Traffic signal lamp recognition system for assisting red-green blindness drivers

Abstract

The utility model provides a traffic light recognition system for assisting a red-green blindness driver, which belongs to the field of image recognition and comprises a main controller, an image acquisition module, a voice module, a display module and a traffic light recognition model, and the image acquisition module, the voice module and the display module are all connected to the main controller. The main controller and the image acquisition module are both wirelessly connected to the traffic light identification model, the traffic light identification model is a target detection model YOLOv8, the image acquisition module acquires images of traffic lights and wirelessly transmits the images of the traffic lights to the traffic light identification model, and the traffic light identification model identifies and detects the images of the traffic lights. The traffic light identification model adopts a target detection model YOLOv8 and can timely, accurately and wirelessly transmit a detection result to the main controller, the display module is used for displaying the detection result of the traffic light, the voice module is used for broadcasting the detection result of the traffic light, and the display module and the voice module form double insurance.

Description

Technical Field

[0001] The utility model relates to the technical field of image recognition, especially traffic signal recognition. Background Art

[0002] Color vision impairment is usually called color blindness. Color-blind patients cannot distinguish various colors or a certain color in the natural spectrum, and those with poor color discrimination ability are called color weakness. The boundary between color weakness and color blindness is not easy to strictly distinguish, only the degree of severity is different. Red-green color blindness is one of them. Red-green color-blind patients cannot distinguish red and green. According to the laws of our country, if a person suffers from red-green color blindness, he or she is not able to obtain a motor vehicle driving license. The main reason for this is that red-green color-blind patients cannot distinguish traffic signals in a timely and accurate manner, posing a great potential traffic hazard. According to statistics, about 7% of male red-green color-blind patients and about 1% of female color-blind patients in our country. With the development of the times, driving a car has become a necessary activity in daily life. Red-green color-blind patients may also obtain the qualification to drive a car in the near future. Therefore, it is very necessary to develop an identification system that can assist red-green color-blind drivers in identifying traffic signals.

[0003] Red-green color blindness is usually a genetic trait. At present, there is no complete cure for this condition, and only some auxiliary methods can be used to relieve the inconvenience in daily life. Color-blind glasses are a relatively effective auxiliary. It adjusts light of specific wavelengths to enhance or filter specific colors, which is beneficial to improving the visual contrast of red-green color-blind patients. Wearing color-blind glasses when driving a motor vehicle is helpful for distinguishing traffic signals, but there are also some deficiencies: the effects of color-blind glasses vary from person to person. The types and degrees of color blindness of different individuals are different, and the color-blind glasses may change under different lighting conditions. Vehicle driving requires a high degree of concentration. Any small mistake may lead to a major accident. The effects of color-blind glasses are not stable enough and can only be used in daily life, and are not well applicable to red-green color-blind patients driving motor vehicles. In the prior art, there is an auxiliary method that captures the status of traffic signals through a camera, converts the red pixels to black, and converts the green pixels to white, and displays them through a liquid crystal display screen. However, this method may cause misjudgment in complex scenarios (such as complex scenarios with traffic flow, noise, etc.), and it is necessary to divert attention to observe the screen, which is not conducive to the driver's concentrated driving. Content of the Utility Model

[0004] The technical problem to be solved by the utility model is how to improve the timeliness and accuracy of red-green color-blind drivers in identifying traffic signals.

[0005] The present utility model solves the above technical problems through the following technical solutions: A traffic signal recognition system for assisting red-green color-blind drivers, comprising a main controller, an image acquisition module, a voice module, a display module, and a traffic light recognition model. The image acquisition module, the voice module, and the display module are all connected to the main controller, and the main controller and the image acquisition module are both wirelessly connected to the traffic light recognition model, and the traffic light recognition model is the target detection model YOLOv8.

[0006] The present utility model acquires the image of the traffic signal through the image acquisition module and wirelessly transmits the image of the traffic signal to the traffic light recognition model. The traffic light recognition model performs recognition and detection on the image of the traffic signal. The traffic light recognition model adopts the target detection model YOLOv8, which has a fast inference speed and can wirelessly transmit the detection result to the main controller in a timely and accurate manner. The display module is used to display the detection result of the traffic signal, and the voice module is used to broadcast the detection result of the traffic signal. The display module and the voice module convey the traffic signal information to the red-green color-blind driver, forming a double insurance to avoid traffic hazards caused by the red-green color-blind driver not obtaining the traffic light status information in a timely manner.

[0007] Preferably, it further includes a cloud platform and a WiFi module. The main controller and the image acquisition module are both wirelessly connected to the cloud platform through the WiFi module, and the traffic light recognition model is located on the cloud platform.

[0008] Preferably, the cloud platform further includes a detection script.

[0009] Preferably, the main controller includes a chip U1, the model of the chip U1 is STM32F103RCT6. The first pin and the second pin of the chip U1 are both connected to the +5V voltage. The third pin and the fifth pin of the chip U1 are connected and then grounded. The seventh pin and the tenth pin of the chip U1 are both connected to the WiFi module. The ninth pin, the twelfth pin, the forty-second pin, the forty-third pin, the forty-fourth pin, the forty-fifth pin, the forty-sixth pin, the forty-seventh pin, the forty-eighth pin, the forty-ninth pin, the fiftieth pin, the fifty-first pin, the fifty-second pin, the fifty-third pin, the fifty-fourth pin, and the fifty-fifth pin of the chip U1 are all connected to the image acquisition module. The fifteenth pin and the sixteenth pin of the chip U1 are connected to the voice module. The twenty-ninth pin and the thirtieth pin of the chip U1 are connected to the display module. The thirty-first pin and the thirty-third pin of the chip U1 are connected and then connected to the +3.3V voltage. The thirty-second pin and the thirty-fourth pin of the chip U1 are connected and then connected to the +3.3V voltage. The fourth pin, the thirty-fifth pin, the thirty-sixth pin, and the fifty-ninth pin of the chip U1 are all grounded.

[0010] Preferably, the image acquisition module is a camera, including chip U2, the model of chip U2 is OV7670. The first pin and the fifteenth pin of chip U2 are connected to +3.3V voltage. The second pin and the sixteenth pin of chip U2 are grounded. The third pin of chip U2 is connected to the ninety-fifth pin of chip U1. The fourth pin of chip U2 is connected to the one-hundred-and-twenty-sixth pin of chip U1. The fifth pin of chip U2 is connected to the fifty-fifth pin of chip U1. The seventh pin of chip U2 is connected to the forty-fourth pin of chip U1. The eighth pin of chip U2 is connected to the forty-fifth pin of chip U1. The ninth pin of chip U2 is connected to the forty-sixth pin of chip U1. The tenth pin of chip U2 is connected to the forty-seventh pin of chip U1. The eleventh pin of chip U2 is connected to the forty-eighth pin of chip U1. The twelfth pin of chip U2 is connected to the forty-ninth pin of chip U1. The thirteenth pin of chip U2 is connected to the fiftieth pin of chip U1. The fourteenth pin of chip U2 is connected to the fifty-first pin of chip U1. The eighteenth pin of chip U2 is connected to the forty-third pin of chip U1. The nineteenth pin of chip U2 is connected to the fifty-fourth pin of chip U1. The twentieth pin of chip U2 is connected to the fifty-second pin of chip U1. The twenty-first pin of chip U2 is connected to the forty-second pin of chip U1. The twenty-second pin of chip U2 is connected to the fifty-third pin of chip U1.

[0011] Preferably, the WiFi module includes chip U3, the model of chip U3 is ESP8266. The first pin of chip U3 is connected to the voice module and connected to 5V power supply. The second pin of chip U3 is grounded. The third pin of chip U3 is connected to the tenth pin of chip U1. The fifth pin of chip U3 is connected to the seventh pin of chip U1.

[0012] Preferably, the voice module includes chip U4, the model of chip U4 is SYN6288. The twelfth pin of chip U4 is connected to 5V power supply. The fourteenth pin of chip U4 is grounded. The twenty-third pin of chip U4 is connected to the fifteenth pin of chip U1. The twenty-fourth pin of chip U4 is connected to the sixteenth pin of chip U1.

[0013] Preferably, the display module is an OLED screen, including chip U5, the model of chip U5 is 0.96OLED. The first pin of chip U5 is grounded. The second pin of chip U5 is connected to +5V voltage. The third pin of chip U5 is connected to the thirtieth pin of chip U1. The fourth pin of chip U5 is connected to the twenty-ninth pin of chip U1.

[0014] The advantages provided by the present utility model are as follows:

[0015] 1. The utility model collects the image of the traffic signal lamp through the image acquisition module, and wirelessly transmits the image of the traffic signal lamp to the traffic signal lamp recognition model. The traffic signal lamp recognition model performs recognition and detection on the image of the traffic signal lamp. The traffic signal lamp recognition model adopts the object detection model YOLOv8, which has a fast inference speed and can wirelessly transmit the detection result to the main controller in a timely and accurate manner. The display module is used to display the detection result of the traffic signal lamp, and the voice module is used to broadcast the detection result of the traffic signal lamp. The display module and the voice module convey the traffic signal lamp information to the red-green color-blind driver, forming a double insurance to avoid traffic hazards caused by the red-green color-blind driver not obtaining the traffic signal lamp status information in a timely manner.

[0016] 2. The traffic signal lamp recognition model of the utility model adopts the object detection model YOLOv8. This model has high requirements for computing power. Introducing the cloud platform can well solve this problem. The use of the cloud platform reduces the performance requirements for the main control chip and greatly reduces the manufacturing cost of the device without affecting the system performance. Brief Description of the Drawings

[0017] Figure 1 is a schematic diagram of the traffic signal lamp recognition system for assisting red-green color-blind drivers provided by the utility model;

[0018] Figure 2 is the circuit diagram of the traffic signal lamp recognition system for assisting red-green color-blind drivers provided by the utility model. Detailed Embodiments

[0019] To make the objectives, technical solutions, and advantages of the utility model clearer and more understandable, the following combines specific embodiments and refers to the attached drawings to clearly and completely describe the technical solutions of the utility model. Obviously, the described embodiments are some, but not all, of the embodiments of the utility model. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the utility model belong to the scope of protection of the utility model.

[0020] As Figure 1 shown, this embodiment provides a traffic signal lamp recognition system for assisting red-green color-blind drivers, including a main controller, an image acquisition module, a voice module, a display module, and a traffic signal lamp recognition model. The image acquisition module, the voice module, and the display module are all connected to the main controller. The main controller and the image acquisition module are both wirelessly connected to the traffic signal lamp recognition model. The image acquisition module collects the image of the traffic signal lamp and wirelessly transmits the image of the traffic signal lamp to the traffic signal lamp recognition model. The traffic signal lamp recognition model performs recognition and detection on the image of the traffic signal lamp, and the detection result is wirelessly transmitted to the main controller. The display module is used to display the detection result of the traffic signal lamp, and the voice module is used to broadcast the detection result of the traffic signal lamp.

[0021] It further includes a cloud platform, a WiFi module. The main controller and the image acquisition module are wirelessly connected to the cloud platform through the WiFi module. The traffic light recognition model is located on the cloud platform. The image acquisition module transmits the acquired image of the traffic light to the cloud platform through the WiFi module, and the traffic light recognition model performs recognition and detection on the image of the traffic light.

[0022] The cloud platform further includes a detection script. The traffic light recognition model performs recognition and detection on the image of the traffic light and generates a detection result file. The detection script reads the detection result file and sends the detection result to the main controller through the WiFi module.

[0023] As Figure 2 shown, the main controller includes a chip U1, the model of chip U1 is STM32F103RCT6, the image acquisition module is a camera, including a chip U2, the model of chip U2 is OV7670, the WiFi module includes a chip U3, the model of chip U3 is ESP8266, the voice module includes a chip U4, the model of chip U4 is SYN6288, and the display module is an OLED screen, including a chip U5, the model of chip U5 is 0.96OLED1.

[0024] The first pin and the second pin of chip U1 are both connected to the +5V voltage. The third pin and the fifth pin of chip U1 are connected together and then grounded. The seventh pin of chip U1 is connected to the fifth pin of chip U3. The ninth pin of chip U1 is connected to the third pin of chip U2. The tenth pin of chip U1 is connected to the third pin of chip U3. The twelfth pin of chip U1 is connected to the fourth pin of chip U2. The fifteenth pin of chip U1 is connected to the twenty-third pin of chip U4. The sixteenth pin of chip U1 is connected to the twenty-fourth pin of chip U4. The twenty-ninth pin of chip U1 is connected to the fourth pin of chip U5. The thirtieth pin of chip U1 is connected to the third pin of chip U5. The thirty-first pin and the thirty-third pin of chip U1 are connected together and then connected to the +3.3V voltage. The thirty-second pin and the thirty-fourth pin of chip U1 are connected together and then connected to the +3.3V voltage. The forty-second pin of chip U1 is connected to the twenty-first pin of chip U2. The forty-third pin of chip U1 is connected to the eighteenth pin of chip U2. The forty-fourth pin of chip U1 is connected to the seventh pin of chip U2. The forty-fifth pin of chip U1 is connected to the eighth pin of chip U2. The forty-sixth pin of chip U1 is connected to the ninth pin of chip U2. The forty-seventh pin of chip U1 is connected to the tenth pin of chip U2. The forty-eighth pin of chip U1 is connected to the eleventh pin of chip U2. The forty-ninth pin of chip U1 is connected to the twelfth pin of chip U2. The fiftieth pin of chip U1 is connected to the thirteenth pin of chip U2. The fifty-first pin of chip U1 is connected to the fourteenth pin of chip U2. The fifty-second pin of chip U1 is connected to the twentieth pin of chip U2. The fifty-third pin of chip U1 is connected to the twenty-second pin of chip U2. The fifty-fourth pin of chip U1 is connected to the nineteenth pin of chip U2. The fifty-fifth pin of chip U1 is connected to the fifth pin of chip U2. The fourth pin, the thirty-fifth pin, the thirty-sixth pin, and the fifty-ninth pin of chip U1 are all grounded. The first pin and the fifteenth pin of chip U2 are connected to the +3.3V voltage. The second pin and the sixteenth pin of chip U2 are grounded. The first pin of chip U3 is connected to the twelfth pin of chip U4. The first pin of chip U3 and the twelfth pin of chip U4 are both connected to the +5V voltage. The second pin of chip U3 is grounded. The fourteenth pin of chip U4 and the first pin of chip U5 are both grounded. The second pin of chip U5 is connected to the +5V voltage.

[0025] The traffic light recognition model of the present utility model adopts the existing object detection model YOLOv8, which can greatly improve the real-time performance and accuracy of image detection of traffic lights. The processing process of the object detection model YOLOv8 for the images of traffic lights is prior art, including preprocessing operations such as resizing the images and normalizing the image values. These operations ensure that the input images can be correctly processed by the model. The preprocessed images are forward-propagated through the deep neural network of the object detection model YOLOv8. In the first half, image features are extracted, and in the second half, multiple detection heads are responsible for generating bounding boxes and class confidence scores. The forward propagation generates a series of bounding boxes and class confidence scores. The bounding boxes with lower class confidence scores are removed to improve the detection accuracy. The overlapping bounding boxes are processed, and the box with the highest confidence is retained to eliminate redundant detection results. The final output is the position information (coordinates of the bounding box) of the object, the class label, and the corresponding confidence score, generating the corresponding detection result file.

[0026] The object detection model YOLOv8 used in the present utility model is a trained model. During the training process of this model, a dataset on traffic lights can be collected through methods such as dash cams and online collection. The dataset is manually labeled using annotation software, with the traffic lights boxed and given labels (red light state, green light state) to complete the dataset. The dataset is divided into a training set and a validation set, and data augmentation is performed on the training set, including operations such as random rotation, flipping, and scaling, to enhance the generalization ability of the model. The object detection model YOLOv8 is used for object detection training. The image features of traffic lights are extracted through a convolutional neural network, and continuous iteration is carried out until the accuracy and recall rate reach a balance, realizing the availability of the model.

[0027] In the present utility model, the traffic light recognition model performs recognition and detection on the images of traffic lights, and the detection results are wirelessly transmitted to the main controller. When the detection result is green, the main controller outputs signals to the display module and the voice module. The display module displays a white circle, and the voice module broadcasts "The current is a green light". When the detection result is red, the main controller outputs signals to the display module and the voice module. The display module displays a black triangle, and the voice module broadcasts "The current is a red light". The display through colors and shapes can improve the recognition rate and facilitate the driver's judgment. The display module and the voice module are enabled simultaneously, providing double insurance to fully ensure that the driver can obtain traffic light information in real time and accurately.

[0028] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A traffic light recognition system for assisting red-green color blind drivers, characterized by: It includes a main controller, an image acquisition module, a voice module, a display module, a traffic light recognition model, a cloud platform, and a WiFi module. The image acquisition module, voice module, and display module are all connected to the main controller. The main controller and the image acquisition module are both wirelessly connected to the traffic light recognition model. The traffic light recognition model is the target detection model YOLOv8. The main controller and the image acquisition module are both wirelessly connected to the cloud platform via the WiFi module. The traffic light recognition model is located on the cloud platform. The cloud platform also includes a detection script.

2. The traffic light recognition system for assisting red-green color blind drivers according to claim 1, characterized in that: The main controller includes a chip U1, the model of which is STM32F103RCT6. The first and second pins of the chip U1 are both connected to a +5V voltage, the third pin of the chip U1 is connected to the fifth pin and then to ground, the seventh and tenth pins of the chip U1 are both connected to the WiFi module, the ninth, twelfth, forty-second, forty-third, forty-fourth, forty-fifth, forty-sixth, forty-seventh, forty-eighth, forty-ninth, fiftieth, fifty-first, fifty-second, fifty-third, fifty-fourth and fifty-fifth pins of the chip U1 are all connected to the image acquisition module, the fifteenth and sixteenth pins of the chip U1 are connected to the voice module, the twenty-ninth and thirtieth pins of the chip U1 are connected to the display module, the thirty-first and thirty-third pins of the chip U1 are connected to the +3.3V voltage, the thirty-second and thirty-fourth pins of the chip U1 are connected to the +3.3V voltage, and the fourth, thirty-fifth, thirty-sixth and fifty-ninth pins of the chip U1 are all grounded.

3. The traffic light recognition system for assisting red-green color blind drivers according to claim 2, characterized in that: The image acquisition module is a camera, including a chip U2, the model of the chip U2 is OV7670, the first pin and the fifteenth pin of the chip U2 are connected to the +3.3V voltage, the second pin and the sixteenth pin of the chip U2 are grounded, the third pin of the chip U2 is connected to the ninety-fifth pin of the chip U1, the fourth pin of the chip U2 is connected to the twelfth pin of the chip U1, the fifth pin of the chip U2 is connected to the fifty-fifth pin of the chip U1, the seventh pin of the chip U2 is connected to the forty-fourth pin of the chip U1, the eighth pin of the chip U2 is connected to the forty-fifth pin of the chip U1, the ninth pin of the chip U2 is connected to the forty-sixth pin of the chip U1, and the tenth pin of the chip U2 is connected to the 1, the eleventh pin of chip U2 is connected to the forty-eighth pin of chip U1, the twelfth pin of chip U2 is connected to the forty-ninth pin of chip U1, the thirteenth pin of chip U2 is connected to the fiftieth pin of chip U1, the fourteenth pin of chip U2 is connected to the fifty-first pin of chip U1, the eighteenth pin of chip U2 is connected to the forty-third pin of chip U1, the nineteenth pin of chip U2 is connected to the fifty-fourth pin of chip U1, the twentieth pin of chip U2 is connected to the fifty-second pin of chip U1, the twenty-first pin of chip U2 is connected to the forty-second pin of chip U1, and the twenty-second pin of chip U2 is connected to the fifty-third pin of chip U1.

4. The traffic light recognition system for assisting red-green color blind drivers according to claim 2, characterized in that: The WiFi module includes a chip U3, the model of which is ESP8266. The first pin of the chip U3 is connected to the voice module and the 5V power supply, the second pin of the chip U3 is grounded, the third pin of the chip U3 is connected to the tenth pin of the chip U1, and the fifth pin of the chip U3 is connected to the seventh pin of the chip U1.

5. The traffic light recognition system for assisting red-green color blind drivers according to claim 4, characterized in that: The voice module includes a chip U4, the model of which is SYN6288. The twelfth pin of the chip U4 is connected to a 5V power supply, the fourteenth pin of the chip U4 is grounded, the twenty-third pin of the chip U4 is connected to the fifteenth pin of the chip U1, and the twenty-fourth pin of the chip U4 is connected to the sixteenth pin of the chip U1.

6. The traffic light recognition system for assisting red-green color blind drivers according to claim 4, characterized in that: The display module is an OLED screen, including chip U5. The model of chip U5 is 0.96OLED. The first pin of chip U5 is grounded, the second pin of chip U5 is connected to +5V voltage, the third pin of chip U5 is connected to the 30th pin of chip U1, and the fourth pin of chip U5 is connected to the 29th pin of chip U1.

Images