共被期刊论文引用了2次
您的检索式：您选中1篇文献正在查看引证文献汇总
|
|
|
题名
|
作者
|
年代
|
出处
|
被引量
|
| 1 | Bird watching in China reveals bird distribution changes显示文摘Using China Bird Report(2003-2007) as data source in combination with descriptions about bird habitats,we built up the China Bird Watching Database.We also developed spherical GIS software 'Global Analyst' to create the point-based database which contains accurate spatial-temporal information.The China Bird Watching Database can reflect the achievement of Chinese bird watchers and complement the basic knowledge of bird distribution.Now a total of 30936 records from 17 Orders,70 Families and 1078 Species of 5 years are included in the database,representing over 80% of all bird species in China.Till 2007,the geographic coverage has encompassed all provincial level administrative districts in China,with the exception of Hong Kong and Taiwan.The China Bird Watching Database also recorded a group of species which are additions at national and provincial levels,including 14 species which are additions to the national checklist and 109 species which appeared outside their original distributions.Comparing the new records with their original distributions,we found the trend that species move to higher latitude and higher elevation regions and some species of waterfowls in Xinjiang Uygur Autonomous Region,including a suite of rare seabirds in the China's Mainland.The majority of bird watchers come from the Eastern Region of China,but their covering range is spreading northwest.At the same time,we appeal to adopting a suite of new technologies for observation,and building up sharing platform of bird watching data to capture the distribution dynamics of birds in China and provide a direct foundation for bird conservation. | LI XueYan LIANG Lu GONG Peng LIU Yang LIANG FeiFei | 2013 | Chinese Science Bulletin2013，58，6： | 13 |
| 2 | 基于Mel子带参数化特征的自动鸟鸣识别显示文摘针对自然复杂声学环境下基于鸟鸣的物种分类问题,提出了一种基于Mel子带参数化特征的鸟鸣自动识别方法。采用高斯混合模型(GMM)拟合连续声学监测数据分帧后的对数能量分布,选取高似然率的数据帧组成候选声音事件完成自动分段。在谱图域对相应片段采用Mel带通滤波器组滤波处理,然后基于自回归模型(AR)分别建模各个子带输出的随时间变化的能量序列,得到能够描述不同种类鸟鸣信号时频特性的参数化特征。最后利用支持向量机(SVM)分类器进行分类识别。基于野外自然环境11种鸟鸣信号开展了自动分段与识别实验,所提方法针对各类鸟鸣的查准率、查全率以及F1度量均不低于89%,明显优于现有基于纹理特征的方法,更适用于野外鸟类连续声学监测领域的自动数据分析需求。 | 张赛花 赵兆 许志勇 张怡 | 2017 | 计算机应用2017，37，4： | 8 |
您所在的机构暂未开通此部分的功能，请联系管理员