图像跟踪中的边缘检测技术

Edge detection technology in imaging tracking

为提高电视图像跟踪系统的图像检测精度,实现对目标的稳定跟踪,研究一种有效、实时的图像检测方法非常必要。本文介绍了边缘检测技术的基本原理,描述了几种边缘检测方法,如传统的基于经典微分算子的边缘检测、LOG滤波器与Marr-Hildreth边缘检测算子、多灰度图像边缘聚焦法、Canny边缘检测算子、基于梯度信息的自适应平滑滤波和基于小波的边缘检测算子等。给出了边缘检测技术在实际图像跟踪中的应用实例,指出实际的电视图像跟踪系统可以根据不同的图像类型,考虑安全性、稳定性、精度噪声等因素,选择最优的边缘检测方法。

To improve the image detection accuracy for TV image tracking system to realize the stable tracking for a space target, it is necessary for the tracking system to research on a high effective and real time image detection technology. In this paper, the basic conception of edge detection is introduce and several kinds of edge detection methods are described, they are traditional edge detection based on classical differential operators, LOG filter and Marr-Hildreth edge detection operator, edge focus for multi - grey image, Canny edge detection operator, adaptive smooth filtering based gradient information, and wavelet based edge detection operator. Then, some examples of edge detection algorithms used in image tracking are given. Finally, this paper suggests that optimal edge detection method can be selected by considering safety, stabilization, accuracy and image noise based different image forms in photoelectric tracking systems.