摘要
实现对圆的快速轮廓检测,是进行圆形运动目标跟踪与定位的前提。针对圆形目标在成像时发生形变而难以准确检测其外形轮廓的问题,提出一种将随机霍夫变换(RHT)和梯度矢量流主动轮廓模型(GVF Snake)相结合的圆形运动目标检测方法。首先利用改进的随机霍夫变换快速获取目标初始轮廓,解决GVF Snake的初始轮廓自动设置问题,然后基于GVF Snake的轮廓逼近能力,解决传统方法中出现的目标轮廓丢失问题,准确完整地提取目标的真实边界。实验结果表明,该方法有效克服圆形目标的形变,能够快速地提取目标轮廓,具备圆形运动目标的实时检测能力。
Fast detection of circle＇ s contour is the premise of tracking and positioning circular moving object. The shape deformation occurred during imaging the circle object will cause difficulty in accurately detecting its contour. For this problem, in the thesis we put forward a circular moving object detection method which combines the randomised Hough transform （RHT） with the gradient vector flow active contour model （GVF Snake）. First of all, the improved RHT algorithm is employed to quickly detect the initial contour of the object to solve the problem of GVF Snake in its automatic initial contour setting. Then, taking the advantage of the GVF Snake in its Capacity of contour approaching, the method solves the object contour losing problem in traditional method and extracts accurate and entire real boundary of the object. Results of the experiments show, the method effectively overcomes the shape deformation of the circle, it can extract the contour of the object quickly and has the capacity of real-time detection on circular moving object.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第1期211-213,282,共4页
Computer Applications and Software
基金
中央高校基本科研业务费专项资金项目(NP2011049)
