Study On Semantic Model Of Fish Behavior In Three-dimensional Spatial

As the world’s water pollution problem is getting worse and worse,various countries are actively involved in the construction of water pollution warning system.Aquatic organisms are widely used in study on variation of water environment and most of these studies used fish as a biological indicator.An important premise of researching fish behavior is to detect and extract fish target.Thus,effectively detecting,accurately separating and locating fish target in three-dimensional space is the key technology to study fish behavior under water.In this paper, experimental platform in three-dimensional space is built and the red carp is used as the research object. Combined with the imaging principle of underwater object, imaging principle of camera, object segmentation, theory of plane geometry and image processing technology, we studied the semantic model of fish behavior in three-dimensional space. The main work in this paper includes:Study on technology of color image segmentation. As the research object of this paper is the red carp, we present a color image segmentation method for experimental object in RGB color space on the basis of predecessors’ work.Accurate location of fish target based on three-dimensional space. According to the coordinate transformation relation between solid fish in water tank and virtual fish in mirror,reducing three-dimensional coordinates of fish target.Identification method of half-occlusion fish body in three-dimensional space.After obtaining the corresponding relations between fish in water tank and in the mirror, identifying the occlusion fish by technology of fish body contour recognition and centroid point extraction. And according to the calculated coordinates of heads,tails and centroid points to identify the bracket and swimming direction of half-occlusion target fish.