Intelligent traffic light identification and real-time prompting system for color blindness driver

Abstract

The invention discloses an intelligent traffic light identification and real-time prompting system for color blindness drivers. The system comprises an environment sensing module; the acquisition module is responsible for acquiring surrounding environment information; the parameter self-adaptive module is responsible for dynamically adjusting identification parameters of the system; the dynamic image processing module is responsible for processing and analyzing images acquired in real time; the signal lamp identification module is responsible for identifying colors and states of traffic lights; the decision and prompt module is responsible for making a corresponding driving decision and prompting a driver; according to the system, through the environment sensing module, the system can obtain and process the information of the surrounding environment in real time, the information comprises illumination conditions, weather conditions and the like, and basic data is provided for subsequent parameter adjustment and image processing. And the parameter adaptive module dynamically adjusts image processing parameters and signal lamp identification parameters according to the environment information, so that the adaptability of the system to different environments is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to an intelligent traffic light recognition and real-time prompting system for color-blind drivers. Background Art

[0002] Color blindness is a common visual disorder. Patients have difficulty distinguishing between red, green, blue and other colors, which brings great inconvenience to their transportation. Especially during driving, traffic lights are an important basis for judging the timing of driving. Color-blind drivers often find it difficult to accurately identify the status of traffic lights, thereby increasing the risk of traffic accidents. Intelligent transportation system refers to the organic combination of road networks, vehicles and equipment through computers, information technology, communication technology and other means to form an Internet of Things system, realize intelligent traffic management, and improve transportation efficiency, convenience and safety.

[0003] After searching, the invention patent with Chinese patent number CN101964114B discloses a traffic light auxiliary recognition system for red-green color blind drivers, which belongs to the field of image recognition technology. It solves the problems of high cost and small application scope of existing color auxiliary recognition technology. The traffic light auxiliary recognition system for red-green color blind drivers includes an image processing module, a display module and an image acquisition device. The image processing module processes the original image information acquired by the image acquisition device into image information recognizable to red-green color blind drivers, and sends it to the display module for display. The specific steps are as follows: a. Decompose each pixel of the original image information into three color channels of red, green and blue; b. Identify the color. When it is judged to be red or green, enter step c, otherwise enter step d; c. Automatically convert the pixel into a color recognizable to the red-green color blind driver; d. Do not process it. The traffic light auxiliary recognition system for red-green color blind drivers has high real-time performance and accuracy.

[0004] However, in actual use of the above invention, it is difficult to automatically adjust the recognition parameters according to environmental changes because the original image information of the traffic lights is collected by the image acquisition device. The traffic environment is complex and changeable. Factors such as lighting conditions and weather conditions will affect the visibility and color performance of the traffic lights, which may lead to misjudgment or missed judgment, thereby affecting driving safety. Therefore, an intelligent traffic light recognition and real-time prompt system for color-blind drivers is proposed. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art that the original image information of traffic lights is collected by image acquisition devices, and it is difficult to automatically adjust the recognition parameters according to environmental changes. The traffic environment is complex and changeable. Factors such as lighting conditions and weather conditions will affect the visibility and color performance of traffic lights, which may lead to misjudgment or missed judgment, thereby affecting driving safety. An intelligent traffic light recognition and real-time prompt system for color-blind drivers is proposed to solve the shortcomings in the prior art that the original image information of traffic lights is collected by image acquisition devices, and it is difficult to automatically adjust the recognition parameters according to environmental changes. The traffic environment is complex and changeable. Factors such as lighting conditions and weather conditions will affect the visibility and color performance of traffic lights, which may lead to misjudgment or missed judgment, thereby affecting driving safety.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] An intelligent traffic light recognition and real-time prompting system for color-blind drivers, comprising:

[0008] Environmental perception module; responsible for obtaining information about the surrounding environment;

[0009] Parameter adaptive module: responsible for dynamically adjusting the identification parameters of the system;

[0010] Dynamic image processing module: responsible for processing and analyzing the images acquired in real time;

[0011] Traffic light recognition module: responsible for identifying the color and status of traffic lights;

[0012] Decision-making and prompting module: responsible for making corresponding driving decisions and prompting the driver;

[0013] User feedback module: responsible for collecting feedback information from drivers;

[0014] The environment perception module is responsible for collecting environment information including images of traffic lights, surrounding light conditions, and weather conditions through sensors. The environment perception module transmits the collected environment information to the parameter adaptive adjustment module and the dynamic image processing module. The parameter adaptive adjustment module is responsible for dynamically adjusting image processing parameters such as exposure, contrast, color saturation, etc. and traffic light recognition parameters such as color threshold, shape features, etc. according to the environment information. The parameter adaptive adjustment module transmits the adjusted parameters to the dynamic image processing module and the traffic light recognition module. The dynamic image processing module processes and analyzes the real-time acquired images using computer vision technology based on the image information provided by the environment perception module and the image processing parameters provided by the parameter adaptive adjustment module. The state image processing module transmits the processed image and the extracted features to the signal light recognition module. The signal light recognition module uses a neural network algorithm to recognize the state of the signal light, such as red light, green light, yellow light, and effectively distinguish between signal lights and other light sources based on the image features and recognition parameters. The signal light recognition module transmits the recognition result to the decision and prompt module. The decision and prompt module makes corresponding driving decisions based on the recognition result and prompts the driver through voice or visual. For example, if a red light is detected, the system will prompt the driver to stop, and if a green light is detected, the system will prompt the driver to pass. The user feedback module collects feedback information from the driver, evaluates the system effect and user experience, and further optimizes system performance. For example, the recognition algorithm and prompt method are continuously improved through feedback.

[0015] The above technical solution further includes:

[0016] Preferably, the environment recognition module includes a camera unit, a lidar unit, an ultrasonic unit, a GPS unit, an IMU unit, a data fusion unit and a result output unit. The camera unit (including a front-view camera, a surround-view camera and a side-view camera) is used to capture images of the surrounding environment in real time, including traffic lights, traffic signs, vehicles and pedestrians. The lidar unit is used to provide high-precision three-dimensional environmental data to help identify the spatial position of traffic lights and surrounding obstacles. The ultrasonic unit is used to detect surrounding obstacles at close range and assist in determining the distance between the vehicle and surrounding objects. The GPS unit is used to provide the geographic location of the vehicle to help the system understand the current road and traffic light position of the vehicle. The IMU unit is used to detect the acceleration and angular velocity of the vehicle to help determine the dynamic state of the vehicle. The data fusion unit is used to integrate data from various units to generate a comprehensive environment perception model. The result output unit is used to pass the generated environment perception model data to the parameter adaptive adjustment module and the dynamic image processing module.

[0017] Preferably, the parameter adaptation module includes a parameter adjustment strategy unit and a parameter execution and feedback unit. The parameter adjustment strategy unit is responsible for formulating and adjusting strategies for image processing parameters and traffic light recognition parameters based on data provided by the environment recognition module, and automatically optimizing and adjusting parameters based on historical data and real-time data to improve recognition accuracy and system robustness. The parameter execution and feedback unit is responsible for passing the adjusted parameters to the dynamic image processing module and the traffic light recognition module, and receiving feedback from these modules during execution, evaluating the effect of the parameter adjustment, ensuring the real-time and accuracy of the parameter adjustment, and at the same time being able to monitor and adjust problems that may arise during the adjustment process, providing a basis for subsequent parameter optimization.

[0018] Preferably, the step of the dynamic image processing module using computer vision technology to process and analyze the image acquired in real time includes:

[0019] Data reception and preprocessing: Obtain real-time image data from the environment perception module, and perform preliminary settings on the image according to the parameters provided by the adaptive adjustment module, such as adjusting brightness, contrast, and filter intensity;

[0020] Image denoising: Apply median filtering to reduce image noise;

[0021] Feature extraction and representation: Extract key features such as color, texture, shape, and edge from the preprocessed image to help identify the target object in the image, and encode and represent the extracted features;

[0022] Result output: The results of preprocessing and feature extraction are output to the traffic light recognition module.

[0023] Preferably, the specific steps of the signal light recognition module using a neural network algorithm to recognize the state of the signal light include:

[0024] Input processing: receiving the output from the dynamic image processing module, that is, the image data that has been preprocessed and feature extracted, and further processing the received data to ensure that the input data meets the input requirements of the neural network model;

[0025] Application of neural network model: load the pre-trained neural network model, input the pre-processed image data into the neural network model, perform forward propagation calculation, and obtain the model's prediction result on the color of the traffic light;

[0026] Result processing and output: The output results of the neural network model are analyzed, and the output is a probability distribution representing the colors of the traffic lights. According to the probability distribution, the color with the highest probability is selected as the actual color of the traffic light, that is, the state of the traffic light (red light, yellow light, green light) is determined, and the recognition results are output to the decision and prompt module.

[0027] Preferably, the specific steps of the neural network model training include:

[0028] Data preparation: Collect a large amount of traffic light image data from various traffic scenarios, ensure that the data covers different lighting conditions, weather conditions, angles and distances, annotate the collected image data, clarify the location and color status of the traffic lights in each image, and preprocess the annotated image data, such as image resizing, normalization, and denoising;

[0029] Model selection and design: Select the convolutional neural network architecture, determine the number of network layers, convolution kernel size, pooling method, and fully connected layer parameters;

[0030] Model training: The labeled image data is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance. The learning rate, batch size, and number of training rounds training parameters are determined. The training set is input into the neural network model to perform forward propagation and back propagation calculations. By continuously adjusting the model parameters, the value of the loss function is reduced and the recognition accuracy of the model is improved. During the training process, the validation set is regularly used to evaluate the performance of the model.

[0031] Model evaluation and optimization: Use the test set to evaluate the recognition accuracy and robustness of the model. By comparing the recognition results under different test conditions, analyze the advantages and disadvantages of the model, and further optimize the model based on the test results and actual needs.

[0032] Preferably, the decision and prompt module includes a state analysis unit, a decision logic unit, a prompt generation unit and a behavior feedback unit. The state analysis unit is responsible for analyzing the output from the traffic light recognition module, including the state and position of the traffic light. The decision logic unit is responsible for making decisions (for example, slowing down, stopping, or continuing to drive) based on the analyzed traffic light state and the current state of the vehicle (such as speed and position). The prompt generation unit is responsible for generating prompt information for the driver to ensure that the information is transmitted clearly and effectively. The behavior feedback unit is responsible for adjusting the system's prompt method according to the driver's response (whether to stop or slow down in time).

[0033] Preferably, the prompt generating unit is implemented as follows:

[0034] Voice prompts: responsible for providing real-time voice prompts to the driver to inform the status of the traffic lights;

[0035] Visual prompts: responsible for communicating the status of signal lights to the driver through a visual interface;

[0036] Haptic prompts: responsible for enhancing the driver's perception of prompts through tactile feedback.

[0037] The present invention has the following beneficial effects:

[0038] 1. In the present invention, through the environment perception module, the system can obtain and process the information of the surrounding environment in real time, including lighting conditions, weather conditions, etc., to provide basic data for subsequent parameter adjustment and image processing. The parameter adaptive module dynamically adjusts the image processing parameters and signal light recognition parameters according to these environmental information, thereby improving the system's adaptability to different environments.

[0039] 2. In the present invention, the dynamic image processing module uses computer vision technology to process and analyze the real-time acquired images, which can effectively extract image features and reduce image distortion or blur caused by environmental factors. The signal light recognition module uses a neural network algorithm to recognize the state of the signal light based on image features and recognition parameters. This algorithm has strong learning and generalization capabilities and can accurately recognize the color and state of the signal light in a complex and changeable traffic environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a system architecture diagram of an intelligent traffic light recognition and real-time prompt system for color-blind drivers proposed by the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0042] like Figure 1 As shown, the present invention proposes an intelligent traffic light recognition and real-time prompt system for color-blind drivers, comprising:

[0043] Environmental perception module; responsible for obtaining information about the surrounding environment;

[0044] Parameter adaptive module: responsible for dynamically adjusting the identification parameters of the system;

[0045] Dynamic image processing module: responsible for processing and analyzing the images acquired in real time;

[0046] Traffic light recognition module: responsible for identifying the color and status of traffic lights;

[0047] Decision-making and prompting module: responsible for making corresponding driving decisions and prompting the driver;

[0048] User feedback module: responsible for collecting feedback information from drivers;

[0049] The environmental perception module is responsible for collecting environmental information through sensors. The environmental perception module passes the collected environmental information to the parameter adaptive adjustment module and the dynamic image processing module. The parameter adaptive adjustment module is responsible for dynamically adjusting the image processing parameters and the signal light recognition parameters according to the environmental information. The parameter adaptive adjustment module passes the adjusted parameters to the dynamic image processing module and the signal light recognition module. The dynamic image processing module uses computer vision technology to process and analyze the real-time acquired images based on the image information provided by the environmental perception module and the image processing parameters provided by the parameter adaptive adjustment module. The dynamic image processing module passes the processed image and the extracted features to the signal light recognition module. The signal light recognition module uses the neural network algorithm to identify the signal light status based on the image features and recognition parameters. The signal light recognition module passes the recognition results to the decision and prompt module. The decision and prompt module makes corresponding driving decisions based on the recognition results and prompts the driver through voice or visual. The user feedback module collects the driver's feedback information and evaluates the system effect and user experience.

[0050] In one embodiment, the environment recognition module includes a camera unit, a lidar unit, an ultrasonic unit, a GPS unit, an IMU unit, a data fusion unit and a result output unit. The camera unit is used to capture images of the surrounding environment in real time, the lidar unit is used to provide high-precision three-dimensional environmental data, the ultrasonic unit is used to detect surrounding obstacles at close range, the GPS unit is used to provide the vehicle's geographic location, the IMU unit is used to detect the vehicle's acceleration and angular velocity, the data fusion unit is used to integrate data from various units to generate a comprehensive environment perception model, and the result output unit is used to pass the generated environment perception model data to the parameter adaptive adjustment module and the dynamic image processing module.

[0051] In the embodiment of the present invention, the camera unit continuously acquires image frames through the front-view camera, the surround-view camera, and the side-view camera, improves the image quality by using the image processing algorithm, and transmits the processed image data to the data fusion unit. The laser radar unit generates point cloud data by emitting lasers and receiving reflected signals. Each point cloud data point contains position coordinates (X, Y, Z). After filtering and aggregation, the point cloud data is transmitted to the data fusion unit to form a three-dimensional model of the environment. The ultrasonic unit ultrasonic sensor emits sound waves and receives reflected waves, calculates the distance according to the time difference, provides the location information of the close-range obstacles according to the generated distance data, and sends the detection results to the data fusion unit. The GPS unit obtains the latitude and longitude information of the vehicle through the satellite signal receiver, and integrates the location information into the data fusion unit. The IMU unit monitors the movement changes of the vehicle through the accelerometer and gyroscope to calculate the direction, speed and posture changes of the vehicle, and transmits these dynamic information to the data fusion unit to enhance the accuracy of the environmental model.

[0052] In one embodiment, the parameter adaptation module includes a parameter adjustment strategy unit and a parameter execution and feedback unit. The parameter adjustment strategy unit is responsible for formulating and adjusting strategies for image processing parameters and traffic light recognition parameters based on data provided by the environment recognition module. The parameter execution and feedback unit is responsible for passing the adjusted parameters to the dynamic image processing module and the traffic light recognition module, receiving feedback from these modules during execution, and evaluating the effect of the parameter adjustment.

[0053] In the embodiment of the present invention, the data received by the parameter adjustment strategy unit includes light intensity, weather conditions, traffic flow, etc., which have an important impact on the selection of image processing parameters and signal light recognition parameters. Based on the received data, the parameter adjustment strategy unit will formulate and adjust strategies for image processing parameters (such as brightness, contrast, denoising intensity, etc.) and signal light recognition parameters (such as recognition threshold, feature weight, etc.). These strategies enable the system to maintain optimal performance under different environmental conditions.

[0054] In one embodiment, the steps of the dynamic image processing module using computer vision technology to process and analyze the real-time acquired image include:

[0055] Data reception and preprocessing: Obtain real-time image data from the environment perception module and perform preliminary settings on the image according to the parameters provided by the adaptive adjustment module;

[0056] Image denoising: Apply median filtering to reduce image noise;

[0057] Feature extraction and representation: Extract key features from the preprocessed image, and encode and represent the extracted features;

[0058] Result output: The results of preprocessing and feature extraction are output to the traffic light recognition module.

[0059] In one embodiment, the specific steps of the traffic light recognition module using the neural network algorithm to recognize the state of the traffic light include:

[0060] Input processing: receiving the output from the dynamic image processing module and further processing the received data to ensure that the input data meets the input requirements of the neural network model;

[0061] Application of neural network model: load the pre-trained neural network model, input the pre-processed image data into the neural network model, perform forward propagation calculation, and obtain the model's prediction result on the color of the traffic light;

[0062] Result processing and output: The output results of the neural network model are analyzed. The output is a probability distribution representing the colors of the traffic lights. According to the probability distribution, the color with the highest probability is selected as the actual color of the traffic light, that is, the state of the traffic light is determined, and the recognition results are output to the decision and prompt module.

[0063] In the embodiment of the present invention, the output from the dynamic image processing module is received, and the image size is adjusted and standardized to ensure that the input data meets the requirements of the neural network model. The formula for standardizing the image can be expressed as:

[0064]

[0065] Where I represents the original image, u and σ represent the mean and standard deviation of the image respectively.

[0066] In one embodiment, the specific steps of neural network model training include:

[0067] Data preparation: Collect a large amount of traffic light image data from various traffic scenes, annotate the collected image data, and preprocess the annotated image data;

[0068] Model selection and design: Select the convolutional neural network architecture, determine the number of network layers, convolution kernel size, pooling method, and fully connected layer parameters;

[0069] Model training: The labeled image data is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance. The learning rate, batch size, and number of training rounds training parameters are determined. The training set is input into the neural network model to perform forward propagation and back propagation calculations. During the training process, the validation set is regularly used to evaluate the performance of the model.

[0070] Model evaluation and optimization: Use the test set to evaluate the recognition accuracy and robustness of the model, and further optimize the model based on the test results and actual needs.

[0071] In one embodiment, the decision and prompt module includes a state analysis unit, a decision logic unit, a prompt generation unit and a behavior feedback unit. The state analysis unit is responsible for analyzing the output from the traffic light recognition module. The decision logic unit is responsible for making decisions based on the analyzed traffic light state and the current state of the vehicle. The prompt generation unit is responsible for generating prompt information for the driver to ensure that the information is transmitted clearly and effectively. The behavior feedback unit is responsible for adjusting the system's prompt method according to the driver's response.

[0072] In the embodiment of the present invention, the state analysis may use Boolean logic to represent the state of the signal light, for example

[0073]

[0074] Among them, S represents the state of the traffic light.

[0075] The decision logic is based on the combination of the state and the current state of the vehicle, and its formula can be expressed as:

[0076] D=f(S,V,T)

[0077] Wherein, D represents the decision result (such as continue driving or stop), S represents the traffic light state, V represents the vehicle speed; and T represents the time from the traffic light.

[0078] The prompt information is generated by weight scoring, and its formula can be expressed as:

[0079] P=W1·S+W2·V+W3·T

[0080] Among them, P represents the prompt information weight, and W1, W2, and W3 represent the corresponding weight coefficients.

[0081] In one embodiment, the prompt generating unit is implemented as follows:

[0082] Voice prompts: responsible for providing real-time voice prompts to the driver to inform the status of the traffic lights;

[0083] Visual prompts: responsible for communicating the status of signal lights to the driver through a visual interface;

[0084] Haptic prompts: responsible for enhancing the driver's perception of prompts through tactile feedback.

[0085] In an embodiment of the present invention, the voice prompt uses speech synthesis technology to convert text information into speech. Design a variety of voice prompt schemes according to actual scenarios, such as "Red light ahead, please stop and wait" and "Green light ahead, you can pass". Adjust the volume, speed and tone of the voice prompt to ensure its clarity and ease of understanding. Visual prompts simulate the actual status of traffic lights by displaying traffic light icons or color blocks on the display. Use obvious color contrast and dynamic effects (such as flashing) to emphasize the importance of prompt information. Tactile prompts design a variety of vibration modes to correspond to different traffic light states. Adjust the intensity and frequency of the vibration to ensure that it is easy to perceive and does not interfere with the driver's attention.

[0086] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent traffic light recognition and real-time prompt system for color-blind drivers, characterized in that: include: Environmental perception module; responsible for obtaining information about the surrounding environment; Parameter adaptive module: responsible for dynamically adjusting the identification parameters of the system; Dynamic image processing module: responsible for processing and analyzing images acquired in real time; Traffic light recognition module: responsible for identifying the color and status of traffic lights; Decision-making and prompting module: responsible for making corresponding driving decisions and prompting the driver; User feedback module: responsible for collecting feedback information from drivers; The environment perception module is responsible for collecting environment information through sensors. The environment perception module transmits the collected environment information to the parameter adaptive adjustment module and the dynamic image processing module. The parameter adaptive adjustment module is responsible for dynamically adjusting the image processing parameters and the signal light recognition parameters according to the environment information. The parameter adaptive adjustment module transmits the adjusted parameters to the dynamic image processing module and the signal light recognition module. The dynamic image processing module processes and analyzes the real-time acquired image using computer vision technology according to the image information provided by the environment perception module and the image processing parameters provided by the parameter adaptive adjustment module. The dynamic image processing module transmits the processed image and the extracted features to the signal light recognition module. The signal light recognition module uses a neural network algorithm to identify the signal light status based on the image features and the recognition parameters. The signal light recognition module transmits the recognition result to the decision and prompt module. The decision and prompt module makes corresponding driving decisions according to the recognition result and prompts the driver through voice or visual. The user feedback module collects the driver's feedback information and evaluates the system effect and user experience.

2. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 1, characterized in that: The environment recognition module includes a camera unit, a laser radar unit, an ultrasonic unit, a GPS unit, an IMU unit, a data fusion unit and a result output unit. The camera unit is used to capture images of the surrounding environment in real time, the laser radar unit is used to provide high-precision three-dimensional environmental data, the ultrasonic unit is used to detect surrounding obstacles at close range, the GPS unit is used to provide the geographic location of the vehicle, the IMU unit is used to detect the acceleration and angular velocity of the vehicle, the data fusion unit is used to integrate data from various units to generate a comprehensive environment perception model, and the result output unit is used to pass the generated environment perception model data to the parameter adaptive adjustment module and the dynamic image processing module.

3. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 1, characterized in that: The parameter adaptation module includes a parameter adjustment strategy unit and a parameter execution and feedback unit. The parameter adjustment strategy unit is responsible for formulating and adjusting the strategy of image processing parameters and traffic light recognition parameters based on the data provided by the environment recognition module. The parameter execution and feedback unit is responsible for passing the adjusted parameters to the dynamic image processing module and the traffic light recognition module, and receiving feedback from these modules during the execution process to evaluate the effect of the parameter adjustment.

4. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 1, characterized in that: The steps of the dynamic image processing module using computer vision technology to process and analyze the real-time acquired image include: Data reception and preprocessing: Obtain real-time image data from the environment perception module and perform preliminary settings on the image according to the parameters provided by the adaptive adjustment module; Image denoising: Apply median filtering to reduce image noise; Feature extraction and representation: Extract key features from the preprocessed image, and encode and represent the extracted features; Result output: The results of preprocessing and feature extraction are output to the traffic light recognition module.

5. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 1, characterized in that: The specific steps of the signal light recognition module using the neural network algorithm to recognize the state of the signal light include: Input processing: receiving the output from the dynamic image processing module and further processing the received data to ensure that the input data meets the input requirements of the neural network model; Application of neural network model: load the pre-trained neural network model, input the pre-processed image data into the neural network model, perform forward propagation calculation, and obtain the model's prediction result on the color of the traffic light; Result processing and output: The output results of the neural network model are analyzed. The output is a probability distribution representing the colors of the traffic lights. According to the probability distribution, the color with the highest probability is selected as the actual color of the traffic light, that is, the state of the traffic light is determined, and the recognition results are output to the decision and prompt module.

6. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 5, characterized in that: The specific steps of the neural network model training include: Data preparation: Collect a large amount of traffic light image data from various traffic scenes, annotate the collected image data, and preprocess the annotated image data; Model selection and design: Select the convolutional neural network architecture, determine the number of network layers, convolution kernel size, pooling method, and fully connected layer parameters; Model training: The labeled image data is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance. The learning rate, batch size, and number of training rounds training parameters are determined. The training set is input into the neural network model to perform forward propagation and back propagation calculations. During the training process, the validation set is regularly used to evaluate the performance of the model. Model evaluation and optimization: Use the test set to evaluate the recognition accuracy and robustness of the model, and further optimize the model based on the test results and actual needs.

7. The intelligent traffic light recognition and real-time prompt system for color blind drivers according to claim 1, characterized in that: The decision and prompt module includes a state analysis unit, a decision logic unit, a prompt generation unit and a behavior feedback unit. The state analysis unit is responsible for analyzing the output from the traffic light recognition module. The decision logic unit is responsible for making decisions based on the analyzed traffic light state and the current state of the vehicle. The prompt generation unit is responsible for generating prompt information for the driver to ensure that the information is transmitted clearly and effectively. The behavior feedback unit is responsible for adjusting the system's prompt method according to the driver's response.

8. The intelligent traffic light recognition and real-time prompting system for color-blind drivers according to claim 7, characterized in that: The implementation of the prompt generating unit includes: Voice prompts: responsible for providing real-time voice prompts to the driver to inform the status of the traffic lights; Visual prompts: responsible for communicating the status of signal lights to the driver through a visual interface; Haptic prompts: responsible for enhancing the driver's perception of prompts through tactile feedback.

Images