The Improved Genetic Algorithm Used to Optimize Gabor Filters for Vehicle Detection

摘要

In this paper we focus on using the improved Genetic Algorithm (IGA) to optimize Gabor filters for vehicledetection. The Gabor filters parameters are encoded in a chromosome and optimized using genetic operators. Considering the deficiency of the traditional crossover operator at global searching, the partheno-crossover operator, which is a new crossover operator, is proposed and used to improve GA. The objective is to build the specifically set of filters which are capable of responding stronger to vehicles than to nonvehicles.Compared with using general GA, the Gabor filter optimized by using the IGA allows better correct rate of vehicle detection.