May to November is the flood season of Taiwan. After the end of monsoon rain season in late May, occasional torrential rains brought by typhoons or southwester monsoon are considerable challenges for the drainage facilities of Taiwan. Although the monsoon rains and typhoons are the major sources of rainfall in Taiwan, persistent heavy rainfall often triggers landslides, debris flows, collapse of flood control facilities, and other major disasters, seriously threaten the safety and property of the people.
Taiwan is often attacked by typhoons in summer and autumn. For flood control and prevention, in addition to relevant engineering and policy-making improvements, non-engineering measures must be taken to reduce flood damage. It becomes increasingly important to establish an integrated flood forecasting system able to provide real-time hydrological and meteorological forecast data, as Taiwan is on the path of the typhoon in the western Pacific area. Regarding hydrological monitoring, accurate judgments and interpretations of water level changes at various water level stations require the use of “water level identification” technology.
Most traditional methods of water level monitoring rely on direct measurements by measuring instruments installed at water edge, such as a visual water gauge or various types of water level gauge include the float type, the pressure type, and the ultrasonic type. These measurement methods require high-volume equipment, which is difficult to install; while equipment installed at a river bank is prone to mechanical failure or flood damage. Therefore, many water conservancy facilities are currently equipped with monitoring equipment to watch for abnormal changes in water levels, particularly, in the typhoon season. However, it is a great challenge to accurately identify the water level through camera lens in heavy rains or a dark environment. In regard to this problem, the NCHC research team developed a new water level identification method, which identifies water levels using water level images captured by monitoring equipment and coupled with image identification technology that further calculates water level values.
ˇ˝ System architecture diagram
The technology of the proposed water level identification method captures water level images with the lens of image capturing equipment, and obtains candidate images. Candidate images are pre-processed to recognize the position of water level, and by further calculating the relative percentage of the height reference value and the processed image, users can obtain a height value. The pre-processing of an image is employed to perform relevant image processing and enhance the identification effects of the candidate image. Thus, water level height monitoring is more accurately and efficiently controlled.
ˇ˝ Identification method and water level value estimation
This method has the advantage of low cost, as it requires no expensive special equipment, but only a lens with the help of software or hardware for image pre-processing. With no limitation on equipment specifications, the equipment could be installed in areas safe from flooding, and effectively lower equipment failure rates. Reducing identification complexity through image processing considerably improves the effect of identifying water levels. It is applicable in all weather conditions and environments and has the advantages of popularity and convenience.
ˇ˝Identification application-daytime ˇ˝Identification application-nighttime
The technology of the proposed water level identification method can be applied in water level monitoring of flood prevention, and meet the demands of relevant enterprises in height measurements and monitoring of any liquid levels. Future applications in pumping station management of water reservoirs, farming and fishery aquaculture, chemical measurements are expected to considerably improve efficiency, thus bringing the infinite potential of the technology into full play.