雨滴微物理特性及降雨动能是揭示降雨物理本质的重要特征量,亦是开展侵蚀定量分析与建立侵蚀量预报模型的基础.采用粒子成像瞬态测量可视化技术观测自然降雨雨滴,结合计算机视觉识别技术解算雨滴微物理特性参数,同时采用虹吸式自记雨量计记录自然降雨降雨强度.研究表明:该次降雨雨滴以中等粒子为主,雨滴直径均值为1.52 mm,降落末速度均值为3.47 m/s,其中直径在1.00~3.00 mm范围内的雨滴占样本总数的87.21%.雨滴直径和降落末速度呈显著的对数关系.基于实测的雨滴微物理特性和降雨强度估算降雨动能,该结果与传统的经验模型估算结果相对误差均值为7.28%.该方法得到的降雨动能较以往的经验模型能更真实的反应雨滴降落过程中的做功大小,为准确计算降雨过程中雨滴所造成的溅蚀量奠定基础.%Microphysical features of raindrop and rainfall energy are the key parameters for study of rainfall physics, which also have great significance in quantitative analysis of soil erosion and in soil erosion prediction model. The existing measurement methods include splash method, immersion method and so on, but most of them have many disadvantages, such as, labor and time consuming, poor real-time response, low precision and so on. Therefore, a new method should be developed urgently. In order to obtain physical properties of raindrops, in this study, the particle imaging transient visual measurement technology, light field and imaging system were designed, image identification, extraction and measurement were investigated, and finally the particle imaging transient visual measurement technology and computer vision identification technology were used to obtain the microphysical features of natural raindrop. The principle of the system could be described as below: a Fresnel lens was installed in the front of the light source. When the lights were passing through the Fresnel lens, the lights from point source would become parallel lights, which would shine on a projecting screen, and then the raindrop would project on the screen during its falling. Specifically, the system consisted of three parts: projection system, image capture system, and image control system. In the image capture system, two cameras were used to capture the raindrop image, one with a fast speed to obtain static image, and the other worked slowly to capture the tailed image. Based on the two images, we calculate the diameter and the fall velocity of raindrops. In order to obtain a clear raindrop image, we must remove the noise in the images. Basically, the image noise removal involved four steps. First, it was statistical filtering; second, the rolling filtering; third, the smooth filtering, and finally, it was the image binarization. Based on the static image and the corresponding tailed image mentioned above, geometric mean value algorithm and outer contour algorithm were used to calculate the diameter and fall velocity of raindrop. In our previous research, we found that the measurement technology had small relative error and it was suitable for the measurement of microphysical features of raindrop. Meanwhile, rainfall intensity was recorded by siphonic pluviograph. The results showed that medium-sized particles were the predominant contributor in the single rainfall. Raindrop diameter and fall velocity in our study were, on average, 1.52 mm and 3.47 m/s, respectively. Specifically, the proportion of raindrops with diameter ranged from 1.00 to 3.00 mm was up to 87.21%. Fall velocity was strong logarithmically related to raindrop diameter, and more precisely, fall velocity grew rapidly with an increase in rainfall diameter when the diameter was below 1.5 mm. As the raindrops fatten, the growth rate of fall velocity was reduced. Moreover, rainfall energy calculated in the present study was compared with the classic statistical model, and the relative error was averaged as 7.28%. In all, microphysical features of raindrop and rainfall energy can be measured precisely by the technique in this study, which sets the basis of estimating rainfall splash erosion.
展开▼
