A Simple And Convenient Arithmetic Of Object Recognizing On Tomato Automatic Harvesting

The machine vision technology obtains the object images by the image sensor, which is instead of the man's eyes. It transforms the images into the data matrixes and analyzes the images by computer, which can instead of the man's brain. At the same time it accomplishes the tasks relating with the vision. On 1970's, the technology arises from the two fruitful applying technologies, remote sensing and biomedicine picture s analyzing. Thereafter, the developed countries have developed the extensive researching using it and applied in various industry fields.Researchers have proved the feasibility of using the computer vision to direct the robot's system of the harvesting machinery. Kartell (1987) included that because the labor force cost increasing, the development of more prompt and effective microcomputer system and new sensors' system decides the economy feasibility of the robots harvesting system.Currently, robots activate in various fields and replace people to do some jobs, which are difficult to do by hand. A manipulator with vision sensors, which recognize the objects by the images' outlines, is a kind of simplest robots. The main researching orientations of the application of manipulators in the agricultural engineering are classification of agricultural products on line, the robots in the gardens for fresh fruits and vegetables' harvesting, packaging, forest pruning, farm and greenhouse and so on.The paper mainly researches the basic work of computer vision controlling the manipulators on the system of harvesting tomatoes that is how to recognize the tomatoes from the complex background and move the video lens towards to the centers of tomatoes, then measure the distance between the lens and tomatoes by laser ,and finally control the manipulators to harvest the tomatoes. The main problem is how to recognize the tomatoes and the leaves.The objective of this work includes:(1) Summarize the national and international research on the computer vision technology for the harvesting of farm products and process automatization.(2) Process the tomato and leave's pictures with the traditional image processing method and look for the rough centroid's position and show the disadvantages of this method using on the harvesting of the farm products.(3) Bring forward to a new idea for the locations of the tomato centroids by the image RGB values.(4) Compare the traditional and the color values to show the new way's superiority on the harvesting of the farm products.The paper takes the new method to pick up the tomato RGB color characters by the color information of the images. Analyzing the pixels of the images, the tomatoes identifying result is R-G>80. Compute the centroid's coordinates of the tomatoes and orientate them on the images using the Square method. It will satisfy quick and exact fruits harvesting.