Fruit picking method and device based on deviation distance value

The invention provides a fruit picking method and device based on a deviation distance value, and the method comprises the steps: carrying out the image collection processing of a to-be-picked fruit tree, so as to obtain a plurality of fruit tree images; generating a plurality of fruit positions; dividing the plurality of fruit positions into a first type of positions corresponding to mature fruits and a second type of positions corresponding to immature fruits; carrying out first picking treatment; recording all the second-class positions in a data pair mode; arranging a plurality of sample fruits to obtain a first sample fruit tree image; obtaining a second sample fruit tree image; generating different deviation distance values corresponding to different clustering categories, and generating an update data pair; the method has the beneficial effects that on the premise of ensuring the picking efficiency, the waste of computing resources is reduced, and the possibility of missing picking of fruits is reduced