苹果幼果图像的分割与识别算法研究

Segmentation and Recognition Algorithms for Image of Young Apple

以苹果的幼果为研究对象,在自然环境下采集其可见光图像,研究幼果图像分割与识别的算法。对幼果图像进行预处理,利用Otsu法对图像进行有效分割。在预处理的基础上,对苹果幼果图像进行识别。利用改进后的连通域标记算法对预处理后的苹果幼果图像进行标记并标号,进一步去除不符合要求的小面积,减少干扰。然后对图像进行轮廓提取,利用圆形度这一形状特征量识别出苹果幼果,并进行自动标记。实验结果证明了相关算法的正确性和有效性。

Apple＇s young fruit is taken as research object, and key segmentation technolo-gies and recognition of young fruit images are studied, the images are captured in visiblelight of natural environment and image processing methods of segmentation and recognitionof the apple＇s fruit are researched based on knowledge of image processing. Firstly, in orderto preprocess the apple images, the image is effectively segmented through the Otsu segmen-tation algorithm. Secondly,on the basis of pretreatment, the image of apple＇s fruit is recog-nized. The preprocessed apple images are marked and labeled with the improved connectedcomponents labeling algorithm which removes small areas that do not meet the requirementsto reduce the interference, and then the image contour is extracted. The apple fruit is identi-fied and automatically marked by the shape characteristics of circular degree. The algorithmsare proved to be effective and correct by experimental results.