摘要
雨滴颗粒的数目、形状、尺寸以及谱分布是准确反演降雨强度的基础。为了获取自然环境下的雨滴微观特征量,采用迭代最大熵分割算法,获取雨滴图像的灰度区域阈值,通过阈值比较,把雨滴颗粒从图像中分割出来,在此基础上,把凹点匹配分割算法引入粘连雨滴的分割研究中,以链码差和曲率作为判断准则,选取凹点进行匹配,实现粘连雨滴的精确分割,以几何均值算法标记不同雨滴颗粒质心位置,记录雨滴特征量,提出了一种新型的雨滴微观特征提取方法。通过实验表明,该算法不仅能够有效地从图像中提取边界区域明朗且颗粒完整的雨滴,而且能够对粘连雨滴颗粒进行分割,从而可精准地从图像中获取雨滴微观特征量。
The number,shape,size and spectral distribution of the raindrop particles are the basis for accurately inverting the rainfall intensity. In order to get the microscopic characteristics of raindrops in the natural environment,this paper took the iterative maximum entropy segmentation algorithm to obtain the gray region threshold of the raindrop image,separated in the particles from raindrop image by comparing the thresholds,on this basis,it introduced pit match segmentation algorithm to the study of separating adhesive raindrop which was based on chain code difference and curvature as the judging criteria and selected match pits to achieve accurate segmentation of adhesion raindrop,after that using the geometric mean algorithm to mark the centroid position of different raindrops and recorded the microscopic features of raindrops,from this,proposed a new type of raindrop micro-feature extraction method. Experimental results show that the proposed method can not only extract the clear and complete raindrops from the image,but also segment the adhesive raindrops,eventually,accurately receiving the microscopic features of raindrops from the images.
作者
行鸿彦
张兰
郭敏
Xing Hongyan;Zhang Lan;Guo Min(Collaborative Innovation Center on Forecast ＆ Evaluation of Meteorological Disasters;Jiangsu Key Laboratory of Meteorological Observation ＆ Information Processing,Nanjing University of Information Science ＆ Technology,Nanjing 210044,China)
出处
《计算机应用研究》
CSCD
北大核心
2018年第6期1883-1886,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(61671248)
江苏省产学研联合创新资金计划资助项目(BY2013007-02)
江苏省高校自然科学研究重大项目(15KJA460008)
江苏省六大人才高峰计划
江苏省信息与通信工程优势学科资助项目
关键词
迭代最大熵
阈值
凹点匹配
雨滴特征量
iteration maximum entropy
threshold
pit matching
characteristic value of raindrop
