The issue designs a high-definition digital processing platform, and presents algorithms of image sensor decoding and image enhancement pre-processing and its implementation. Firstly, the framework of the high-definition digital processing platform is introduced. Then the image sensor decode algorithm is given in detail. Finally the paper introduces the framework of the pre-processing algorithm .The major function block of the pre-processing algorithm is discussed in detail, including image filtering, image sharpening, gamma correction, white balance. Experiment results prove that the proposed method improves the image quality captured by CMOS sensor effectively.Keywords-digital processing; decode algorithm; pre-processing; image filtering; image sharpening; white balance I.

Based on the study of the causes of nocturnal traffic accidents, it is found that the poor lighting of automobile headlights at night is one of the main reasons. In order to solve this problem, an intelligent adjustment system of headlights based on machine vision is designed. Through the binocular camera to collect images, its detection and identification processing, prejudge the road ahead is straight, up and down slope, or bend, and whether there are people with the car in front, so as to drive the lights to make corresponding adjustments. This system can realize the adjustment of the headlights in the horizontal and vertical direction, according to the road conditions to do real-time intelligent adjustment, but also to achieve the brightness of the lamp adjustment, to achieve antiglare effect. In order to achieve safe lighting, improve the safety of night driving.