摘要
针对大规模空间封闭图形的高效识别,考虑与识别有关的各个阶段,给出基于转向法的空间封闭图形识别总体方案。围绕封闭图形唯一性和确定性要求,基于转向法定向封闭几何事实,提出一种顺(逆)时针搜索方法来识别单个封闭图形,并辅以区域划分排序方法、动态优化搜索目标、优先级设置等实现大规模空间封闭图形的高效识别,实际应用表明该方法的有效性。
Aiming at efficient recognition of large-scale closed space graph, an overall scheme of recognition based on turning is given, considering the various stages related to recognition. Centering on the uniqueness and determinacy of single closed graph, a counter-clockwise or clockwise recognition method for single closed graph based on beamed-close of turning is put forward. With the region division rules, dynamic optimization rules and priority setting rules, the efficient recognition of large-scale closed space graph is achieved. The effectiveness of the method is proved by the practical application.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第14期184-186,211,共4页
Computer Engineering
基金
"十一五"国家科技支撑计划基金资助项目(2006BAF01A27)
关键词
图形识别
空间封闭图形
转向法
区域划分
动态优化
graphics recognition
closed space graph
turning method
region division
dynamic optimization
