现阶段常用的基于图像灰度的点特征算子有Moravec算子、Harris算子、Forstner算子和Forstner改进算子,不同算子的提取速度、定位准确性、稳定性存在一定的差异。通过人造测试影像和真实拍摄影像测评各算子的提取速度、定位准确性、稳定性、参数自适性,比较得出各算子的优缺点,并在此基础上提出一种改进算子。研究结果表明:改进算子具有稳定性强、定位精度高和自动设置阈值的特点。
This method based on image' s gray to detect the interest pointis mainly includes Moravec operator, Harris operator, Forstner operator and the improved Forstner operator nowadays, and there are some differences in the detect speed, positioning accuracy, stability between different operators. In this paper the detect speed, posi- tioning accuracy, stability and parameters' suitability of each operator are evaluated by testing a artificial test imase and a really taken image, and the test results show the advantages and disadvantages of each operator. A improved operator is proposed based on the test results. The study results show that the improved operator has the characteris- tics of strong stability, high positioning accuracy and auto-setted threshold.