Image recognition method of partially occluded citrus fruit

Abstract

The invention discloses a partially covered citrus fruit image recognition method, comprising the steps of: collecting the image and cutting it; converting the cut image into a black and white binary image and removing black holes and blackening processing of white false targets; The obtained fruit outline width is single-pixelized, and the fruit outline coordinates in the marked area are read to obtain the citrus fruit outline coordinate matrix; a set of citrus fruit outline coordinates arranged in a clockwise direction is formed; find out that the abscissa and ordinate are different from the previous point Contour corner points, generate citrus fruit contour corner point sets; find intersection points, inflection points, endpoints, breakpoints; rough filter check points to form a set of feature points to be selected and fine filter check points to form a set of feature points; curve fitting, discard Fake fruit targets and remove overfitting curves; calculate curve parameters. The invention is not sensitive to the angle of image acquisition, ensures the accuracy of the contour extraction of the covered fruit area, and meets the requirement of simultaneously detecting round fruits and oval fruits when the fruit contour is restored.

Description

technical field [0001] In particular, the present invention relates to a partially occluded citrus fruit image recognition method. Background technique [0002] The real-time online identification of fruits on trees is one of the core technologies of fruit harvesting robots, and it is also a basic problem of fruit information management. The rapid identification of occluded fruits is the first difficult problem to be overcome among these problems. [0003] The machine recognition of partially occluded fruit on the tree mainly includes three steps: extraction of the contour segment of the unoccluded fruit area, restoration of the fruit contour, and calculation of fruit position and size parameters. The impact is relatively large, and there is a limitation of a single property curve when the fruit outline is restored. Contents of the invention [0004] The object of the present invention is to, aiming at the deficiencies of the above-mentioned prior art, provide a kind of p...

AI Technical Summary

Problems solved by technology

[0003] The machine recognition of partially occluded fruit on the tree mainly includes three steps: extraction of the contour segment of the unoccluded fruit area, restoration of the fruit contour, and calculation of fruit position and size parameters. The influence is large, and at the same time, there is a limitation of a single property curve when the fruit outline is restored

Images