Visual Acquisition Algorithm of Current Crop Location Information for Intra-Row Weeding Control in Paddy Fields

Crop location information is the key to intra-row weeding control in the paddy fields. The acquisition of current crop location information provides accurate basis for intra-row weeding and bypassing seedlings. In order to obtain accurate and reliable information of current crop location, a method to recognize and locate current crop was proposed in this paper. Firstly, the image captured is binarized by ExG index and Otsu algorithm. And then according to the rice intra-row distance, the edge of the region of interest (ROI) is determined. After that the position and centers of crop are marked by connected region segmentation. Finally, the current crops are determined by current crop area ratio between the current image frame and the previous image frame which is a complete crop image. If the area ratio is bigger than 1/2, the crop near the lower edge of the view is the current crop. Or the crop further is the current crop. Through the acquisition of rice crop images, the algorithm experiments of the collected images were carried out. The test results show that the recognition rate of current crop location is more than 85.79%, which can meet the needs of intra-row weeding control in paddy fields.