CAJ | 학술논문

本文提出一种基于改进蜂群算法的视频目标检测方法,首先对两幅图像进行优化获得最大互信息值,进而获得最佳空间匹配参数,最后通过三帧差分法检测出目标.该算法相对传统算法,能够抑制背景残留噪声,而且不需要对图像进行预处理、特征选取以及背景更新,降低了算法复杂度.通过与传统蜂群算法的结果对比,证明了改进算法的有效性和可靠性.
본문제출일충기우개진봉군산법적시빈목표검측방법,수선대량폭도상진행우화획득최대호신식치,진이획득최가공간필배삼수,최후통과삼정차분법검측출목표.해산법상대전통산법,능구억제배경잔류조성,이차불수요대도상진행예처리、특정선취이급배경경신,강저료산법복잡도.통과여전통봉군산법적결과대비,증명료개진산법적유효성화가고성.
Here,a video object detection method based on an improved bee colony algorithm is presented.First,the maximum mutual information values of two images are obtained through optimization.Then,the best spatial matching parameters are acquired,and finally the target is detected through the three frame difference method.Compared to the traditional algorithm,the proposed algorithm can restrain the residual background noise,and does not require the image pre-processing,feature selection and background updating,which reduce the complexity of the algorithm.Compared with the results based on the traditional bee colony algorithm,the effectiveness and reliability of the improved algorithm are demonstrated.