Road vehicle shadow feature extraction method

Abstract

The invention discloses a road vehicle shadow feature extraction method. Firstly, data needing to be processed are reduced and road surface areas of an image are extracted through a geometric constraint method, gauss convolution is conducted on the road surface areas, gray level upper limit threshold values of shadows are obtained, the image is scanned in lines through the threshold values, collection of shadow lines is acquired, and location information of the shadow lines is recorded. The shadow lines are divided into standard shadow lines and ordinary shadow lines according to difference value between an upper neighborhood and a lower neighborhood, the standard shadow lines are stored in a result set, shadow features of vehicles are simulated according to location relation fusion between the standard shadow lines and the ordinary shadow lines, and areas which may include the vehicles are extracted. By means of the shadow lines extracting and fusing to simulate existing feature extraction method based on the areas, the extraction effect of most of the vehicles is good. The consuming time is small, the rate of missing detection is low, the detection effect is good for distant vehicles particularly, and the detection rate of a system is improved.

Description

technical field [0001] The invention belongs to the technical field of feature extraction, in particular to a road vehicle shadow feature extraction method, in particular to a vehicle shadow feature extraction method based on shadow line fusion. Background technique [0002] With the rapid development of transportation, traffic safety is becoming more and more important. Detection of vehicles ahead in advanced driver assistance systems is a central problem of the system. Vehicle feature extraction technology (Hypothesis Generation) is a key part of the front vehicle detection technology. It can realize efficient vehicle detection by extracting the feature area that may contain vehicles in the image. The shadow feature of the vehicle is a continuous and large shadow area at the bottom of the vehicle in front. The average gray level of the shadow feature of the vehicle in the image is always lower than its neighbors and corresponds to the vehicle one by one. It is located at ...

AI Technical Summary

Problems solved by technology

This type of method has a high accuracy rate for the shadow detection of close-range vehicles. However, for the less obvious shadows of distant vehicles, the above-mentioned fixed threshold shadow area segmentation extraction and area screening algorithms are prone to miss detection, which seriously affects the vehicle detection performance.

Images