消防炮射流轨迹图像分步逐级筛选方法  

Stepwise Selection Method of Fire-fighting Cannon Trajectory

作　　者：苏浩[1] 赵敏[1] 

机构地区：[1]南京航空航天大学,江苏南京210016

出　　处：《机械制造与自动化》2015年第1期217-219,共3页Machine Building & Automation

基　　金：航空科学基金(2012ZC52042)

摘　　要：在基于逐点搜索的计算机视觉智能消防系统的消防炮射流轨迹的识别中,围绕多起点搜索技术搜索射流轨迹后筛选最优轨迹的问题,鉴于已有筛选方法寻优能力和抗干扰能力不强,提出了一种分步骤的逐级筛选寻优方法。采取分步处理方式,第一步进行基础筛选,采取两种不同来源的信息并行进行基础筛选,保留优质的轨迹的同时,去除完全不符合射流特征或偏差很大的轨迹;第二步根据粒子群算法的适应度寻优思想和射流轨迹本身的特点,设置适应度值计算规则,再根据适应度值的计算结果,进行筛选择优,使整个筛选方法有较强的可靠性和对环境有较好的适应性,并已在visual studio平台下实现,为智能消防系统的轨迹搜索提供了一种较为理想的筛选方案。In the intelligent fire control computer vision system,based on the cannon jet path recognition by point search method,around the problem of screening the optimal trajectory,this paper proposes a stepwise filtering optimization method. It this method,stepwise processing is adopted,first of all the basis is selected,that two different sources of information are taken out and filtered to remove the trajectory which is completely not in conformity with the jet characteristics or great deviation,and at the same time to retain one with high quality,then according to the fitness of particle swarm optimization（ PSO）,and the characteristics of the jet path,the rule for calculating the fitness value is set up to choose the best trajectory. The whole screening method has strong reliability and good adaptability. It is applied to the visual studio platform and an ideal filter scheme is provided for intelligent fire control system of trajectory search.

关 键 词：消防炮射流轨迹 分步筛选 并行筛选 基础筛选 目标提取 适应度值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]