【作者】 唐晶磊； 何东健； 朱兆龙；

【Author】 Jinglei Tang1,Dongjian He2,Zhaolong Zhu2(1.College of Information Engineering 2.College of Mechanical and Electronic Engineering,Northwest A&F University,Yangling,Shaanxi,712100,China)

【机构】 西北农林科技大学信息工程学院； 机械与电子工程学院；

【摘要】 以农田场景图像为基础,依据灰度图像的分割效果选取G-R因子实现图像的灰度化,并对比分析了均值和中值滤波的适用条件;为避免光照不均导致阈值偏差对分割效果的影响,对直方图滤波后,采用直方图阈值化方法实现了绿色植物与土壤背景的分割;对比分析了3种不同结构元素和形态学运算操作,基于形态学滤波实现分割图像的后处理。因此,在实现绿色植物与土壤背景图像分割时,需依据图像和目标特点分析选取图像处理方法,以获取较好的分割效果。更多还原

【Abstract】 G-R factors were selected to graying according to segmentation result based on the farmland scene image.Moreover,mean filter and median filter applicable conditions were comparative analyzed.Histogram threshold method was used to segment plants and soil background after filtering the histogram,which to avoid effect on segmentation caused by threshold deviation from poor illumination.Three different structure element and morphological operations were comparative analyzed,and post-processing of image segmentation was achieved based on morphological filtering.Therefore,characteristics of images and goals should be considered to obtain a better segmentation results in image segmentation of plants and soil background.更多还原