Research On Automobile Distance Measurement System On Highway Based On Vehicle-mounted Camera

Poor visibility brings the safety trouble to the automobile running onhighway in fog weather. The main work of this research is a part of " SceneryRecognition of highway in fog weather". The contents include research oncalibration method of vehicle-mounted camera, and how to acquire the distancebetween the two automobiles on highway.After the camera calibration, the relationship between the 2D imagecoordinates and the 3D space coordinates can be established. Using the cameraparameters and the geometry information of the object in real world, the distancebetween this vehicle and the preceding one can be obtained. The main works arefollowing:1,More research is concentrated on the image processing methods andobject recognition methods involved in vehicle-mounted camera calibration. Abackground expunction algorithm based on morphology is proposed. Theadvantages and disadvantages of several edge detection operations are comparedand one of the fittest operations is selected. Then the auto-detection of the cornerpoint is discussed. According to the above steps, the process of cameracalibration can be accomplished automatically.2,Aiming at the application of this subject and considering the highwayscenery, the line detection algorithm is implemented to detect the road marks.3,Using camera parameters obtained by camera calibration based oncheckerboard image and image information on highway, and the relationshipamong image coordinates , 3D coordinates and geometry information, thedistance between autos on highway can be obtained.This paper is concentrated on calibration method of vehicle-mounted camera.A process that can calibrate a vehicle-mounted camera automatically is proposed.This method can eliminate the factitious factors completely, so as to estimate thespace between this automobile and the preceding one on the highway. Thisresearch will be a basis for the further investigations on calibration ofvehicle-mounted camera.