| As the automatic driving constantly upgrades,more sensors like high-definition camera and lidar will be mounted in a vehicle.Acquiring accurate and reliable detection signal in driving environment is the precondition for implementing intelligent driving function,and maintaining a "clear field of view" is essential for camera to capture the driving environment and provide data.In the case of muddy roads and rainy days,the vehicle-mounted camera’s lens is inevitably covered with the dirts like stain,dust,water-drop and mud.If the dirts are not found and cleaned in time,they will seriously affect the quality of captured images and the recognition capability of sensors,and then cause the in-vehicle computer to identify error message,interfering with driver’s judgment and leading to an accident.To ensure the stable and secure operation of intelligent driving function in any conditions,it is necessary for intelligent driving vehicle to keep camera’s lens clean.Therefore,the automatic cleaning system is essential for camera and lidar in vehicle.On the basis of previous studies,this study designs a vehicle-mounted camera automatic cleaning testing system including stains treatment system,cleaning system and image acquisition system,builds a mock-up of this system and defines relevant test parameters.Then,this study improves the automatic cleaning testing system,establishes the conditions for basic test and determines the operation steps of test.With the help of catia,matlab and python,this study analyzes the experimental results by applying machine learning algorithm and statistical method of pixels in grayscale digital image.In addition,this study explores the relationship between the degree of dirt on camera’s lens and the grayscale value of captured image,develops the standards for degree of dirt,and studies the effect of such parameters as spraying angle,cleaning time,spraying pressure and spraying distance on the cleaning efficiency of automatic cleaning testing system.The obtained results indicate that the degree of dirt on camera’s lens is associated with the grayscale value of captured image,specifically speaking,dirtier lens leads to lower grayscale values of captured image,while cleaner lens induces higher grayscale values of captured image;with other conditions unchanged,the cleaning efficiency linearly increases with the increasing spraying angle;with the increasing spraying distance of cleaning fluid,the cleaning efficiency rapidly increases and gradually stabilizes,but decreases when the spraying cleaning fluid is beyond a certain distance;higher spraying pressure of cleaning fluid causes higher cleaning efficiency;but the cleaning efficiency is influenced more by the change of spraying distance than by the change of spraying pressure.The cleaning efficiency of the automatic cleaning testing system reaches the highest(87.75%)at the conditions as follows:0.24MPa spraying pressure,8 cm spraying distance and 90° spraying angle.This study is expected to provide reference for the practical engineering application and optimization of the camera automatic cleaning testing system in intelligent driving vehicle. |
