Vision Navigation Of Automated Guided Vehicle

Among the navigation methods of AGV (Automated Guided Vehicle) ,vision navigation is developing rapidly because it has such merits as plenty of information and flexibility, where image processing is important. Therefore, improving exactness and robustness of landmark recognition, can not only decrease tracking error, but also increase adaptability and automation of vision-navigated AGV.Aimed at image processing techniques of guidance images, related research is implemented. Based on template convolution, resolve the noise disturbance of image. Image segmentation based on differential operator implements edge detection of road mark. Landmark recognition is the key technique in image processing. Based on genetic algorithm, the algorithm of landmark recognition is suggested, the location of AGV can be obtained through coordinate system transformation.According to the above methods, the image collection card is developed using Visual C++ language, making it implement image collection, landmark recognition and serial communication. The experiment results show that the vision navigation system can efficiently perform the tracking of land mark, and that the recognition algorithm can satisfy the requirements of the system.