Research on Target Recognition Technology of Caragana Automatic Harvesting Robot

By analyzing the color characteristics of Caragana korshinskii images collected by the robot in the RGB space, a method of recognizing Caragana korshinskii image is proposed. This scheme consists of four stages. At first, the brightness of the image is converted. And then the uniformity of the brightness can be realized. The appropriate segmentation threshold is determined. At the same time an auxiliary image is generated to realize the image segmentation. For the unreasonable discontinuities in the image of Caragana korshinskii after segmentation due to illumination, this paper proposes a new branch compensation method to eliminate the unreasonable discontinuities in the Caragana korshinskii image. The realization of the target recognition of the Caragana korshinskii lays the foundation for the follow-up root positioning and harvesting of the Caragana korshinskii harvesting robot.