Font Size: a A A

Object Intelligent Segmenting And Recognizing In Tomato Automatic Harvesting Robot

Posted on:2006-03-22Degree:MasterType:Thesis
Country:ChinaCandidate:J H TangFull Text:PDF
GTID:2168360155467223Subject:Agricultural mechanization project
Abstract/Summary:PDF Full Text Request
Autonomous robot for harvesting farm produces (such as fruits) has a good application foreground. In abroad, the research in this field has been done for many years and has been had some results. But in our country, the research development is slow. There is a long way to apply the robot to practice. In the study of my paper, my study object mostly is outdoor tomato, and the paper achieves an intelligent method to identify tomato in outdoor environment. The paper mainly has been accomplished work as followed:1. The outdoor environment is changeful, so a lot of pictures were grabbed in some kinds of outdoor environment that we could have a general understand about actual problems in identifying tomato.2. Calculating different tomato histogram, then analyzing the color feature of tomato in outdoor environment. Based on the analysis result of color, two methods that are respectively based on boundary and region have been designed and tested in RGB (HLS) images and some images, which are derived from RGB, and HLS images. The experimentations have proved that these methods are not good because the arithmetic's' flexibility is discontented.3. Putting forward an artificial neural network (ANN) method to segment tomato image. First, the paper analyses the feasibility of ANN, then the paper design the type and structure of ANN. Finally, the parameters of ANN have been optimized. The experimentations have proved that this method is better than two above methods, but it also need more study.4. Analyzing the fact problems in harvesting, such as several tomatoes overlapped and tomato part covered etc. Then using distance arithmetic and watershed arithmetic to separate several tomatoes. Putting forward two methods to calculate tomato's center of gravity when tomato is part covered.Finally, the paper summarizes the major research work, and discusses the future about this research.
Keywords/Search Tags:Computer vision (Machine vision), Tomato harvesting, Color image processing, Artificial neural network, Pattern recognition
PDF Full Text Request
Related items