农田视觉导航基准线的识别与提取方法研究

Study on the Method of Visual Navigation Baseline Identification and Extraction of Farmland

提出了一种基于农田图像处理的视觉导航基准线的提取方法。该方法首先使用加权平均法(2G-RB)对原图像进行灰度化处理;采用改进最大化类间方差分割灰度图像;再用小面积法去噪;识别去噪后的图像的中心点,确定导航中心点位置。基于中心点的分布规律,将图像中识别出的中心点进行筛选,得到有效的导航基准中心点,使用改进霍夫变换与最小二乘法相结合的方法提取导航基准线。本文方法能够准确地识别和提取导航基准线,实验证明了该方法的有效性和实时性。

In this paper a method is presented to extracts visual navigation baseline based on the image processing of farmland.Firstly, the original image was gray scale processed using the weighted average method ( 2 G-R-B ) in this method.And then improved maximize between-class variance was used for gray scale image segmentation.Then the bi-nary image of the division was denoised with small area method.The center points were identified on denoised image, to determine the location of the center navigation.Based on the distribution of the center point, image obtained effective nav-igation center points after filtering.The navigation baseline was generated by using Hough transform.The proposed method and vertical projection method were used to handle different images of the crops.The conclusion is mentioned method is effective and has the real-time property.