Abstract
In this sturdy, an image-based disdrometer has been developed. The system uses a CCD camera to capture falling raindrops and applies proper image processing techniques to determine the characteristics of raindrops. The observed data can be used to analyze raining condition and increase the accuracy of analyzing precipitation in small area. In the system, a telecentric lens is applied with the camera to reduce the blur effect and distortion from out of focus, which is easier to measure the shape more accurately. However, the telecentric lens needs enough illuminance to capture raindrops, so a light source is also applied. Due to the limitation of camera frame rate, it is difficult to measure the raindrop shape and velocity in one measurement. That is, two different exposure time measurements, 100us and 2000us, are applied to measure shape and velocity respectively. Furthermore, the details of image processing procedures are also introduced. Finally, three kinds of experiments were done to verify the system. Marble experiments verified the feasibility and performance of the system structure and image processing procedures. Water sprinkling experiments simulated the raining condition with more uncertainty than the marble experiment. Last, real raining condition pictures were taken and analyzed by the system. The analyzed data was compared with the data from Hsinchu weather station, Central Weather Bureau, Taiwan. An error rate of 14.42% in rain rate was obtained.