Plant pressing is a form of dried flowers in a flat shape, combining high artistic and economic value. However, its complex creation techniques currently limit large-scale production. This paper employs a process of “material collection—processing and integration—color preservation and dyeing—pressing—shape design” to prepare plant materials. Geometric correction and grayscale conversion are applied to the original plant images to reduce errors caused by photography, thereby obtaining computer vision images of the plant materials. Based on this, coordinate transformation technology is used to convert the tangent point position coordinates into world coordinates at the time of collision. Ray detection technology is employed to solve the shortest distance from all array points to the object, thereby completing the extraction of virtual shape information from the plant materials. Red magnolia is selected as the experimental material, and its plant leaf images are preprocessed using the method described in this paper to extract leaf features from different growth regions and construct virtual reality scene images. Compared with three commonly used similar methods, the structural similarity of the reconstructed plant images using the method proposed in this paper ranges from 0.941 to 0.988, demonstrating more reliable practical performance.
