Validation of Harris Detector and Eigen Features Detector

International Conference on Aerospace and Mechanical Engineering 2017

航空航天工程;机械制造

Kok, K.Y.^1 ; Rajendran, P.^1

School of Aerospace Engineering, Universiti Sains Malaysia, Engineering Campus, Nibong Tebal, Pulau Pinang, Malaysia^1

关键词: Application research;    Built-in detectors;    Common features;    Computational costs;    Corner detector;    Harris detector;    Stereo matching;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/370/1/012013/pdf
DOI  :  10.1088/1757-899X/370/1/012013

学科分类：航空航天科学

来源: IOP

【 摘 要 】

Harris detector is one of the most common features detection for applications such as object recognition, stereo matching and target tracking. In this paper, a similar Harris detector algorithm is written using MATLAB and the performance is compared with MATLAB built in Harris detector for validation. This is to ensure that rewritten version of Harris detector can be used for Unmanned Aerial Vehicle (UAV) application research purpose yet can be further improvised. Another corner detector close to Harris detector, which is Eigen features detector is rewritten and compared as well using same procedures with same purpose. The simulation results have shown that rewritten version for both Harris and Eigen features detectors have the same performance with MATLAB built in detectors with not more than 0.4% coordination deviation, less than 4% & 5% response deviation respectively, and maximum 3% computational cost error.