摘要
对土壤背景进行有效分割是玉米苗期田间杂草识别的前提和基础。本研究利用颜色分量G-R和G-B与灰度直方图来实现玉米苗期图像与背景图像的分割,解决了玉米苗期田间杂草识别中受影响及环境适应性差等问题。通过对不同环境下苗期玉米图像的植被颜色和背景颜色分量的统计分析表明,采用G-R和G-B双阈值颜色特征分割,进行土壤背景分离取得了很好的效果。
Soil background segmentation is the premise and basics of weed identification at maize seedling stage.In this paper,using color components G-R,G-B and histogram,corn seedling image and background image segmentation were realized to solve the problem of affects and environment adaptability on weed identification at maize seeding stage.Statistics analysis of vegetation color and background color of maize seedling image under different environments,by using G-R and G-B double threshold segmentation of the color characteristics,it showed good effect.
出处
《农业网络信息》
2012年第8期22-24,共3页
Agriculture Network Information
关键词
玉米图像
颜色特征
双阈值
图像分割
maize image
color characteristics
dual threshold
image segmentation
