武夷山九曲溪湿地鸟类物种辨识①-计算机系统应用.PDF

计 算 机 系 统 应 用 2013 年 第 22 卷 第 9 期 武夷山九曲溪湿地鸟类物种辨识① 石贵民 (武夷学院 数学与计算机系, 武夷山 354300) 摘 要: 唐庆和, 唐兆园等人在 2006 年 12 月～2008 年 3 月采用“点样线法”对九曲溪生态保护区湿地鸟类多样性 进行了调查研究, 共记录鸟类 65 种. 近年来, 随着数字图像识别技术的快速发展, 对动、植物物种的鉴别已成为 一个热门的研究领域. 本文在保护九曲溪水生鸟类的前提下, 使用图形分析技术对鸟类物种进行辨识研究; 使用 SVM 决策树法, 对鸟类颜色图表(包括形状, 颜色和纹理)进行深入分析和研究, 提出一套分类方法. 该方法的应 用将更有效的识别九曲溪湿地鸟类物种, 为鸟类生物多样性研究及保护提供科学依据. 关键词: 九曲溪; 湿地; 鸟类; 生态; SVM 决策树 Recognition of Bird Species in Nine Turns Stream Wetland SHI Gui-Min (Department of Mathematics and Computer, Wuyi University, Wuyishan 354300, China) Abstract: From March 2006 to Decembe2008, sample and line transect methods were used to investigate bird diversity in Nine Turns Stream wetland in Fujian Province, A total of 65 species were recorded. With the rapid development of digital graphic recognition technology, in recent years, more attention is given to the subject of animal and plant species identification which has become a popular researching field. The article mainly discusses about the method to apply graphic analyzing technology for bird identification research in the case of aquatic bird protection in the marsh of Nine Turns Stream. It also put forward a set of classification case with SVM Decision Tree for protesting information through deep analysis and study on bird color graph in terms of shape, color and texture. Moreover, the appearance of the case contributes to the effective recognition and classification of birds in species Nine Turns Stream wetland,and then provides the scientific basis for bird biodiversity research and conservation. Key words: nine turns stream; wetland; bird species; ecology; decision tree of SVM 湿地是自然界生物多样性最为丰富的生态系统, 闽赣两省边界, 地理位置介于 117035～117048.79E、 具有调蓄洪水、调节气候、涵养水源、净化水质等多种 27035～27047N, 总面积 43.400 公顷.