关键词: 昆虫, 图像处理, 图像分割, 图像识别, 自动鉴定, 形态分类学

Abstract: Automatic identification on insect images is a rapid identification method, and image segmentation is a key process in insect image automatic identification. Based on searching and sorting the recent references on insect image segmentation, we found that the researches on insect image segmentation are increasing rapidly. With the development of image processing technology, all kinds of segmentation methods such as level set, edge-flow and JSEG (a segmentation method integrating shape, color and texture information of images) had been tested in insect image segmentation. Although these methods performed well on their own test sets, there is no sharing image dataset or rules to evaluate these methods. It is necessary to build a system to evaluate the increasing insect image segmentation methods. In this review, we suggest creating some standard datasets and doing researches on evaluation of segmentation methods which are desiderated in insect image segmentation.

Key words: Insect, image processing, image segmentation, image recognition, automated identification, morphological taxonomy