The Research Of Features Extraction Technology Of Aircraft On Geometry Characteristic

Automatic Target Recognition (ATR) is one of the important parts of computer vision, and has been widely applied to many fields like military and civil affairs, automatization, such as attack and guard of the air's target, monitor and control in the air, the management of aerodrome have widely outlook of application. The technology of Feature Extraction and Analysis is an important step of ATR. The relation between both is supplement each other. The result of Feature Extraction will have effect on the accuracy of ATR directly, on the contrary, the outcome of ATR will reflect the availability of these features.In this paper, firstly we study the image preprocessing of airplane in its 2-D image, including image smoothing, image filter, image edge detection, and binarization. And then on the basis of image edge detection and binarization, we detect the image's contour and track the boundary of image. So we get the closed boundary curve of airplane image. To receive the contour polygon of airplane image we approximate the closed boundary curve. Then we extract the geometry features of image, including: size features, figure features, the ratio of long to broad and the triangle area ratio of the same hemline etc.Finally, these geometry features, which is the input of Support Vector Machine (SVM), is used to recognize the category of the airplane. So testify the validity of features.