摘要
针对基于计算机视觉技术对无土栽培番茄缺乏营养元素智能识别研究中 ,如何提取缺素叶片纹理特征问题 ,提出了差分百分率直方图法。特征有效性不受叶片大小、形状差异和叶片图像中叶片周边白色背景的影响 ,实验验证该方法能较好地提取出缺素叶片纹理特征。最后利用 K近邻模式识别法进行模式识别 ,识别的准确率在 80
In this paper, the percent histogram of differentiation was put forward in order to extracting texture feature of tomato's leaves in soiless agriculture for diagnosing diseases intellectively that were short of a kind of nutrient. The better suitable differentiating operator was studied by analyzing and tests. The method can get rid of the disturbance of the white background noises in the leaf image by dealing with the histogram, and remove the interferes of leaf's size by changing the histogram into percent histogram. At the same time the influence of shape was put out in the histogram. The way of looking for the better differentiation position in percent histogram for extracting effective texture feature was researched. The experimental results of extracting texture feature of the tomato's leaves proved that the method was effective in separating the tomato's different kinds of leaves, and the result of experiment showed that the accuracy is above 80%.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2003年第2期76-79,共4页
Transactions of the Chinese Society for Agricultural Machinery
基金
江苏省自然科学基金项目 (项目编号 :BK2 0 0 10 89
关键词
信息处理技术
番茄
叶
差分
提取纹理特征
Information processing technology, Tomatoes, Leaves, Difference, Extracting texture feature
