Development of strawberry harvest robot with 3 dimension space recognition function

• Project/Area Number: 17580228
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Agricultural information engineering
• Research Institution: Utsunomiya University
• Principal Investigator: KASHIWAZAKI Masaru Utsunomiya University, Agriculture, Associate Professor (00282385)
• Project Period (FY): 2005 – 2008
• Project Status: Completed (Fiscal Year 2009)
• Budget Amount: ¥4,040,000 (Direct Cost: ¥3,800,000、Indirect Cost: ¥240,000); Fiscal Year 2007: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000); Fiscal Year 2006: ¥1,000,000 (Direct Cost: ¥1,000,000); Fiscal Year 2005: ¥2,000,000 (Direct Cost: ¥2,000,000)
• Keywords: Strawberry Fruit / Photometric Stereo Method / 3 dimensional shape / 3 dimension space recognition / Machine Harvest / ルートプランニング / イチゴ / 3次元形状認識

Research Abstract

This study aims at the development of the strawberry harvest robot with the object recognition function in 3 dimension space. The photometric stereo method (PS method) was assumed to be a basic theory in the study period, the image based 3D modeling system (IB3D system) was constructed theoretically, it made far trial purposes, and the object recognition system was developed intended for the strawberry. First of all, the IB3DM system was designed by the PS method. This system is composed of one high sensitivity CCD camera and some omnidirectional point lights and computer. RGB filter was built into the lens mount in the camera point and the acquisition of the color image was enabled. As for the IB3DM system that developed, it was small, and it was possible to acquire highly accurate image information under slight light such as LED. In usual color CCD camera, the distinction of the strawberry fruit's coming in succession is difficult. The strawberry fruit that exists coming in succession can be distinguished in this system. This algorithm enables the distinction of an individual strawberry based on the shape information in three dimension space. The end effector that cut the stem of the strawberry fruit, held the stem, and harvested was made, and it integrated it with IB3DM.
A basic function was confirmed by installing the IB3DM system in the strawberry harvest robot as an external world recognition sensing system of the manipulator mount type, and using the model of the strawberry fruit and the stem and leaf. As a result, the grasp of three dimension shape and the position of the strawberry fruit and the stem and leaf became possible. However, because the presumption accuracy of three dimension shape is different in the image of the single color and the two or more color image, it is necessary to improve the PS method algorithm.