基于无人机图像分形特征的油松受灾级别判定

利用无人机采集油松样地图像，提取图像中的单株样本树图像，计算单株样本树图像的多个纹理特征值，对纹理特征值进行灾害分级，与地面基于失叶率调查的灾害分级进行比对，探索能准确描述油松受灾情况的无人机图像纹理特征，实验结果表明：受灾油松图像的三种分形特征，即分形维数、缝隙量及维数升降因子能较好地反应油松的失叶率状况，可作为油松受灾级别的图像判定特征，同时上述分形特征也适用于整块油松样地的受灾级别判定。

The judgment on the disaster classification of Chinese pine based on the fractal features in UAV image

The unmanned aerial vehicle (UAV) is useful in taking remotely sensed images for forestry applications. We use the UAV to acquire images of Chinese-pine field and then segment the individual sample pine-tree from images. The texture features are extracted for each individual sample pine-tree. Based on the calculated texture features and the leaf loss rate, the degree of disaster can be classified and compared with the ground survey results to exploring the most appropriate texture features for this study. Experimental results show that three fractal texture features including Fractal Dimension, lacunarity and Fractal Dimension Gradient can be used for predicting the leaf loss rate of individual tree accurately. With this experiment, we conclude that this approach can be applied to the disaster classification issue of the Chinese-pine field for images obtained by the UAV.