Abstract:In order to solve the shortcomings that the existing methods of measuring the area of plant leaves, for example, the traditional direct contact measurement can cause damage to the plant, and the laser scanner or structured light to reconstruct the 3D model is expensive and complicated. A non-destructive and multi-category measurement method of irregular leaf area was proposed. Firstly, smart phone was used to get images of plants, based on these images, the three-dimensional point cloud of plants was reconstructed by using structure from motion, and the noise of the leaf point cloud was removed by using threshold segmentation algorithm with HSV color space as features. Secondly,the three-dimensional point cloud was classified by using the fuzzy C-means clustering algorithm to segment the single leaf point cloud, and then, the surface mesh model of the single leaf was reconstructed by using the Delaunay triangular meshing algorithm. Finally, the leaf area was obtained by calculating the mesh area. To prove the proposed method, five different types of plant leaves with different shapes were measured, and it was compared with the real leaf area. The average error of the calculation result of the proposed method was 6.25% in terms of leaf overlap data, and from the perspective of the complexity of the blade shape, the average error was 4.81%. In addition, the samples had been added that actually overlap each other between the leaves for experiments, which increased the reliability of the experiment.The results showed that the method was stable and had high precision, meeting the needs of measuring the leaf area of irregular plants in multiple categories.