3D Reconstruction of Orchid Based on Virtual Binocular Vision Technology

It is normally challenging to model flowering plants, which is affected by illumination, temperature, moisture, soil, etc. To further improve the accuracy and efficiency of such 3D modeling, a 3D reconstruction method for orchid flower, based on virtual binocular stereo vision, was proposed in this paper, where two cameras with baseline and fixed focal length were applied to constitute a binocular stereo vision system. The improved Tsai two-step calibration method was also implemented to calibrate the cameras; a stereo matching algorithm based on the improved Harris corner points and regional joint was adopted to match features. Subsequently, the cloud information of 3D space points was obtained through calculation. The Delaunay triangulation method of 3D space point set was then improved, by applying the local 3D Delaunay algorithm of mesh boundary construction, to achieve 3D reconstruction based on the OpenGL platform. The simulation results suggest that, the proposed method is characterized by high operational speed, good robustness and high modeling accuracy, valuable as reference for 3D modeling of other flowering plants.