双亮差法测量能见度的改进研究

Improved Study on Meteorological Visibility Measuring by Duel Differential Luminance Method

数字摄像测量能见度系统(DPVS)以天空背景和人工目标计算能见度。但提取天空背景的算法很难实现所有环境的适用性,使DPVS不适用于复杂环境的自动化能见度观测。为了解决上述问题,首先基于DPVS的双亮度差法提出了一种新型目标物。其目标部分与背景部分由20%和40%反射率的标准漫反射板构成,外观结构易于图像处理算法识别。其次,根据目标物图像特征,设计了高鲁棒性的识别算法。对比观测试验中,改进方法的观测数据在1~2km范围内与人工观测数据一致,在2〜10km范围内同长春气象站的能见度观测数据具有较好的线性关系。在鲁棒性测试中,改进的双亮差法依然能够保证其数据与对比观测数据的一致性。综上可知,基于新型目标物的DPVS提升了复杂环境下的自动化观测性能。

The meteorological visibility calculated by digital photography visibility system(DPVS),is based on the sky background and artificial target,but the algorithm of extracting the sky background did not work well for all kinds of circumstances,which limits the DPVS works in complex environment.In order to solve the problems above,a brand new designing target based on duel differential luminance method is proposed,which are composed of 20%and 40%standard diffuser as target part and background part.The appearance of new target is easy to identify by image processing algorithm.Secondly,according to the characteristics of new target,a high robust recognition algorithm is designed.In the comparing observing test,the data of improved DPVS are consistent with the artificial observation data in the visibility range of 1 km to 2 km.There is a good linear relationship in the visibility range of 2 km to 10 km compared with the visibility data from Chang Chun meteorological station.In the robustness test,the improved method can still ensure the consistency with the visibility observation data.Above all,this DPVS based on the new target can improved the automatic observation performance in complex environment.