Abstract: A method for positioning tea-bud picking points based on fused thermal images and RGB images is provided and includes: firstly, image pairs of several tea buds are acquired by using an image acquisition device, and are each labeled to obtain a tea-bud object detection database and a tea-bud keypoint detection database; secondly, in order to obtain a trained object detection model and a trained keypoint detection model, the tea-bud object detection database and the tea-bud keypoint detection database are inputted into an object detection model and a keypoint detection model for training, respectively; finally, the trained object detection model and keypoint detection model are used to sequentially process the tea-bud image pairs to obtain tea-bud keypoint positions, and then tea-bud picking point positions are obtained in combination with the tea-bud growth characteristics. The positioning accuracy and efficiency of the tea-bud picking points can be improved.

Abstract: A method for positioning tea-bud picking points based on fused thermal images and RGB images is provided and includes: firstly, image pairs of several tea buds are acquired by using an image acquisition device, and are each labeled to obtain a tea-bud object detection database and a tea-bud keypoint detection database; secondly, in order to obtain a trained object detection model and a trained keypoint detection model, the tea-bud object detection database and the tea-bud keypoint detection database are inputted into an object detection model and a keypoint detection model for training, respectively; finally, the trained object detection model and keypoint detection model are used to sequentially process the tea-bud image pairs to obtain tea-bud keypoint positions, and then tea-bud picking point positions are obtained in combination with the tea-bud growth characteristics. The positioning accuracy and efficiency of the tea-bud picking points can be improved.