Based On The Characteristics Of Insect Wing Recognition Technology Research

After 300 million years of evolution of species on Earth, insect has become the most widely distribution of animals. On present, there are more than one million insects. In the traditional, classification and identification of insects is a very hard work. The expert uses the eye to found the differences in different insects. But it demands technical skills of identification of insect species. It is unable to meet the current needs of society. With the development of computer technology, computer image processing and pattern recognition have been applied to the rapid identification and classification of insects. Through the image recognition technology, the rapid establishment of insect classification system is not only scientific significance, but also have extensive use of practical work. Currently, we have every year a large-scale insect diseases, how to quickly identify the types of insects, which is a prerequisite for effective pest control. The same problem is in the customs. Customs must check so much imported crops, which may contain some insects. Some insects may be harmful to our country. Customs is lack of a large number of entomology experts. It can not quickly identify the types of insects. Therefore, this paper will explore the use of insect wing veins for image recognition, to develop a smart and highly efficient insect identification system. Thesis research work are as follows:(1) The insect wing veins image pre-processing. This article describes the filtering and denoising, normalization, gray and so on pretreatment technology for insect wing vein images for the specific purpose of the application and to achieve results. Insect wing veins for the particularity of the image to compare a variety of edge detection operator, select the image Canny edge detection operator. (2) Veins ridge crossing point and the endpoint of the extraction. This article describes the cross-point and the endpoint of the structural characteristics. According to their structural characteristics, design a practical algorithm can effectively extract the cross-point and the endpoint.(3) Wing vein image registration. This article describes a variety of alignment principles, because the veins outside the outline of a clear, therefore propose to focus on the outer contour-based registration.(4) The wing vein images compared. The image into a feature points than the comparison of the characteristic point is the intersection of veins ridge point and the endpoint. Feature points based on wing veins than with efficient and accurate characteristics.(5) The system software part of the implementation. This paper briefly describes the system structure. System is divided into feature extraction and characteristics of the client than the server-side. It gave details of client-side implementation of the various functions.