Application of BP Neural Network and Genetic Algorithm in Lock Detection of Freight Train

To realize automatic detection and recording system of freight train, it's necessary to detect from the train image whether the locks exist. First the projection feature, moment vector feature of edge image, and the feature of gray histogram are extracted, and then BP neural network is used for detecting and locating the lock. At the same time, genetic algorithm is used to speed up searching the whole image with parallel and robust processing. Experiment results show that the method can overcome the problems of variety in lock types, image deformation and variation of environmental brightness, and enhances the speed and success rate of detection.