Font Size: a A A

Research On An Automatic Navigation System For Field Weeding Robot

Posted on:2012-06-27Degree:MasterType:Thesis
Country:ChinaCandidate:L L JiangFull Text:PDF
GTID:2218330344950678Subject:Detection Technology and Automation
Abstract/Summary:PDF Full Text Request
With the purpose of reducing herbicide usage, protecting the environment as well as decreasing the labor intensity. This paper carried out the research of the key technology of weeding robot on the basis of reviewing literatures of the weeding robot navigation technology. This paper studied the automatic navigation system based on the combination of machine vision and GPS technology.Firstly, This paper analysised the color characteristics of maize from different point of view, used Extra-green method to grayscale image, and then thresholded to obtain binary images using the method of Otsu. After processing the image, a median filter and morphological processing method was applied. According to the characteristics of the line of corn seedlings, image was segmented based on the height of pixels, then area transaction, navigation points identification were processed. Locating points were classified using the target points classification algorithm. This paper fitting a straight line by using Hough transform, and then fitting once more with the method of least squares. The method was proved reliable through the maize seedling image analysis under different situations,completed the process of identifying the navigation line machine vision-based.Second,requirements for weeding robot navigation,simple calibration template was produced.Tsai two-step method used to calibrate the internal and external camera parameters obtained.fuzzy control rules were designed using MATLAB fuzzy toolbox on the platform of robot Voyager II. In addition, a vision-based navigation control software was developed by VC++. Large number of weeding robot navigation test have confirmed that the fuzzy control algorithm has achieved the goals with great efficiency,basically achieved weeding robot navigation machine vision-based.Finally, weeding robot navigation outside the fields was builded a GPS/digital compass nonlinear joint Kalman filter model, and then a integrated navigation software was designed to achieve the theoretical study on navigation technology.The work provided theoretical and technical basis for further research of the subject.
Keywords/Search Tags:Machine Vision, Weeding robot navigation, GPS, Digital compass, Fuzzy control
PDF Full Text Request
Related items