Design And Implementation Of Vehicle Intelligent Fog Lamp Control System Based On Machine Learning

In the weather of dense fog,turning on the vehicle fog lamp can act as a safety warning and help improve the safety of road traffic.However,the fog lamp is not used properly by the most driver,which increases the risk of traffic accidents.In order to improve the driving safety of the car,this article combines image recognition technology to design a set of vehicle intelligent fog lamp control system to realize the intelligent control of fog lamps.First,clarify a clear design scheme and goal of vehicle intelligent fog lamp control system,this is,the system utilizes machine learning image recognition technology to recognize the weather images collected by the camera,and automatically controls the fog lamp to turn on and off according to the recognition results.Secondly,based on the three algorithms of support vector machine,convolutional neural network and capsule network to design three kinds of foggy day recognition algorithm models,then built an image test set of four weather types including sunny,cloudy,rainy and dense fog to test recognition effect of three algorithms in different weather environments.The test results show that the recognition accuracy of the dense fog recognition algorithm based on the capsule network reaches 97.5%,which is higher than the support vector machine and the convolutional neural network.Therefore,the research finally chose the fog recognition algorithm based on the capsule network.Finally,hardware devices such as Jetson Nano development board,IMX219 camera,Raspberry PI display and LED indicator(used to simulate fog lamp)are selected to build the control system.And transplant the dense fog recognition algorithm based on the capsule network to Jetson Nano for debugging and testing.During the test,the driving video is played on the display screen to simulate the foreground of the car during the driving process,and the intelligent fog lamp control system complete image acquisition and image recognition.When the screen play content is foggy,the probability value of dense fog days output by the algorithm model fluctuates around the value of0.9,while the probability value of dense fog days fluctuates in the range of 0.3? 0.8 when the screen play content is non foggy.According to the result of the probability value fluctuating in different weather categories,the turn-on threshold and turn-off threshold of the fog lamp are set to 0.9 and 0.6,respectively.Under the condition of threshold value,the control system can automatically turn on / off the LED indicator in dense fog / non fog environment,which meets the functional design requirements.