Fast geometry figure recognition algorithm based on edge pixel point eigenvalues
基于边缘点特征值的快速几何图形识别算法

In view of some shortcomings about frequently-used currently shape recognition algorithms such as large amount of calculation, long processing time, single figure recognition or demanding to pre-set templates,this paper presented a fast geo-metry figure recognition algorithm based on edge pixel point eigenvalues.The polygon apexes and its rank orders were quickly recognized firstly based on the different variation laws of the eigenvalues of polygon apexes and other edge pixel point and the exact shape recognition of the polygon was finished as well.Then the figure center and radius, the length of major and minor axle could be worked out by eigendistance and the equation of a circle or ellipse was constructed to make a fast recognition for a circle or ellipse be done. The simulation result shows the algorithm merits such as recognizing rich kinds of figure, lower computational complexity, higher processing speed, no pre-setting template, realizing easyly, obtaining more information.