苹果图象的低层处理及尺寸检测

Apple image processing and size detecting:

分析了苹果品质检测中的图象低层处理方法 ,并根据苹果的外形特征进行尺寸检测。利用白色背景下 ,图象的B分量灰度直方图进行图象与背景的分离 ,结合苹果表面光滑易反射的特点 ,用链符法搜索并细化轮廓 ,采用最小外接矩形法求横径、纵径 ;测量尺寸与实际尺寸的相关系数为 0 .955。

The general image processing methods in the apple quality detecting were analyzed.The image gray histogram of B component in the white background was used to segment the fruit from the background.The Chain Code Method was adopted to search the thinned boundary line.According to the shape character of apple,the error of detected apple size versus real apple size was in 2 mm,and the correlation coefficient was 0.955.It is of great significance in promoting the application of computer vision technique for fruit quality detecting and grading.