims refers to any combination and all possible combinations of one or more of the related listed items, and includes these combinations.

[0026] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of an automated test method for an automotive lamp control system provided by an embodiment of the present application. As Figure 1 shown, the specific steps of the automated test method for the automotive lamp control system include: S101 - S104.

[0027] S101. Input a preset atmosphere simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs multiple atmosphere lights.

[0028] Exemplarily, preset atmosphere simulation instructions are sent to the automotive lighting control system to be tested. These instructions are carefully designed to trigger the output of various types of atmosphere light. To obtain better test results, the switching between multiple modes may also be added to the instruction set to test the mode conversion efficiency of the automotive lighting control system. The preset atmosphere simulation instructions usually contain multiple parameter combinations, such as color values (RGB or HSV color spaces), brightness levels, fade rates, color temperature ranges, lighting modes (such as pulsating, fading, breathing, etc.), and special scene simulations (such as sunset, aurora, ocean, etc.). These instructions can be sent to the lighting control unit in the form of digital signals through the CAN bus, LIN bus, or a dedicated test interface. In actual operation, the test system will establish an instruction library covering all possible atmosphere light scenarios that the vehicle may use. For example, when the user selects the "romantic mode", the automotive lighting control system will send the instruction "MODE: ROMANTIC; COLOR: #FF69B4; BRIGHTNESS: 70; STYLE: BREATHING; CYCLE: 3s" to the lighting controller, which will trigger the lighting to produce a slow breathing effect in pink. During the test process, the instructions will be sent sequentially according to a predetermined sequence to ensure that every possible atmosphere light scenario is tested. By this method, the response ability of the automotive lighting control system in different working modes and the accuracy of the output effect can be comprehensively verified, and at the same time, the stability of the system when receiving complex instructions or rapid instruction switching can also be detected.

[0029] S102. Receive the atmosphere light through the event camera at the preset position to generate an atmosphere event stream.

[0030] Exemplarily, special event camera technology is utilized to capture the ambient light changes emitted by automotive lamps and convert these changes into a sequential event data stream. Different from traditional cameras, an event camera does not acquire complete images at a fixed frame rate. Instead, it only records the pixel positions where the brightness in the scene changes and the direction of the change (increase or decrease), featuring a microsecond-level time resolution and a high dynamic range. In the test setup, the event camera is fixed at pre-planned camera positions, which are carefully designed to comprehensively cover the illumination areas of all lamps inside the vehicle. For example, when testing the center console ambient light, the event camera may be located between the driver's seat and the passenger seat; when testing the door ambient light, the camera is placed directly opposite the door panel. The parameters of the event camera, such as the contrast threshold (usually set to ±15% brightness change) and the time resolution (up to 1 μs), are precisely configured to ensure that all subtle light changes can be captured. When the lamp executes the "fade from blue to red" instruction, the event camera generates a series of event data, recording the brightness changes of each pixel point during the fade from blue to red. These raw event data are then converted into a standardized event stream, retaining the spatio-temporal information of the light changes. This event-based acquisition method has significant advantages compared to traditional video acquisition. It can not only accurately capture rapidly changing lighting effects but also greatly reduce the data volume and improve the analysis efficiency.

[0031] S103. Input the atmosphere simulation instruction and the atmosphere event stream into the trained LSTM model to obtain test anomaly items.

[0032] Exemplarily, the atmosphere simulation instructions and the atmosphere event stream obtained in the first two steps are simultaneously input into a pre-trained deep learning model for anomaly detection. The LSTM (Long Short-Term Memory) model used is specifically designed to process sequential data and can effectively capture the features and patterns of the atmosphere light changing over time. During the training phase, the model uses a large amount of normal operation data of standard lighting systems and manually labeled anomaly data for learning, establishing a mapping relationship between the instructions and the expected light output. In actual testing, the model receives two inputs simultaneously: one event camera is used to receive the atmosphere simulation instructions, and the other is the atmosphere event stream captured by the event camera. The LSTM model extracts the instruction features and understands their semantic meanings (for example, the instruction "MODE: PARTY; COLOR: RAINBOW; SPEED: FAST" is parsed as a fast-switching colorful lighting effect); analyzes the event stream features to identify the actual light change patterns; and finally compares the correlation between the two to detect potential anomalies. For example, when the instruction requires "blue gradient", but the event stream shows that the light is "red flashing", the model will mark this inconsistency as an anomaly item. The anomaly items output by the model include not only the anomaly types (such as color deviation, brightness anomaly, response delay, etc.), but also the quantitative scores of the anomaly degree, providing detailed basis for subsequent analysis. This machine learning method greatly improves the automation and accuracy of testing and can detect tiny anomalies that are difficult to detect manually.

[0033] S104. Determine the abnormal automotive lamp control system according to the test anomaly items and output a test anomaly report.

[0034] Exemplarily, the module responsible for report generation classifies and organizes the abnormal items output by the LSTM model, and summarizes them into different types of abnormalities, such as the color not matching the instruction (for example, the instruction requires blue but actually shows green), brightness abnormality (such as the instruction requires 50% brightness but actually only has 30%), response time delay (such as the response is more than 200 ms after the instruction is issued), etc. According to the preset abnormal judgment criteria, a severity level is assigned to each type of abnormality, which is usually divided into three levels: minor (does not affect the function but has aesthetic problems), moderate (the function is affected but still usable), and severe (the function fails). The number and severity of abnormalities occurring in each tested vehicle lamp control system are counted. When the cumulative abnormalities of a certain lamp system exceed the preset threshold (such as severe abnormality ≥ 1 time, or moderate abnormality ≥ 3 times), the system is marked as an "abnormal system". For example, when the ambient light control system of a certain vehicle model has the problem of "color deviation" 5 times in 10 tests, and the deviation value exceeds 30% of the color gamut space, it will be determined that there is a severe abnormality. After performing test anomaly analysis through the above process, a test anomaly report is automatically generated according to the analysis results. The report includes a detailed list of abnormal lamps, a description of the abnormal type of each lamp, the specific scenario where the abnormality occurs, the quantitative indicators of the abnormality, and the accompanying event camera data archive. This report not only intuitively displays the test results but also provides the abnormal reproduction conditions, facilitating engineers to perform subsequent fault diagnosis and system optimization.

[0035] An embodiment of the present application provides an automated test method for an automotive lamp control system. The method includes: inputting a preset ambient simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs various ambient lights; receiving the ambient lights through an event camera at a preset position to generate an ambient event stream; inputting the ambient simulation instruction and the ambient event stream into a trained LSTM model to obtain test abnormal items; determining the abnormal automotive lamp control system according to the test abnormal items, and outputting a test abnormal report. Through the above method, the event camera is used to accurately capture the changes in ambient light, establish an objective and unified test standard, avoid visual fatigue, realize the quantitative analysis of light characteristics, combine with the LSTM model for anomaly recognition, realize the full automation and standardization of the test, ensure the consistency of test results, accurately locate problems by automatically generating an abnormal report, reduce the test cost, reduce the manpower input, and at the same time improve the test accuracy, achieving the test goals of high efficiency, low cost, and high quality.

[0036] To more clearly introduce the technical solution of the present application, the technical solution of the present application will also be introduced through specific embodiments below. It should be noted that the specific embodiment is used to expand the description of the technical solution of the present application, rather than limiting the present application.

[0037] In some embodiments, an event camera at a preset position receives ambient light and generates an ambient event stream, including: S1021 - S1028.

[0038] S1021. Configure the parameters of the event camera at the preset position, and set the contrast threshold and time interval parameter as the event trigger conditions.

[0039] Exemplarily, by setting the contrast threshold (usually ±15% brightness change), the camera is triggered to record only when the pixel brightness change exceeds this threshold. At the same time, set the time interval parameter (such as 1000 microseconds) to define the minimum time granularity of event sampling. In the ambient light test of luxury cars, the contrast threshold can be set to a lower value (±10%) to capture subtle gradient effects; while in the dynamic light test of sports cars, the threshold can be appropriately increased (±20%) to focus on significant changes. This refined parameter configuration ensures that the event camera can optimize the acquisition for different types of ambient lights.

[0040] S1022. Based on the event camera after parameter configuration, obtain the pixel point log intensity change value of the ambient light, and compare it with the contrast threshold to obtain the original event data representing the brightness change direction.

[0041] Exemplarily, after the configuration is completed, the event camera starts to work, continuously monitors the light intensity received by each pixel point, and calculates the logarithmic difference between the current pixel brightness value and the previous value in real time to form the log intensity change value. Taking the door ambient light as an example, when the light changes from dark blue (RGB value: 0, 0, 128) to bright blue (RGB value: 0, 0, 255), the pixel log intensity may change by 0.3 units. The data system connected to the event camera compares this change value with the contrast threshold (such as ±0.2). If it exceeds the threshold, an event tuple containing the pixel coordinates (x, y), timestamp t, and polarity p (+1 or -1) is generated, indicating an increase or decrease in brightness at that position at a specific time point. This method accurately records the spatial position, time, and direction information of the light change.

[0042] S1023. Temporally partition the original event data according to the timestamp, and aggregate the events captured between two adjacent frames into multiple uniform time bins to obtain a time-discretized event set.

[0043] Exemplarily, the original event data is asynchronous and continuous, and needs to be divided by time sequence for subsequent processing. The data system divides the event stream according to time windows (such as 10 milliseconds) to create multiple evenly distributed time bins. For example, when testing the "rhythm mode" ambient light, the data of one second may be divided into 100 time bins. All events that occur within each time bin are aggregated, and the time order of the events is retained. If the center console ambient light performs a breathing effect, the data system will record a large number of positive events generated during the brightness increase stage, and then a set of negative events in the brightness decrease stage will appear in the time bin. This method of dividing by uniform time intervals converts the continuous event stream into a discrete time series, facilitating subsequent pattern recognition and feature extraction.

[0044] S1024. Obtain the polarity information of the latest timestamp event at each pixel position within each time bin, and perform spatial aggregation on the time-discretized event set according to the polarity information to obtain multiple event slices represented in two-dimensional space.

[0045] Exemplarily, traverse each time bin, and for each pixel position (x, y), only retain the polarity value (+1 or -1) of the latest-occurring event. For example, when testing the "rainbow mode" of the ambient light, multiple events may be recorded successively at position (320, 240) within one time bin, and the data system only retains the polarity value of the last event. Perform this operation on the entire image space to form a two-dimensional matrix with the same resolution as the camera, which is called an event slice, where the matrix element value is +1 (brightness increase), -1 (brightness decrease), or 0 (no change). Perform the same operation on each time bin to generate a series of event slices, and each slice reflects the spatial distribution of light changes within a specific time window. This spatial aggregation method effectively compresses the data volume while retaining key information, converts the time-sequential events into a spatial distribution representation, and reveals the spatial pattern of light changes.

[0046] S1025. Perform normalization processing on multiple event slices represented in two-dimensional space to obtain a standardized sequence of event frames.

[0047] Exemplarily, the event slices need to be normalized to eliminate device differences and environmental impacts. The data system uses a statistical normalization method to adjust the numerical distribution of each two-dimensional event slice to the standard range [-1, 1] to ensure data consistency and comparability. For example, when capturing the "starry sky ceiling" ambient light effect in the car, the event density in the top area may be much higher than that in the side area, and this difference can be balanced through normalization processing. The normalization process also includes noise suppression and outlier handling, such as smoothing sudden brightness changes (which may be environmental light interference). For the multi-area linked ambient light effect in the car, normalization can ensure that the light changes in different areas are equivalently represented regardless of their original brightness levels.

[0048] S1026. Input the standardized event frame sequence into a preset global spatial dependency extractor. Through the self-attention mechanism of the global spatial dependency extractor, calculate the correlation matrix of each pixel position with all other pixel positions to obtain the global spatial feature map.

[0049] Exemplarily, for each pixel position, calculate its degree of association with all other pixel positions in the image. For example, when capturing the "wrap-around" ambient light effect, even if these areas are spatially far apart, the self-attention mechanism can identify the cooperative pattern of the front-row and rear-row light changes, generate the Q (query), K (key), and V (value) matrices, obtain the attention weights through matrix multiplication and softmax operation, and form a feature map representing the global spatial dependency relationship. This method is particularly suitable for analyzing complex ambient light effects. For example, in the "rhythm tracking" mode, the synchronous changes of lights in different areas can be accurately captured. Global spatial dependency analysis enables the system to understand the mutual influence between distant pixels and reveals the overall pattern of light changes.

[0050] S1027. Input the standardized event frame sequence into a preset GM-LSTM network for sorting to generate the hidden state at each time step.

[0051] Exemplarily, the GM-LSTM (gated memory long short-term memory) network is designed specifically for processing time series data and can effectively capture the temporal features in the event frame sequence. The standardized event frame sequence is input into the network in chronological order. For example, when testing the "music interaction mode" ambient light, 50 consecutive standardized event frames are fed into the network in sequence. GM-LSTM selectively retains or forgets specific information through its special gating structure (including the input gate, forget gate, and output gate). For example, it can identify the pattern that the brightness of the ambient light increases at low syllable beats. The network generates a corresponding hidden state vector for each input frame. These vectors contain the cumulative temporal information up to the current time point and encode the short-term and long-term dependency relationships. These hidden state vectors form the temporal dependency feature representation.

[0052] S1028. Through differential operation and spatial attention mechanism, according to the global spatial feature map, extract the discriminative clues of the hidden states at adjacent time steps to obtain the ambient event stream.

[0053] Exemplarily, the difference between the hidden states of adjacent time steps is calculated to highlight the state changes. For example, when the vehicle's "welcome mode" is activated, the state differences between consecutive time steps reveal the pattern of the lights gradually turning on from the outside to the inside. Combining with the global spatial feature map, the difference features are weighted through the spatial attention mechanism to emphasize the changes in important regions. For example, when the ambient light switches from the "blue calm mode" to the "red passion mode", the attention mechanism highlights the regions with significant color changes. The processed features form the ambient event stream, which integrates information from both spatial and temporal dimensions and accurately represents the complete features of the light changes.

[0054] In some embodiments, the standardized event frame sequence is input into a preset global spatial dependency extractor. Through the self-attention mechanism of the global spatial dependency extractor, the association degree matrix between each pixel position and all other pixel positions is calculated to obtain the global spatial feature map, including: S261 - S267.

[0055] S261 is constructed as a two-dimensional feature matrix with the shape of H×W×C based on the standardized event frame sequence to obtain the spatial linear feature representation.

[0056] Exemplarily, the standardized event frame sequence is reorganized into a three-dimensional feature matrix, where H and W respectively represent the height and width of the image (such as 1920×1080), and C represents the number of feature channels. When processing the "starry sky mode" of the in-vehicle ambient light, each pixel position contains multiple feature channels, such as brightness change values, time information, etc. Through this structured representation, the system can integrate multi-dimensional feature data while maintaining the spatial position information. For example, the flashing pattern of the ceiling ambient light is represented as a sequence of feature vectors at specific spatial positions in this matrix.

[0057] S262 performs a three-way feature projection on the spatial linear feature representation to obtain the query matrix, the key matrix, and the value matrix.

[0058] Exemplarily, a three-way feature transformation is performed on the spatial linear feature representation to generate the query matrix Q, the key matrix K, and the value matrix V. When analyzing the "gradient mode" of the door ambient light, the query matrix represents the currently concerned features, the key matrix represents the possible matching features, and the value matrix contains the actual feature information. Each matrix is generated through different linear projection weights to ensure the expression of features in different spaces.

[0059] S263 performs a transposed matrix multiplication operation on the query matrix and the key matrix, and divides the product result by a scaling factor for amplitude adjustment to obtain the original attention map.

[0060] Exemplarily, the product of the query matrix Q and the transposed matrix of the key matrix K is calculated to obtain a score matrix representing the similarity between pixels. When analyzing the "rhythm mode" of the center console, a high similarity score indicates that the light change patterns at two positions are similar. Adjustment is made through a scaling factor (usually the square root of the feature dimension) to avoid the problem of vanishing gradients caused by overly large values.

[0061] S264. Apply the softmax normalization function to the original attention map to convert the similarity values into a probability distribution, obtaining a normalized spatial attention weight matrix.

[0062] Exemplarily, apply the softmax function to normalize the original attention map, converting the similarity scores into a probability distribution. When processing the "ambient atmosphere mode", softmax ensures that the sum of the attention weights at each position is 1, and a high probability value indicates a strong correlation between that position and the currently focused position. This probabilistic representation enables highlighting important spatial associations.

[0063] S265. Perform a matrix multiplication operation on the normalized spatial attention weight matrix and the value matrix to obtain a spatial dependence feature map.

[0064] Exemplarily, multiply the normalized attention weight matrix by the value matrix V to obtain a feature representation that integrates global information. When analyzing the "music interaction mode", the output feature at each position contains weighted information from all relevant positions, reflecting the spatial dependence relationship of the light changes.

[0065] S266. Perform a temporal transfer process on the spatial dependence feature map, connecting the feature state at the current time step with the hidden state at the previous time step to obtain an enhanced feature representation.

[0066] Exemplarily, perform a temporal correlation process on the feature map, combining the feature at the current time step with the state information at the previous time step. When analyzing the "light chasing mode", this connection enables the system to understand the temporal pattern of the light changes, such as the propagation direction and speed of the light effects. The integration of temporal information enhances the expressive ability of the features.

[0067] S267. Restore the enhanced feature representation to a three-dimensional tensor H×W×C to obtain a global spatial feature map.

[0068] Exemplarily, reshape the enhanced feature into the original three-dimensional tensor format (H×W×C) to restore the spatial structure. When processing the "full vehicle atmosphere mode", the reconstructed feature map retains the complete spatial layout information, while containing global dependence relationships and temporal features, facilitating subsequent anomaly detection analysis.

[0069] In some embodiments, through differential operations and spatial attention mechanisms, based on the global spatial feature map, discriminative cues of the hidden states at adjacent time steps are extracted to obtain an ambient event stream, including: S281 - S285.

[0070] S281. Perform a subtraction operation on the output state of the previous time step and the input state of the current time step of the hidden state, and obtain discriminative features by calculating the state difference tensor.

[0071] Exemplarily, perform an exact differential calculation on the hidden states at adjacent time steps. During the calculation process, perform an element-wise subtraction operation on the output state vector (with a dimension of 256×256×64) of the previous time step t - 1 and the input state vector of the current time step t. When processing the "rhythm mode" of the in-vehicle ambient light, the state difference tensor can capture a drastic change in the light intensity from 80% to 20%, or a subtle color difference from blue (RGB: 0, 0, 255) to purple (RGB: 128, 0, 255). The difference calculation uses 32-bit floating-point precision to ensure that no information is lost due to numerical truncation. The generated discriminative features contain key information about temporal changes and provide a basis for subsequent feature enhancement.

[0072] S282. Perform a non-linear transformation on the discriminative features through a convolutional layer, batch normalization, and ReLU activation function to obtain an enhanced representation of the discriminative features.

[0073] Exemplarily, apply a 3×3 convolutional kernel to the discriminative features for spatial feature extraction, set the convolutional stride to 1, and use the SAME mode for padding to keep the feature map size unchanged. The momentum parameter of the batch normalization layer is set to 0.99, and epsilon is set to 1e - 5 to standardize the feature distribution. The ReLU activation function truncates negative values to 0 and retains positive features to enhance the non-linear expression ability of the feature representation. When processing the "gradient tracking" mode, this step can effectively extract the spatial continuity features of the light change, such as the movement trajectory and diffusion pattern of the light beam. The enhanced feature map more prominently expresses the dynamic characteristics of the light change, and the dimension remains 256×256×64.

[0074] S283. Perform a channel mean operation and a Sigmoid activation function conversion on the enhanced representation of the discriminative features to obtain a spatio-temporal awareness mask.

[0075] Exemplarily, calculate the mean of the 64 channels of the enhanced features respectively to obtain a channel attention vector. Map the attention values to the [0, 1] interval through the Sigmoid function to generate a 256×256×1-dimensional spatio-temporal awareness mask. The high-value regions (such as above 0.8) in the mask indicate that there are significant temporal changes at that position, and the low-value regions (such as below 0.2) indicate that the changes are not obvious.

[0076] S284. Weight the output state of the previous time step and the input state of the current time step respectively according to the spatiotemporal perception mask to obtain a motion change correction vector.

[0077] For example, the spatiotemporal perception mask is expanded to the same dimension as the state tensor (256×256×64) through a broadcast mechanism, and element-wise multiplication is performed with the output state of the previous time step and the input state of the current time step, respectively. The multiplication operation enhances the features of the areas with significant changes and suppresses the features of the areas with insignificant changes. The weighted state tensor more accurately reflects the motion change characteristics of the ambient light.

[0078] S285. Perform channel-level connection operation on the motion change correction vector, and perform feature fusion through the convolution layer to obtain an atmosphere event stream.

[0079] Exemplarily, the motion change correction vectors are concatenated in the channel dimension to form a 256×256×128 feature tensor. The concatenated features are reduced in dimension through a 1×1 convolution layer, with 64 convolution kernels, a step size of 1, and an output dimension of 256×256×64. The convolution operation combines the feature information of the previous and next time steps, and the generated atmosphere event stream contains complete temporal change features, which can be directly used for subsequent anomaly detection analysis.

[0080] In some embodiments, the atmosphere simulation instructions and atmosphere event streams are input into a trained LSTM model to obtain test anomalies, including: S1031-S1035.

[0081] S1031. Input the atmosphere simulation instructions and atmosphere event streams into the trained LSTM model, wherein the trained LSTM model includes multiple anomaly detection networks and decision networks, each anomaly detection network includes a first long short-term memory network, a second long short-term memory network and a threshold recurrent network, and the decision network includes a fully connected layer and a Sigmoid function.

[0082] Exemplarily, the trained LSTM model adopts a multi-level anomaly detection architecture, and the model as a whole consists of N parallel anomaly detection networks and 1 decision network. Each anomaly detection network contains two independent long short-term memory networks and a threshold recurrent network, where the first long short-term memory network is configured with 128 hidden units to extract the temporal characteristics of atmosphere simulation instructions; the second long short-term memory network is configured with 256 hidden units to specifically process dynamic change information in the atmosphere event stream. The threshold recurrent network adopts a GRU structure with a hidden layer dimension of 512, which is used to fuse the feature representation of instructions and events. The decision network consists of 3 fully connected layers with 1024, 512 and 256 neurons respectively, and the end is connected to the Sigmoid activation function for binary classification prediction.

[0083] S1032. Extract the hidden state features of the atmosphere simulation instruction through the first long short - term memory network in each anomaly detection network to obtain the instruction feature representation vector.

[0084] Exemplarily, the first long short - term memory network processes the input atmosphere simulation instruction sequence. The network adopts a bidirectional LSTM structure, with 64 memory units in both the forward and reverse directions. The forgetting gate controls the degree of forgetting of historical information, the input gate determines the update ratio of new information, and the output gate adjusts the output amount of information. The network updates the cell state and hidden state at each time step. The cell state maintains long - term memory, and the hidden state captures short - term features. After multiple layers of non - linear transformation and state transfer, an instruction feature representation vector with a dimension of 128 is generated, which encodes the key temporal patterns of the atmosphere simulation instruction.

[0085] S1033. Extract the hidden state features of the atmosphere event stream through the second long short - term memory network in each anomaly detection network to obtain the event feature pyramid.

[0086] Exemplarily, the second long short - term memory network adopts a hierarchical feature extraction strategy to construct multi - scale feature representations for the atmosphere event stream data. The network contains 3 LSTM layers, and the number of hidden units in each layer is 256, 128, and 64 in sequence. The bottom - layer LSTM focuses on local detailed features, the middle - layer LSTM integrates medium - scale temporal dependencies, and the top - layer LSTM captures global semantic information. Through the residual connection mechanism, it ensures the effective transmission of feature information at different levels. The output event feature pyramid contains feature maps of multiple scales, forming a multi - granularity representation of the atmosphere event stream.

[0087] S1034. Input the instruction feature representation vector and the event feature pyramid into the gated recurrent network for feature integration to generate the correlation vector matrix of instructions and events.

[0088] Exemplarily, the gated recurrent network receives the instruction feature representation vector and the event feature pyramid as inputs, and dynamically adjusts the feature fusion process through the update gate and the reset gate. The update gate controls the influence degree of the current input information on the hidden state, and the reset gate determines the retention ratio of historical information. The network sets a 512 - dimensional hidden state. Through multiple rounds of iterative operations, it gradually establishes the spatio - temporal correlation relationship between the instruction features and the event features, and generates a correlation vector matrix with a dimension of 1024.

[0089] S1035. Input the correlation vector matrix into the decision network, and perform anomaly prediction on each instruction and the corresponding event through the fully - connected layer and the Sigmoid function in the decision network to obtain the test anomaly items.

[0090] Exemplarily, the decision-making network adopts a multi-layer perceptron structure, and reduces the dimension and extracts features from the associated vector matrix through three fully-connected layers. The output of the fully-connected layer is processed by batch normalization and the ReLU activation function, effectively alleviating the problem of gradient disappearance. The Sigmoid function at the end of the network maps the features to the interval of 0-1, and the samples with output values greater than the set threshold of 0.5 are determined to be abnormal. The test abnormal items represent the degree of abnormality of each instruction-event pair in the form of probability values, providing a quantitative basis for subsequent abnormality handling.

[0091] In some embodiments, the instruction feature representation vector and the event feature pyramid are input into a gated recurrent network for feature integration to generate an associated vector matrix of instructions and events, including: S341-S345.

[0092] S341. Project the instruction feature representation vector into parallel feature subspaces to obtain projected feature representation vectors, and perform a weighted combination operation on the projected feature representation vectors to obtain an instruction coding representation for context awareness.

[0093] Exemplarily, the construction of the projected feature subspaces adopts a multi-head attention mechanism. The instruction feature representation vector is divided into 8 attention heads, each with a dimension of 64. The features are mapped to different representation spaces through learnable projection matrices. Position encoding information is introduced during the projection process, and position embedding vectors generated by sine and cosine functions are used to capture the relative position relationships in the sequence. The weighted combination operation adopts an adaptive weight scheme, and the weight coefficients are normalized by the softmax function to ensure that the sum of the contribution degrees of the features in each subspace is 1. The instruction coding representation integrates local semantic information and global context information, with the dimension remaining 512, effectively enhancing the feature expression ability.

[0094] S342. Apply hierarchical feature recalibration to the event feature pyramid to obtain multi-scale discriminative event features through channel excitation and spatial selective filtering.

[0095] Exemplarily, the hierarchical feature recalibration network includes three branches, which respectively process feature maps of different scales. The channel excitation module calculates the global average pooling response of the feature map to generate a channel descriptor, learns the correlation between channels through a two-layer fully-connected network, and outputs a channel weight vector. The spatial selective filtering uses a 3×3 convolutional kernel with a stride of 1 and a padding of 1 to generate a spatial attention map, highlighting the feature responses in important regions. The multi-scale discriminative event features are integrated through a cross-scale feature fusion module, maintaining the hierarchical structure of the original feature pyramid while enhancing the discriminative ability of the features. This process significantly improves the expression quality of the event features, laying a foundation for subsequent feature fusion.

[0096] S343. Concatenate the instruction encoding representation and the multi-scale discriminative event features in a tensor to obtain a mixed feature matrix.

[0097] Exemplarily, during the generation of the mixed feature matrix, the instruction encoding representation and the multi-scale discriminative event features are concatenated in the channel dimension. The concatenation operation preserves the spatial resolution of the features, and padding operations are used to ensure the alignment of feature dimensions. The dimension of the concatenated feature matrix is H×W×(C1 + C2), where H and W represent the height and width of the feature map respectively, and C1 and C2 represent the number of channels of the instruction features and event features respectively.

[0098] S344. Perform threshold screening and non-linear transformation on the mixed feature matrix to obtain the correlated features with noise suppressed.

[0099] Exemplarily, a dual-threshold strategy is adopted for extracting the correlated features with noise suppressed. Set a high threshold th high = 0.8 and a low threshold th low = 0.2 to screen the elements in the mixed feature matrix. Features greater than the high threshold are directly retained, features less than the low threshold are suppressed, and features between the two thresholds are smoothed through a non-linear function. The non-linear transformation uses the LeakyReLU activation function with a negative slope set to 0.01, effectively avoiding the problem of gradient vanishing. This process significantly improves the robustness of the features and reduces the impact of noise on subsequent processing.

[0100] S345. Pass the correlated features with noise suppressed through a linear transformation layer, which includes a pre-trained weight matrix and a bias vector matrix, to obtain a correlated vector matrix.

[0101] Exemplarily, the weight matrix of the linear transformation layer is obtained through a pre-trained model, initialized with a normal distribution, with a mean of 0 and a standard deviation of 0.02. The bias vector matrix is initialized with zeros to ensure that the feature mapping does not shift in the initial state. The linear transformation process preserves the spatial structure of the features, and the dimension of the output correlated vector matrix is the same as the input. The vector at each position encodes the correlation relationship between the instruction and the event.

[0102] In some embodiments, determine the abnormal automotive lamp control system according to the test abnormal items and output a test abnormal report, including: S1041 - S1045.

[0103] S1041. Classify the test abnormal items into color and atmosphere deviation, brightness and atmosphere non-correspondence, brightness change and atmosphere non-correspondence, and color change and atmosphere non-correspondence.

[0104] Exemplarily, systematic classification processing is performed on the test abnormal items. By using a multi-dimensional abnormal feature analysis method, the abnormal items are divided into four main categories: color and atmosphere deviation types (including parameters such as color temperature deviation and chromaticity shift), brightness and atmosphere non-corresponding types (involving abnormal illuminance values, abnormal brightness uniformity, etc.), brightness change and atmosphere non-corresponding types (including dynamic features such as abnormal gradient process and abnormal response time), and color change and atmosphere non-corresponding types (involving abnormal color transition, dynamic color synchronization, etc.). During the classification process, real-time data collected by a spectral analyzer is combined to establish an abnormal feature vector space to ensure the accuracy and reliability of the classification results.

[0105] S1042. Determine the abnormal levels of various abnormal items of the automotive lamp control system according to a preset abnormal threshold.

[0106] Exemplarily, the abnormal level determination system conducts a quantitative evaluation based on a preset threshold, and sets the tolerance range of key parameters: color temperature deviation ±200K, illuminance deviation ±10%, response time deviation ±50ms, chromaticity shift △E≤3.0. The abnormal degree is divided into three levels: minor (Level 1), medium (Level 2), and severe (Level 3). Through a numerical mapping algorithm, the measurement data of various abnormal items is converted into a standardized score to achieve an accurate determination of the abnormal level.

[0107] S1043. Count the occurrence times of each type of abnormal item and the abnormal level, and set the automotive lamp control system in which the occurrence times of the abnormal item exceed a preset number threshold or the abnormal level exceeds a preset level as an abnormal automotive lamp control system.

[0108] Exemplarily, establish an abnormal evaluation matrix for the automotive lamp control system, record and analyze the distribution of the occurrence times of abnormal items and the abnormal levels of each automotive lamp control system. Set the abnormal determination criteria: when the cumulative occurrence times of "abnormal items" of a single automotive lamp control system exceed a preset number threshold, such as 5 times, or there are 2 or more occurrences of the abnormal level of Level 3 (preset level), or when the weighted score of the cumulative abnormal level exceeds a preset threshold of 15 points, mark this automotive lamp control system as an "abnormal automotive lamp control system". The weighted scoring method is adopted during the evaluation process, with Level 1 scored 1 point, Level 2 scored 3 points, and Level 3 scored 5 points, to ensure the scientificity and rationality of the abnormal determination through comprehensive scoring.

[0109] S1044. Associate the abnormal automotive lamp control system with its corresponding abnormal type to generate an abnormal correspondence table.

[0110] Exemplarily, an abnormal correspondence database is constructed, and a relational data structure is used to record the identification information, abnormal type code, abnormal level mark, and related test parameters of the abnormal lamp control system. The database design includes a primary key index, a foreign key constraint, and a trigger mechanism to ensure data integrity and consistency. The SQL query statement is used to achieve the rapid retrieval and statistical analysis of abnormal information, providing data support for the generation of abnormal reports.

[0111] S1045. Generate a test abnormal report according to the abnormal correspondence table. The test abnormal report includes a list of abnormal lamps, abnormal types, and abnormal degrees.

[0112] Exemplarily, the test abnormal report generation system adopts a modular design to convert the data in the abnormal correspondence table into a structured test report. The report content includes the detailed information of the abnormal lamp (model, location, control module, etc.), the specific description of the abnormal type (including the quantitative analysis of abnormal parameters), the level assessment of the abnormal degree (accompanied by a scoring standard description), and relevant test data charts. The report adopts a standardized format and supports multiple output formats such as PDF and HTML, facilitating data sharing and archive management. The abnormal report also includes test environment parameters, test equipment information, and test personnel certification information to ensure the integrity and traceability of the report.

[0113] Please refer to Figure 2 , Figure 2 FIG. is a schematic block diagram of an automatic test device for an automotive lamp control system provided by an embodiment of the present application. The automatic test device 200 for the automotive lamp control system is used to execute the aforementioned automatic test method for the automotive lamp control system. Among them, the automatic test device 200 for the automotive lamp control system can be configured in a server.

[0114] Among them, the server can be an independent server, a server cluster, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0115] As Figure 2 shown, the automatic test device 200 for the automotive lamp control system includes: a test control module 201, an event generation module 202, an abnormal identification module 203, and a result output module 204.

[0116] The test control module 201 is used to input a preset atmosphere simulation instruction to the automotive lamp control system to be tested, so that the automotive lamp control system outputs various atmosphere lights.

[0117] An event generation module 202, configured to receive ambient light through an event camera at a preset position and generate an ambient event stream.

[0118] An anomaly recognition module 203, configured to input an ambient simulation instruction and an ambient event stream into a trained LSTM model to obtain test anomaly items.

[0119] A result output module 204, configured to determine an abnormal automotive lamp control system according to the test anomaly items and output a test anomaly report.

[0120] An embodiment of the present application provides a test device, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement an automated test method for an automotive lamp control system according to any one of the embodiments of the present application when executing the computer program.

[0121] An embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor is caused to implement an automated test method for an automotive lamp control system according to any one of the embodiments of the present application.

[0122] As described above, the foregoing is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or substitutions within the technical scope disclosed by the present application, and these modifications or substitutions should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An automated testing method for an automobile lighting control system, characterized in that: The method comprises: Inputting a preset atmosphere simulation instruction to the automobile lamp control system to be tested, so that the automobile lamp control system outputs a variety of atmosphere lights; The event camera at the preset position is configured with parameters, and the contrast threshold and time interval parameters are set as event triggering conditions; based on the event camera after parameter configuration, the pixel log intensity change value of the ambient light is obtained, and compared with the contrast threshold to obtain the original event data representing the direction of brightness change; the original event data is divided into time series according to the timestamp, and the events captured between two adjacent frames are aggregated into multiple uniform time bins to obtain a time-discrete event set; the polarity information of the latest timestamp event of each pixel position in each time bin is obtained, and the time-discrete event set is spatially aggregated according to the polarity information to obtain multiple two-dimensional spatial tables. The method comprises the following steps: inputting the event slices represented by the two-dimensional space into a preset global spatial dependency extractor, calculating the correlation matrix between each pixel position and all other pixel positions through the self-attention mechanism of the global spatial dependency extractor, and obtaining a global spatial feature map; inputting the standardized event frame sequence into a preset GM-LSTM network for sorting, and generating a hidden state for each time step; extracting the distinguishing clues of the hidden states of adjacent time steps according to the global spatial feature map through differential operation and spatial attention mechanism, and obtaining an atmosphere event flow; Inputting the atmosphere simulation instruction and the atmosphere event stream into the trained LSTM model to obtain a test abnormality item; The abnormal automobile lighting control system is determined according to the test abnormality item, and a test abnormality report is output.

2. The automated testing method for an automobile lighting control system according to claim 1, characterized in that: The standardized event frame sequence is input into a preset global spatial dependency extractor, and the correlation matrix between each pixel position and all other pixel positions is calculated through the self-attention mechanism of the global spatial dependency extractor to obtain a global spatial feature map, including: According to the standardized event frame sequence, a two-dimensional feature matrix with a shape of H×W×C is constructed to obtain a spatial linear feature representation; Performing three-way feature projection on the spatial linear feature representation to obtain a query matrix, a key matrix, and a value matrix; Performing a transposed matrix multiplication operation on the query matrix and the key matrix, and dividing the product by a scaling factor to adjust the amplitude, thereby obtaining an original attention map; Applying a softmax normalization function to the original attention map, converting the similarity value into a probability distribution, and obtaining a normalized spatial attention weight matrix; Performing a matrix multiplication operation on the normalized spatial attention weight matrix and the value matrix to obtain a spatially dependent feature map; Performing a time-series transfer process on the spatially dependent feature map, connecting the feature state of the current time step with the hidden state of the previous time step, to obtain an enhanced feature representation; The enhanced feature representation is restored to a three-dimensional tensor H×W×C to obtain the global space feature map.

3. The automated testing method for an automobile lighting control system according to claim 1, characterized in that: The method extracts the distinguishing clues of the hidden states of adjacent time steps according to the global spatial feature map through differential operation and spatial attention mechanism to obtain the atmosphere event stream, including: Performing a subtraction operation on the previous time step output state of the hidden state and the current time step input state, and obtaining a discriminative feature by calculating a state difference tensor; The discriminative features are nonlinearly transformed through convolutional layers, batch normalization, and ReLU activation functions to obtain enhanced representations of the discriminative features; The enhancement of the discriminative features is performed by performing a channel mean operation and a Sigmoid activation function conversion to obtain a spatiotemporal perception mask; According to the spatiotemporal perception mask, the output state of the previous time step and the input state of the current time step are weighted to obtain a motion change correction vector; A channel-level connection operation is performed on the motion change correction vector, and feature fusion is performed through a convolutional layer to obtain the atmosphere event stream.

4. The automated testing method for an automobile lighting control system according to claim 1, characterized in that: The step of inputting the atmosphere simulation instruction and the atmosphere event stream into the trained LSTM model to obtain a test abnormality item includes: Input the atmosphere simulation instruction and the atmosphere event stream into a trained LSTM model, wherein the trained LSTM model includes a plurality of anomaly detection networks and a decision network, each of the anomaly detection networks includes a first long short-term memory network, a second long short-term memory network and a threshold recurrent network, and the decision network includes a fully connected layer and a Sigmoid function; Extract hidden state features of the atmosphere simulation instruction through the first long short-term memory network in each of the anomaly detection networks to obtain an instruction feature representation vector; Extract hidden state features of the atmosphere event stream through the second long short-term memory network in each of the anomaly detection networks to obtain an event feature pyramid; Inputting the instruction feature representation vector and the event feature pyramid into the threshold recurrent network for feature integration to generate a correlation vector matrix of instructions and events; The association vector matrix is ​​input into the decision network, and abnormal prediction is performed on each instruction and the corresponding event through the fully connected layer and the Sigmoid function in the decision network to obtain the test abnormal items.

5. The automated testing method for an automobile lighting control system according to claim 4, characterized in that: The step of inputting the instruction feature representation vector and the event feature pyramid into the threshold recurrent network for feature integration to generate an associated vector matrix of instructions and events includes: Projecting the instruction feature representation vector to a parallel feature subspace to obtain a projected feature representation vector, and performing a weighted combination operation on the projected feature representation vector to obtain an instruction encoding representation for context awareness; Applying hierarchical feature recalibration to the event feature pyramid, obtaining multi-scale discriminative event features through channel excitation and spatial selective filtering; Performing tensor concatenation of the instruction encoding representation and the multi-scale discriminative event features to obtain a mixed feature matrix; Performing threshold screening and nonlinear transformation on the mixed feature matrix to obtain correlation features for suppressing noise; The noise-suppressed correlation features are passed through a linear transformation layer, wherein the linear transformation layer includes a pre-trained weight matrix and a bias vector matrix, to obtain the correlation vector matrix.

6. The automated testing method for an automobile lighting control system according to claim 1, characterized in that: The determining of an abnormal automobile lamp control system according to the test abnormal item and outputting a test abnormality report includes: Classifying the test abnormal items into color and atmosphere deviation, brightness and atmosphere non-correspondence, brightness change and atmosphere non-correspondence, and color change and atmosphere non-correspondence; Determining various abnormal items of the automobile lighting control system according to a preset abnormality threshold and determining the abnormality level; Counting the number of occurrences and the abnormality level of each type of abnormal item, and setting the automobile lamp control system whose number of occurrences of the abnormal item exceeds a preset number threshold or whose abnormality level exceeds a preset level as an abnormal automobile lamp control system; Associating the abnormal automobile lamp control system with its corresponding abnormal type to generate an abnormality correspondence table; The test abnormality report is generated according to the abnormality correspondence table, and the test abnormality report includes a list of abnormal lamps, abnormality types and abnormality degrees.

7. An automatic test device for an automobile lighting control system, characterized in that: The automatic testing device for the automobile lamp control system is used to execute the automatic testing method for the automobile lamp control system according to any one of claims 1 to 6, and the automatic testing device for the automobile lamp control system comprises: A test control module, used for inputting preset atmosphere simulation instructions to the automobile lamp control system to be tested, so that the automobile lamp control system outputs a variety of atmosphere lights; The event generation module is used to configure the parameters of the event camera of the preset camera position, set the contrast threshold and time interval parameters as event triggering conditions; based on the event camera after parameter configuration, obtain the log intensity change value of the pixel point of the ambient light, and compare it with the contrast threshold to obtain the original event data representing the direction of brightness change; divide the original event data into time series according to the timestamp, aggregate the events captured between two adjacent frames into multiple uniform time bins, and obtain a time-discrete event set; obtain the polarity information of the latest timestamp event of each pixel position in each time bin, and spatially aggregate the time-discrete event set according to the polarity information to obtain multiple binary The method comprises the following steps: first, inputting the event slices represented by the two-dimensional space into a preset global spatial dependency extractor, and calculating the correlation matrix between each pixel position and all other pixel positions through the self-attention mechanism of the global spatial dependency extractor to obtain a global spatial feature map; inputting the standardized event frame sequence into a preset GM-LSTM network for sorting to generate a hidden state for each time step; and extracting the distinguishing clues of the hidden states of adjacent time steps according to the global spatial feature map through differential operation and spatial attention mechanism to obtain an atmosphere event flow; An abnormality identification module, used for inputting the atmosphere simulation instruction and the atmosphere event stream into the trained LSTM model to obtain a test abnormality item; The result output module is used to determine the abnormal automobile lighting control system according to the test abnormal items and output a test abnormality report.

Images