A mango picking point recognition method

Abstract

The invention discloses a mango picking point recognition method which comprises the following steps: collecting mango images, and establishing a mango picking image library in a natural scene; establishing a mango fruit segmentation model based on a Mask R-CNN network ; calculating the long axis, the short axis and the mass center of each fruit; judging whether a cluster is formed or not by usinga bottom-up hierarchical clustering method; If the mango fruits are clustered, cluster fruit mother branches are identified, and picking points are positioned on the mother branches; and if the mangois a single fruit, segmenting and identifying fruit stems of the fruit, and determining picking points on the fruit stems. Mask-based R-is utilized in the invention. The mango fruit segmentation model of the CNN network carries out fruit instance segmentation, solves the detection segmentation problem caused by light change, shielding and overlapping in a natural orchard scene, and has the advantages of accurate segmentation and multiple application scenes.

Description

technical field [0001] The invention belongs to the technical field of fruit picking point recognition, in particular to a method for mango fruit segmentation, stem recognition and picking point location based on Mask R-CNN. Background technique [0002] my country's mango is one of the original production areas, and it is a large mango-producing country. Mango occupies an important position in the development of my country's fruit industry. At present, mango picking still relies on a lot of manpower. With the continuous expansion of mango production areas in my country and the increasing shortage of agricultural labor, it is urgent to improve the level of mechanization, automation and intelligence of mango picking. The effective identification of mango fruit picking points is the premise of intelligent mango picking. [0003] In terms of fruit segmentation and picking point recognition, traditional computer vision and the currently popular deep convolutional neural networ...

AI Technical Summary

Problems solved by technology

However, there are very few researches on fruit instance segmentation based on deep convolutional neural network, and fruit instance segmentation is an important step in picking point recognition

Images