水田旋耕路径的识别方法

Method of identifying rotary tillage path in paddy cultivation

为减少耕作过程中产生的重耕、漏耕现象,提高耕作效率,针对导航视觉系统采集的水田非结构化自然环境彩色图像设计一种旋耕机耕整路径识别方法,探讨适宜于水田的彩色特征模型的选择。结合水田土壤区域噪声小而秸秆区域噪声大的特点,利用过量补偿法对水田图像进行去噪,再对图像进行逐行扫描,得到每个像素点的彩色特征值。利用土壤区域和秸秆区域的彩色特征值差异得到边界点信息,最后利用最小二乘法将得到的边界点拟合成一条直线,求得直线的倾角即为旋耕机的转向角,从而确定旋耕机的前进方向即确定路径信息。

To reduce the cultivation process of heavy plow, leakage phenomenon of tillage,improve work efficiency, a method of identifying path for rotary cultivator based on the paddy field unstructured environment color image collected by visual navigation system was designed. Choosing the suitable color feature of paddy field was discussed. Considering that the soil area has less noise and the straw area has more noise, the over and compensation method was used to denoise the paddy image. The feature value of each pixel color was acquired by progressively scanning the image. The border information was obtained by using difference of the color feature between the straw area and soil area. The least squares technique was used to make a line and acquire the elevation angle which is the steering angle of rotavator. The rota- vatofs direction of advance which is the path information was determined. This method achieved good re- sults in static image processing. It will be improved by the field test.