基于图像融合技术的大豆图像滤波方法研究  

Soybean Image Filtering Method Based on Image Fusion Technology Research

作　　者：柴玉华[1] 曹晓达[1] 谭克竹[1] 丁然[1] 

机构地区：[1]东北农业大学电气与信息学院,哈尔滨150030

出　　处：《农机化研究》2015年第8期62-65,69,共5页Journal of Agricultural Mechanization Research

基　　金：黑龙江省自然科学基金重点项目(ZD201303)

摘　　要：基于机器视觉的大豆外观品质检测一直是近年来研究的热点,其中大豆图像的滤波是大豆外观品质检测的重要工作内容之一。为了更好地去除大豆图像的噪声,提出了一种基于图像融合技术的大豆图像滤波方法。该算法对一个样本图像分别进行维纳滤波和形态学滤波,在此基础上进行基于小波变换的图像融合算法,有效解决了图像边缘毛刺现象。实验证明,此方法保证了图像的细节和边缘的完整性,图像滤波效果良好。Based on machine vision of appearance quality detection of soybean is a hotspot research in recent years, the filtering of soybean image is one of the important work of appearance quality detection of soybean .In order to remove the noise of soybean image better, this research puts forward a kind of soybean image filtering method based on image fusion. The algorithm put Wiener filter and morphological filter on a sample image respectively, then go on image fusion algo-rithm based on wavelet transform, effectively solve the image edge burr phenomenon, the experiment proves that this method guarantees the integrity of details and the edge of the image , and good effect of filtering .

关 键 词：大豆 小波变换 图像融合 形态学滤波 维纳滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]