摘要
针对REVS-50M小型光电跟踪系统基于颜色阈值检测所存在的误跟踪问题,设计了形状检测实验。在实验过程中,提出采用Hough变换检测圆形手动靶标的方法。但常规Hough变换存在许多缺点,并不适用于较为复杂的情况。进而引导学生采用随机Hough变换,并针对其随机采样带来的缺陷,对随机Hough变换做出了一些改进,利用了梯度信息减少无效累积,并在判断真圆时缩小了验证范围,从而优化了算法。
Appointing to the mis-tracking problem of REVS-50M miniature optical tracking system,the paper designs an experiment of shape detection.In the experiment,Hough transform method is used to detect the circle.But conventional Hough transform has many flaws.It- not able to deal with complex problems.Then we guide students to use randomized Hough transform.Considering the flaw of the randomized Hough transform by random sampling,some improvements including utilizing the gradient information to reduce the invalid accumulation and refining the process are used to verify the real circle.These improvements optimize the algorithm.The teacher guides students gradually to improve the experimental method such that students＇ability of analyzing problems and solving problems are enhanced.The courses achieve good teaching results.
出处
《实验室研究与探索》
CAS
北大核心
2014年第7期130-133,共4页
Research and Exploration In Laboratory
基金
北京市教育委员会共建重点实验室资助项目(CSYS100070417)
