Night green apple recognition method of apple picking robot

Abstract

The invention discloses a night green apple recognition method of an apple picking robot. The method comprises the following steps: 1) supplementing light to green apples by using an annular LED lamp,and collecting images by using a color CCD camera; 2) performing Gaussian filtering on the image, performing contrast-limited adaptive histogram equalization on the image, and then using gamma correction to improve the contrast; 3) performing super-pixel segmentation on the image processed in the step 2) by using an SLIC algorithm; 4) extracting color and texture features of the superpixels; 5) classifying the superpixels by using a pre-trained SVM classifier to obtain an apple template image, and removing a connected region with the area smaller than 300 pixels; and 6) identifying and positioning the apple by using a Hough circle transformation method. According to the method, the problems that the whole night image is dark, the contrast ratio is not high, and the edge is fuzzy are solved, and the problem that the green apple and the leaf are difficult to segment is solved, so that the apple picking robot can normally pick the green apple at night.

Description

technical field [0001] The present invention relates to the technical field of agricultural machinery, in particular to an image recognition method for a fruit and vegetable picking robot, and more specifically, to a green apple recognition method for an apple picking robot at night. Background technique [0002] Apple is a relatively common fruit in daily life. It is rich in various nutrients such as sucrose, malic acid, B vitamins, vitamin C and trace elements. It has a crisp taste and sweet and sour taste, and is deeply loved by people. [0003] At present, most of the research objects of apple recognition are red apples. For green apples, it is difficult to distinguish them from leaves by color, which brings great difficulties to the identification of green apples. At the same time, the fruit recognition algorithm research is mainly carried out under natural light conditions during the day, and the night time accounts for one-third of the whole day. Realizing the normal...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the current problem that green apples cannot be picked normally at night, and proposes a method for identifying green apples at night by an apple picking robot, which solves the problems of overall dark images at night, low contrast, and blurred edges, as well as the problem of green apples and green apples. The problem of indivisible leaves allows apple picking robots to pick green apples normally at night

Images