关于山羊牧道生态景观发育程度衡量算法的研究

发布时间：2018-02-08 23:16

本文关键词： 山羊牧道景观深度扩散元胞自动机连通域深度生态景观　出处：《兰州大学》2015年硕士论文　论文类型：学位论文

【摘要】：山羊牧道广泛分布于中国黄土高原干旱半干旱丘陵地区,其形成和发育对水土流失有着很重要的影响。邹梅等首次使用元胞自动机模拟了山羊牧道生态景观的形成和发育,其模拟结果可处理成包含山羊牧道裸地和植被的二值图像。而在使用元胞自动机模拟山羊牧道景观的形成和发展中,获取景观的发育程度信息有助于进一步实验和分析。本文提出了“纵向连通深度”的概念用于衡量山羊牧道的发育程度。通过为模拟结果图像中的每个像素定义一个纵向的连通深度值,可以估算景观发育程度的整体分布信息。同时本文提出了两种用于计算这种“纵向连通深度”的算法——“深度扩散算法”和基于连通域标记的算法。“深度扩散算法”通过模拟液体压强分布特征来计算“纵向连通深度”,简单明了且容易实现；而基于连通域标记的算法首先对模拟结果图形做连通域标记,然后在此基础上计算每个像素的纵向连通深度。实验结果表明两种算法都可以有效地计算山羊牧道景观模拟结果图像中每个像素的纵向连通深度。但基于连通域标记的算法具有较好的稳定性,但整体性能较低；深度扩散算法的性能表现很好,一般情况下比基于连通域的算法快20倍左右,但稳定性较差,在特定情况下甚至出现运行时间指数级别增长。
[Abstract]:Goat herding is widely distributed in arid and semi-arid hilly areas of the Loess Plateau of China, and its formation and development play an important role in soil erosion. Zou Mei et al used cellular automata for the first time to simulate the formation and development of goat pastoral ecological landscape. The simulation results can be processed into binary images containing goat pastoral bare land and vegetation. In the process of using cellular automata to simulate the formation and development of goat pastoral landscape, Obtaining the developmental degree information of the landscape is helpful for further experiment and analysis. This paper proposes the concept of "longitudinal connectivity" to measure the development of goat herding. By defining each pixel in the simulated result image. A longitudinal depth of connectivity, The global distribution information of landscape development degree can be estimated. At the same time, two algorithms for calculating the "longitudinal connectivity depth", "depth diffusion algorithm" and "depth diffusion algorithm based on connected domain marker", are proposed. "the" longitudinal connection depth "is calculated by simulating the distribution of liquid pressure, which is simple and easy to realize. However, the algorithm based on connected domain labeling firstly marks the graph of simulation results in connected domain. The experimental results show that both algorithms can effectively calculate the longitudinal connectivity depth of each pixel in the goat pastoral landscape simulation image, but based on the connected domain. The labeling algorithm has good stability. But the overall performance is low and the performance of the depth diffusion algorithm is very good. In general, it is about 20 times faster than the algorithm based on connected domain, but the stability is poor, and the running time exponential level increases even under certain circumstances.
【学位授予单位】：兰州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41

【相似文献】