Automated 3-D tracking of video-captured movement using the example of an aquatic mollusk

We describe a computer program that tracks the three-dimensional coordinates of a snail (moving within a tank) from images captured by a single video camera. The tank and a mirror placed at 45° to one side are viewed to provide direct and reflected images of a pond snail,Lymnaea stagnalis, moving around the tank. The output of the camera is captured with an IBM-compatible PC using a frame-grabbing VideoBlaster board. The main features of our Windows program is the frame-by-frame analysis of the captured video, enhancement of contrast, location of dark “snails,” discrimination by size, trigonometric determination of thex, y, z coordinates, and the final export of data in a spreadsheet-ready format. After the user has outlined the position of the tank on the screen and set the discrimination thresholds, no further user activity is required. This makes the analysis of slowly moving mollusks much less operator intensive and less tedious. The method is applicable to other faster moving organisms (e.g., fish) because the analysis is performed on captured video sequences.