基于小波分解的纹理分割的农机化应用研究  

Resarch of Wavelet Based Texture Segmentation for Agricultural Machanization Application

作　　者：丁伟杰[1] 周国民[1] 金弟 

机构地区：[1]浙江警察学院计算机与信息技术系,杭州310053 [2]中国石油集团杭州地质研究所计算中心,杭州310023

出　　处：《农机化研究》2008年第11期176-179,共4页Journal of Agricultural Mechanization Research

基　　金：浙江省自然科学基金资助项目(ZD0205)

摘　　要：农业机械自动导航是农业自动化的重要标志之一,其中图像分割技术已成为农业机械导航的重要环节。为此,根据已收割和未收割的农作物图像具有不同纹理特征的性质,采用小波分解方法,对农作物图像的自动分割进行了大量实验研究,并且与纹理分割常用的Laws能量特征法进行了实验效果和运行效率对比。理论分析和实验结果均表明:基于小波分解的纹理分割方法,完全适用于农业机械导航中的农作物图像自动分割。Agricultural automatic guidance is one of important symbols of automation of agriculture. Image segmentation has become crucial in Agricultural machinery guidance. According to the difference between cut and uncut crops textural, this article explored an wavelet transform way in automatic segmentation of crop images , and carried through a great deal of experiments. Effect and efficiency also had been compared between the wavelet transform way and the Laws texture energy statistical way. Both theory analysis and experimental results show that the wavelet based segmentation is effective and practicable.

关 键 词：机械导航 小波分解 纹理分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]