Fruit distinguishing and locating method with laser scanning and machine vision combined

Abstract

The invention discloses a fruit distinguishing and locating method with laser scanning and machine vision combined. According to the fruit distinguishing and locating method with laser scanning and machine vision combined, through the characteristics that laser scanning is not influenced by the light condition and can simultaneously obtain multiple pieces of information such as the distance, the reflectivity and the coordinate, fruits are distinguished from branches and leaves in a reflectivity graph generated after scanning through a clear outline formed by the obvious difference of reflectivities. According to the color characteristic difference between a color image and a characteristic point of the reflectivity graph, the mature fruits and the immature fruits are distinguished, wherein the mature fruits and the immature fruit are similar in reflectivity. According to the difference distances of multiple mutually-overlapped fruits in a depth map, the mature target fruit is distinguished from the fruits overlapping with the mature target fruit. According to the fruit distinguishing and locating method with laser scanning and machine vision combined, the multiple pieces of information, obtained after laser scanning, such as the reflectivity, the depth and the coordinate and the color image information of machine vision are combined, the purposes that the mature target fruit is distinguished from the branches and the leaves, the immature fruits, and the fruits overlapping with the mature target fruit and is located are achieved comprehensively. The fruit distinguishing and locating method with laser scanning and machine vision combined is simple and rapid, high in real-time performance and practicability, and applicable to distinction and locating of the fruits on a plant.

Description

Technical field [0001] The invention relates to the field of agricultural robots, in particular to a fruit identification and positioning method combining laser scanning and machine vision. Background technique [0002] In robot picking, the identification and positioning of fruits is a key technical problem. At present, machine vision is widely used to identify and locate fruits, but machine vision is sensitive to lighting conditions. Conditions such as backlight and backlight will cause difficulties and errors in extracting fruit edge information in the image, and the effects of supplementary light and other measures are affected by actual conditions. Strict restrictions on lighting conditions. At the same time, it is difficult to distinguish the ubiquitous overlapping of fruits in machine vision images, and only limited segmentation can be performed through complex algorithm processing, which is seriously insufficient in real-time and accuracy. A small number of multi-...

AI Technical Summary

Problems solved by technology

At present, machine vision is widely used to identify and locate fruits, but machine vision is sensitive to lighting conditions. Conditions such as backlight and backlight will cause difficulties and errors in extracting fruit edge information in the image, and the effects of supplementary light and other measures are affected by actual conditions. Strict restrictions on lighting conditions

At the same time, for the ubiquitous overlapping of fruits, it is difficult to distinguish in machine vision images, and only limited segmentation can be performed through complex algorithm processing, which is seriously insufficient in real-time and accuracy

A small number of multi-spectral technology is used for fruit identification, its accuracy and real-time performance are difficult to guarantee, and it cannot meet the needs of overlapping fruit identification and fruit positioning

In recent years, the development of 3D visual sensors with laser light source has been carried out at home and abroad (Liu Zhaoxiang et al., Design of 3D visual sensor for apple picking robot, Journal of Agricultural Machinery, 2010, 41, 2: 171~175; I Dewa Made SUBRATA et al., 3-D Vision Sensor for Cherry Tomato Harvesting Robot, JARQ, 1997: 257-264; Kanae Tanigakia et al., Cherry-harvesting robot, Computer and Electronics in Agriculture, 2008: 65~72.), using the difference in the reflection of fruit trees to laser light to realize fruit recognition, The principle of laser triangulation is used to locate the fruit, but it is difficult to make a correct judgment due to the lack of information when the reflectance of ripe fruit and immature fruit is similar.

Images