Research on leaf recognition and simulation

This study addresses an algorithm to simulate some attributes of a leaf, including its contour, main vein and lateral vein. It covers the synthesis techniques including image segmentation, identifying object zone, and picking up zone shape. Firstly, its contour is extracted from its 2D image with a Sobel operator. Secondly, its contour and vein is simulated with a 1/f noise modeling method integrating with a curve fitting technique. Finally, its whole attributes is simulated by adopting a leaf area filling method. The experiments show that this algorithm is effective and can meet the user´s demand. This study may contribute to furthering micro-irrigation project design, virtual reality, and special species identification to some extent.