Simulation and experiment of machine vision guidance of agriculture vehicles

This paper aimed at the key problems of machine vision guidance on agriculture vehicles. The two problems of image processing and path control were studied in detail. Guidance features in natural environment were extracted and measured based on the Hough transformation algorithm and dynamic window technology. A simulation model was created according to the system identification result and the characteristics of the machine vision guidance on agriculture vehicles. At last a test system was realized on a general purpose wheeled tractor to test all the image processing and control algorithm and the test results were satisfactory.