An approach of near infrared image accurate orientation and recognition

This paper proposes an edge extraction algorithm of near infrared image based on 2-spline wavelet transform. Through Hough transform of edge image, object image's circle structure can be detected, and object image also has an accurate orientation. So we can extract special figure features of object image. Then near infrared images are classified by statistical template matching. All algorithms which are simple and efficient are compiled in VC++ 6.0. The experimental results show that exact digital image processing method can be applied in the field of differentiating paper currency instead of manual judging. This paper is about a new field of infrared image application and a new direction in differentiating the true and the counterfeit.