自然环境下水果图像分割与定位研究

研究了自然环境下苹果、梨、桃子、杏子、李子、石榴、枣、樱桃等水果图像的果实和背景颜色特征，提出了用2r-g-b分量进行图像分割的方法。为了使Hough变换能够检测出水果果实，该文对随机圆Hough变换算法进行了改进，增加了检测圆的梯度约束和半径约束。试验表明对多数水果的图像利用2r-g-b进行分割，并用二值形态学方法对图像进行滤波后，用随机圆Hough变换可以有效地检测出了水果果实。

Study on the Segmentation and Orientation of Fruit Image Under Natural Environment

A new method of image segmentation with 2r-g-b component is proposed in this paper. It is based on the study of color characteristic of fruit and background in a fruit image such as apples, pears, peaches, apricots, plums, megranates, Chinese dates and cherry etc. In order to examine the fruit using Hough transform, the algorithm of Random Circle Hough Transform has been improved in this paper by adding grads restriction and radius restriction of examined circle. A lot of statistical tests show that after 2r-g-b segmentation, binary morphology filtration and random Hough transformation, it is effective to examine the fruit from a fruit image for most fruits.