3-D image measurement of the sea for disaster prevention

Abstract

To reduce the loss of human life in disasters, we have developed an early warning system for typhoons and tsunamis. The system is designed to monitor the change of the sea level from a distance of 20 km. And it can alert a coming tsunami 20 min before reaching the coast. Now, GPS Buoys and observation networks are used to forecast tsunami. However, the GPS buoys are not conducive to monitoring wide ranges of the sea. The observation networks have the disadvantages of high installation costs and difficult maintenance. In this paper, a measurement system of long distance sea level height for disaster prevention has been constructed based on a 3-D image measurement technique. First, to monitor a wide range of the sea surface, this system utilizes the pan and tilt stages and angle sensors to rotate the cameras. To improve the accuracy of long distance 3-D image measurement, a long-distance calibration method of camera system has been proposed based on the scale poles. In addition, all the equipment can be put in one place so it is very convenient to maintain. In this experiment, a dynamic threshold method is proposed in blocks to extract the waves automatically, accompanied by a wave feature matrix matching method in two long-distance images. Finally, the average value of wave height is used as the height of sea level. As a practical use result, a comparison has been made between the measurement results and the data of Japan Meteorological Agency (JMA).