科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2015年
8期
111-113
,共3页
运动图像%权值迭代%特征融合%定位跟踪
運動圖像%權值迭代%特徵融閤%定位跟蹤
운동도상%권치질대%특정융합%정위근종
moving image%weighted iterative%feature fusion%location tracking
为有效对运动图像的目标特征进行较好定位与跟踪,提出了一种基于多权值相似均衡的运动图像循环定位跟踪算法.该算法首先利用颜色与纹理构建运动图像的空间特征分布跟踪模型,通过特征聚类对运动图像自动划分目标空间,然后根据目标模型的相似函数计算多个特征权值对目标模型进行选择性均衡,并结合特征权值自适应迭代定位跟踪信息,以最大程度地对运动图像的细节特征进行定位跟踪.实验结果表明,该算法优于传统的定位跟踪方法,尤其在跟踪目标受到外界干扰时能够得到比较理想的跟踪效果.
為有效對運動圖像的目標特徵進行較好定位與跟蹤,提齣瞭一種基于多權值相似均衡的運動圖像循環定位跟蹤算法.該算法首先利用顏色與紋理構建運動圖像的空間特徵分佈跟蹤模型,通過特徵聚類對運動圖像自動劃分目標空間,然後根據目標模型的相似函數計算多箇特徵權值對目標模型進行選擇性均衡,併結閤特徵權值自適應迭代定位跟蹤信息,以最大程度地對運動圖像的細節特徵進行定位跟蹤.實驗結果錶明,該算法優于傳統的定位跟蹤方法,尤其在跟蹤目標受到外界榦擾時能夠得到比較理想的跟蹤效果.
위유효대운동도상적목표특정진행교호정위여근종,제출료일충기우다권치상사균형적운동도상순배정위근종산법.해산법수선이용안색여문리구건운동도상적공간특정분포근종모형,통과특정취류대운동도상자동화분목표공간,연후근거목표모형적상사함수계산다개특정권치대목표모형진행선택성균형,병결합특정권치자괄응질대정위근종신식,이최대정도지대운동도상적세절특정진행정위근종.실험결과표명,해산법우우전통적정위근종방법,우기재근종목표수도외계간우시능구득도비교이상적근종효과.
In order to better location and tracking for moving image target characteristic, a location tracking algorithm for moving target image based on multi-weighted tterative is proposed. The algorithm firstly according to the spatial characteris-tics of color and texture to construct the motion image distribution tracking model, through the feature clustering on a moving image automatic segmentation of target space, and then a plurality of feature weights on the target model of selective equilib-rium calculation according to the similarity function model of the target, and make use of the characteristic of adaptive weight iterative location tracking information, to the greatest extent of the moving image the details of the characteristics of the positioning and tracking. The experimental results show that, the positioning of the algorithm is superior to the traditional tracking methods, especially in the tracking target by the outside interference can get relatively ideal tracking effect.