Nighttime double fruit overlap tomato identification method based on overlap edge detection under active illumination

Abstract

The invention discloses a nighttime double fruit overlap tomato identification method based on overlap edge detection under active illumination. The method comprises the following steps of collecting a nighttime double fruit overlap tomato image under active illumination; using an OTSU method to perform image segmentation based on R-G color difference; performing image enhancement by using a homomorphic filtering method for the grayscale of the double fruit overlap tomato region extracted after image segmentation; extracting the edge, the contour and the overlapping edge of a spot area of the double fruit overlap tomato region after the image enhancement; judging the anteroposterior position relationship between the front unshielded tomato and the later-blocked tomato in the double fruit overlap tomato using a distance nearest method; and circularly fitting the overlapping edge and the front unshielded tomato contour to achieve recognition of the front unshielded tomato in the double overlap tomatoes. The method can realize the recognition of the front unshielded tomato in the double fruit overlap tomatoes at night, and has good recognition effect on overlap tomatoes with different overlapping rates, especially for double fruit overlap tomatoes with the large overlap rate.

Description

technical field [0001] The invention relates to a method for recognizing overlapping tomatoes in image processing, in particular to a method for recognizing double-fruit overlapping tomatoes at night based on overlapping edge detection under active lighting. Background technique [0002] Fruit and vegetable picking robot is a hot spot in the field of agricultural robot research. The vision system is an important part of the fruit and vegetable picking robot, which can realize fruit and vegetable identification and 3D positioning. Due to the complex and variable growth of fruits and vegetables, the vision system often encounters different growth states of fruits and vegetables such as occlusion of branches and leaves, overlapping of fruits, etc., which brings great challenges to the picking operation of fruit and vegetable picking robots. Among them, overlapping fruit recognition is one of the difficulties in the current research on the vision system of fruit and vegetable p...

AI Technical Summary

Problems solved by technology

However, in the current research on the vision system of fruit and vegetable picking robots at night, the research on the recognition algorithm of fruit and vegetable is mainly aimed at the single fruit recognition algorithm at night, and the research on the recognition algorithm of overlapping fruits at night is less

In addition, most of the current overlapping fruit recognition algorithms cannot realize the front and rear positional relationship judgment between the front unoccluded fruit and the rear occluded fruit in the overlapping fruit, which leads to the fact that when the fruit and vegetable picking robot picks the occluded fruit first, it is easy to cause the front unoccluded fruit to be occluded. The end effector of the fruit and vegetable picking robot manipulator is damaged, and for overlapping fruits with a large overlap rate, recognition errors are prone to occur, and the recognition effect is poor. The overlap rate refers to the ratio of the covered part of the back fruit to the total area of ​​the back fruit

Images