Abstract: To deal with difficulty and randomness of artificial visibility evaluation,an automatic quantitative evaluation method of visibility for moving observed object was proposed.The surfaces of observed object were partitioned into triangular meshes,and a virtual field-of-vision model that was similar to realistic field-of-vision was constructed,in which a number of test lines for simulating the realistic line-of-sight were filled into the virtual field-of-vision.The visible triangular meshes were screened by line-triangle intersection test.A layer-by-layer detection algorithm was proposed to improve the detection efficiency,and a quantitative evaluation algorithm was proposed to evaluate visibility factors of each moment in real time.Membership function was also introduced to provide a method for transforming digital visual assessment into linguistic one.The proposed method had been applied to a self-developed virtual assembly prototype system,and the practical results proved that the method could provide accurate and high efficient evaluation in real time,which could provide supporting data for automatic assembly location technology by considering human factors.

Key words: visibility evaluation, moving observed object, automatic quantification, layer-by-layer detection algorithm, dynamic co-simulation

摘要： 针对目前可视性人工评估难度大且评估结果随机性强、不稳定等问题,提出一种对动态观测物进行可视性自动量化评估的方法。将观测物表面划分成三角网格评估单元,构造了接近人眼视野生理特征的视域模型,并在视域模型中均匀地布置用于模拟人眼视线的检测线簇。通过检测线与三角网格的求交运算筛选出实际可见的三角网格来表征观测物的实际可见部分。提出一种基于邻接性的分层检测算法,极大地降低了检测计算量。设计了一套兼顾准确性和计算效率的评估算法模型,对每一仿真时刻的可视性各因素进行实时量化评估。通过引入隶属度函数提供了将可视性数字化评估结果转化为语言型评估结果的机制。该方法已应用于某自主开发的虚拟装配原型系统中,实践证明其评估结果准确、评估效率高,可以为考虑人机因素的自动装配引导与定位技术提供数据支持。

关键词: 可视性评估, 动态观测物, 自动量化, 分层检测算法, 联合动态仿真