Research On Parameter Extract And Trace Controlling Of Agriculture AGV Vision Navigation

With the development of electronic technology and automation, various of agricultural robots have emerged. Meanwhile, many navigation techniques have also come forth after the appearance of robots, and studies are focused on vision navigation. Here we present a vision navigation of agricultural AGV based on KH-100 AGV vehicle. This paper contains mainly as follows:1. The calibration of CCD. By analysing the existing methods of CCD calibration, as well as the requirement of calibration to vision navigation, we propose a twe-stage method to calibration the CCD of vision system. In the process of calibration, we use circle target, extract the feature points in the image, and get CCD datas in-and-out by experiment and computer. Tested by the already calibrationed parameters, the precision of calibration is finally acquired.2. Image Pre-processing and extracting of boundaries. In the processing, using the normalized excess-green feature and the fixed threshold to segment images of cucumbers grown in rows in greenhouse and grapes grown in rows in field; Comparing with several filters, we adopt the media-filter to remove the noise of images is better; Filling the holes in the segmented images inord to remove the effect of holes; At last, mathematical morphlogy is used to extract the boundary of plants.3. Extracting the navigation parameters. Discuss the pros and cons of the existing methods of extracting navigation parameters, and present Radon transformation to extract navigation parameters to meet real-time control.In order to avoid the limitation of traditional Radon transformation and reduce processing time, we modify the existing one by combining with regions of interest. This method has been tested feasible in both simulations and real fields.4. Designing controller. Based on the structure of AGV vehicle, the kinematic model of agricultural AGV is constructed. But it is difficult to construct the accurate mathmatic model because of uncertain enviroment. Fuzzy logic control is proposed to control AGV. At last, the simulation of tracking a line trace has been researched. The simulation results shows fuzzy controller can quickly get expected position. The fuzzy controller satisfied the need of agriculture AGV.