Research on genetic algorithm based on tabu search for landmark image recognition

Abstract

To solve the problem of monocular-vision-oriented landmark recognition for mobile intelligent robot marching with complex background, a kind of landmark extraction algorithm which fused with transcendental color information and based on tabu-genetic search algorithm is proposed, and the convergence analysis of this algorithm is presented. Firstly, a novel experiment threshold based on RGB vision model is used for the detection step on the target image, after extracting image edge, the tabu-genetic search algorithm starts to work for recognizing landmark image primitive shape (primitive element), and so landmark contours are extracted for next position and recognition steps. Then, the position information as feedback is conveyed to the grayscale image, the whole landmark is extracted accurately with the second segmentation on the local image that is exclusive to landmark area. Finally, landmarks are recognized by Hopfield neural network. Results issued from experiments on a great number of images in deferent scenes show that a satisfied recognition effect has reached that landmarks can be recognized fast and efficiently.