The Target-detection Algorithm Based On The Fractal Features

With the rapid development of computer technology and extensive application of computer vision theory, it is become hotter and hotter to research the target detection and real-time tracking using the computer image processing technology. Dynamic real-time tracking of targets and detection have a very wide range of application in intelligent transport systems, intelligent monitoring systems, military target detection and surgical instruments location in the medical-guided surgery and so on .Target detection has been considered as a complex issue since it appeared, and now it is also a difficulty in the development of Image Processing. Fractal Theory extends our knowledge and enlarges the horizons in research methods; it offers a new technical method for target detection technology. Object detection based on Fractal has its pre-advantages in the implementation and application, using the nature differences between man-made targets and natural backgrounds.This paper mainly studies target detection based on the fractal characteristics, and simulation the target detection algorithm using the commonly fractal characteristics which are fractal fitting error and high-order fractal feather. As deficiencies in the study present algorithms, an improved and effective detection algorithm is introduced in this paper, which combines the wavelet transform and higher-order fractal feature. First, decomposing the original image via wavelet transform, then extracting the wavelet coefficients in low frequency and calculating its high-order fractal parameter .It detects the target position as the fractal characteristic has a sharp different between the target region and the background. The results show that the effect is obvious using this improved algorithm.