基于颜色特征的作物与背景分割的对比研究

化学除草剂和杀虫剂的运用对环境造成很大的冲击,有选择地对准目标精确喷药能够有效地减少化学药剂的使用量,对准精确喷药的关键是运用机器视觉技术正确地识别目标,而作物与背景的快速并准确地分割是正确识别目标的基础。文章介绍几种图像的颜色模型和转换公式,运用VC++软件设计出图像背景分割的程序。选择南京农业大学江浦农场的玉米田、花生田和校园内的小麦田作为研究对象,运用自行设计的程序对图像进行分割并计算出分割图像所需的时间。用处理耗时与处理效果两个指标来评定用于背景分割的几种颜色特征。研究结果表明,过绿的2G-R-B是最好的背景分割因子,这个分割因子具有耗时少、分割效果好且不受光照强度影响的特点,能够满足农业智能生产中实时性的要求。

Contrast research on segmentation of farm crop and background based on color characteristic

The usage of chemistry herbicide and insecticide brings impact on environment. The usage quantity of herbicide and insecticide can be reduced by selecting the target for precision-spraying. The key technology of precision-spraying is how to identify the target accurately. The foundation work for identifying the target is fast and accurate segmentation. A few color image models and the conversion formula were introduced in this paper. Program of image background segmentation was projected based on VC++ 6.0. The corn and peanut plot of Jiangpu Farm and wheat plot located in Nanjing Agricultural University was picked out for research object. The image was segmented and the segmented time was computed using the program. The color characteristics model of background segmentation was assessed by segmented quality and processing time. The results showed that the excess green 2G-R-B was the best index for segmentation. The index had characteristic of consuming little time, good quality of segmentation and not being influenced by sunshine intensity. It can satisfy the request of agricultural intelligence production in-time.