Automatic water-level measurement system for confined-space applications.

Abstract

To prevent flooding from threatening people's lives and property, surveillance cameras have been extensively installed along rivers to monitor the water level. In this study, an automatic water-level measurement system based on a single camera is proposed. The system uses the images captured by the camera to verify the water levels of the rivers. The preliminary positions of the current water marks are first identified from the images of the staff gauges installed on site using the basic digital image processing technology. To establish the correct water levels from the preliminary water mark positions, the system uses the image histogram of the staff gauge region of interest to provide a reference for the water-level determination mechanism. The camera position typically results in a nonorthogonal angle between the camera's optical axis and the on-site staff gauge plane because it causes perspective distortion that eventually leads to deviations. When the water level is low, the distance of the water line deviation from the orthogonal position becomes larger, and the actual distance represented by the pixel also becomes larger, resulting in greater deviations. To address this issue, this system adopts the inverse perspective mapping method to rectify the perspective distortion. By setting the reference water level and control points, this system can accurately calculate water levels. Experimental results demonstrate that the proposed automatic water-level measurement system as applied in confined spaces can reduce the image measurement error caused by the perspective distortion as well as verify and identify the current water levels.