Author(s): Kwonkyu Yu; Namjoo Lee; Taewook Kang
Linked Author(s):
Keywords: Flow velocity; Image processing; Surface image velocimeter; Perspective transform; Camera model
Abstract: The present study’s aim is to develop a real-time SIV (Surface Image Velocimeter) system on a smartphone with an Android operating system. Unlike the traditional SIV system, this new SIV system performs image capturing, image processing and post processing all in one device and the water surface velocity is calculated almost instantly. The site location is detected by the smartphone’s GPS. The azimuth, pitch and roll angles of the instrument are measured automatically and the intrinsic, extrinsic camera parameters are calculated. The only parameter that needs to be manually set by the user is the height of the phone from the surface of the water. After the parameters are set, a series of images are taken and for the normal NTSC type video camera, 30 frames per second are captured. With the help of Open CV, an open source computer vision library, the video frames are split and the images are analyzed to obtain the velocity field of the water surface. The image processing algorithm was based on an algorithm for the correlation analysis of spatiotemporal images. The code was implemented with Java to the Android operating system. Since all the image processing is performed solely by smartphone, water surface velocity is measured literally in real-time. Utilizing the measured camera orientation parameters, coordinate transform parameters are automatically calculated by the system. A test performed in a small channel demonstrated the effectiveness and the convenience of the developed system. Measured results showed that the system was fairly useful for river flow measurement work.
Year: 2017