摘要
设计了一种基于平面图像识别的袋装粮数量智能识别方法。对粮库现场采集的图像样本,运用数字图像处理技术对图像进行降噪和对象体特征突显;然后运用区域增长法对图像中每一闭合像素区域进行聚类分割;最后根据相应几何特征值判定粮袋身份并通过其椭圆度的范围,以实现粮袋所属面的判定,达到对粮堆重量的自动识别。
This paper designed a bag grain quantity recognition method, which was based on plane image recognition. Used digital image processing technique for collected image sample from the scene of grain warehouse; in order to reduce the noise and make the object' s character highlighter. Then made use of the growth method to the each closed pixel district of the pic- ture, so as to carry on the pixel clustering and image segmentation. At last according to the value of corresponding geometric moment of this character to judge the identity of each grain bag, and made the classification of the surface each grain bag be- longed to come true by the value of its eccentricity, so as to achieve the automatic identify of grain' s weight.
出处
《计算机应用研究》
CSCD
北大核心
2009年第4期1572-1574,1590,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(60603027)
重庆市财政局重点科技资助项目(810080)
重庆市重点软科学资助项目(CSTC,2007CE9006)
关键词
图像识别
噪声消除
区域增长
几何矩
数量识别
image recognition
noise eliminate
region growing
geometric moment
quantity recognition