Research On The Skeleton Extraction Of Corn Crop Rows In Precision Spraying System

In the system of precision spraying, the accuracy of agricultural machinery navigation route determined the effectiveness of spraying. It is the key to extract crop skeleton lines for navigation line recognition of precision spraying system, and it is also the basis of accomplishing precision toward target spraying technology. This work was financially supported by the open fund project of NERCITA(KFZN2012W12-012), key science and technology project of Science and Technology Department of Henan Province(132102110150), and Science and Technology Innovation Fund Project of Zhengzhou University of Light Industry graduate student(2014003).The growth of mid corn crops were regarded as the research object in this paper. Visual C++ 6.0 was the software system that could handle the corn crops. The objective of the paper was to seek a kind of practical skeleton algorithm that had strong robustness and stability. The extraction of crop row skeleton based on image processing was an important research direction of machine vision navigation in agriculture. In order to accurately extract the crop skeleton line to do precise target spraying, the main content of this paper was summarized as follows:1. Based on the extensive literature research, the present status and progress of skeleton extraction algorithm at home and abroad were summed up. Four kinds of classic skeleton extraction algorithms were introduced and the advantages and disadvantages of every skeleton extraction algorithm were analyzed.2. In the process of crop image preprocess, a kind of super green gray algorithm improved(1.75G-R-B) was proposed according to the color features of corn crop image in the natural environment. Compared to the traditional algorithm, the algorithm proposed in the paper greatly reducing the interference of background noise, which reduced unnecessary trouble for subsequent filtering operation and morphological operations.3. On the basis of analyzing and summarizing the previous research results, three kinds of algorithms for skeleton extraction of crop were proposed. At the same time, in order to accurately extract the central line of the crop skeleton in precision spraying, this paper presents a kind of minimal tangent circle algorithm that could extract the skeletons of the central crop lines. The main idea was find out the point of tangency of all minimum tangent circles from the binary image that had been thinned.4. Because of the recognition of machine vision was the linear structure. There were a lot of methods about straight line fitting of central crop row. In this paper, the randomized Hough transform algorithm was used to fit a straight line because of its advantage, and set up the experimental platform. Finally, compared with other skeleton algorithms to detect the linear accuracy and verify the algorithm proposed in this paper could be applied in the actual precision spraying system.