| abstract |
The invention relates to a potato image acquisition device based on an RGB-D camera and a bud eye identification and positioning method. The image acquisition device includes three RGB‑D cameras, which collect color images and depth images of potato samples from three different angles. In the image processing part, the color image is preprocessed and the mask-based target extraction method is used to obtain the potato target image, and then the classifier is trained using the Adaboost algorithm and Haar‑like features to identify the bud eye area and obtain the two-dimensional coordinates of the bud eye area ; Calibrate the RGB-D camera to obtain the internal parameters and external parameters of each RGB-D camera, and generate a point cloud in combination with the depth map and color map collected by the corresponding camera; The three-dimensional model of the potato sample is obtained by registration; then the two-dimensional coordinates of the obtained bud eye are converted into the coordinates in the three-dimensional space where the three-dimensional model of the potato sample is located, and the three-dimensional positioning of the potato bud eye is realized, which lays the foundation for realizing the automatic cutting of potato seed potatoes. |