Matching location algorithm of corn image projection based on over-segmentation

Fast and accurate positioning of corn image is premise for variable spray between plants. In order to improve the real-time and stability of matching algorithm for complex scenes tracking in the field, firstly, over-segmenting method is applied to eliminate the interference of weeds and noise in this paper. Then, the datum template which is morphological stable root image is adopted to match and locate with the target block by projection method. This method greatly reduces the complexity of the conventional template-matching algorithm. Therefore, the adaptability and stability of location algorithm are greatly improved.