从利用图像信息的角度，系统分析了图像特征检测问题的研究文献，将图像特征检测方法分为两大类，即基于梯度信息的方法和基于相位信息的方法。论述了每类方法的特点，并对主要的特征检测方法进行详细而全面的论述。指出了目前研究中存在的主要问题以及今后的研究方向。

Image feature detection is a key technique in image processing, pattern recognition, content-based image retrieval, and so on. The typical image feature detectors are analyzed. From the viewpoint of information used, the literatures on image feature detection are analyzed systematically, and image feature detectors can be categorized in to two kinds, that is, gradient-based ones and phase-based ones. Characteristics of each kind are discussed. The key issues of the important image feature detector are explained extensively and comprehensively. Based on the survey, some concerned issues needed further research are presented, and the future research trends are also given.