文献类型：期刊文献

中文题名：基于面向对象的石漠化山区植被信息提取及分布特征研究

英文题名：Research on Vegetation Information Extraction and Distribution Characteristics of Rocky Desertification Area by Means of Object Oriented Approach

作者：郑惠茹[1] 罗红霞[1] 向海燕[1] 程玉丝[1] 杨任飞[1] 程全英[2]

第一作者：郑惠茹

机构：[1]西南大学地理科学学院;[2]中国林业科学研究院资源信息研究所

年份：2018

卷号：27

期号：3

起止页码：648-657

中文期刊名：长江流域资源与环境

外文期刊名：Resources and Environment in the Yangtze Basin

收录：CSTPCD;;国家哲学社会科学学术期刊数据库;Scopus;北大核心:【北大核心2017】；CSSCI:【CSSCI2017_2018】；CSCD:【CSCD2017_2018】；

基金：国家自然科学基金(41201436)

语种：中文

中文关键词：多尺度分割;面向对象分类;地形因子;植被空间分布

外文关键词：multi-scale segmentation; object-oriented classification; terrain factor; vegetation spatial distribution

分类号：Q948

摘要：重庆东北部地区是重庆岩溶石漠化重点治理区,该区地形复杂,山高坡陡,植被退化现象严重。了解该地区的植被分布特征,对该区环境的改善和石漠化治理具有十分重要的现实意义。基于Landsat OLI等数据,运用面向对象分类方法对研究区植被信息进行提取,然后对分类后的数据进行统计、制图和分析,并在空间布局上进行总结和探讨,旨在了解该区域植被的空间分布特征和规律。结果表明:(1)在e Cognition 9.0软件中进行多尺度分割,再结合地物类型特征使用隶属度函数法进行分类,该方法符合山地地物类型的分布规律和特点,分类精度达到81.35%;(2)研究区属典型的中山地区,海拔在500~1 500 m之间的地区约占64.49%,林地和耕地是该区域主要的地物类型,所占总面积为8 872.22 km^2,占研究区总面积的96.50%,各地物类型分布受地形地势的影响较大;(3)综合研究区地形因子(高程和坡度)与植被分布的相关性可知,耕地和草灌主要分布在高程为200~1 500 m且坡度等级在2~4级(5°~35°)之间,该区域人类活动频繁,故受人类活动影响较大,植被覆盖度低,群落生长不稳定,容易受到干扰。针阔混交林主要分布在高程>500 m且坡度等级在2~4级(5°~35°)之间。马尾松林、阔叶林和柏木林主要分布在高程大于500 m且坡度等级在2~5级(5°~45°)之间。
Northeastern Chongqing is the key harnessing area of Karst rock desertification,where there is complex terrain,steep slopes and serious vegetation degradation. Knowledge of vegetation distribution pattern in this area is of great practical significance for the improvement of the environment and the control of rock desertification. By using Landsat 8 OLI remote sensing image,30 m-resolution DEM data and field survey data,we extracted the vegetation information with an object-oriented image classification method. Then,statistics,mapping and analysis on the classified data were implemented to comprehend the characteristics and law of vegetation spatial distribution pattern in this area. The results showed that：（ 1） after multi-scale segmentation with e Cognition 9 software,land cover was classified by using the membership function method based on land cover features. This approach is applicable to the land cover classification in the Karst rocky mountain areas,the classification accuracy reached 81. 35%.（ 2） The study area is typical of the middle mountain area. The area at an altitude between 500 and 1 500 meters accounted for 64. 49%. Woodland and cropland are the major land cover types in this area,which cover an area of 8 872. 22 km^2,accounting for 96. 50% of the total area of the study area. The distribution of various land cover types is greatly influenced by the topography.（ 3） The results showed that cropland,grass and bush mainly distributed between the elevation of 200 and 1 500 meters,with the slope grade between level 2 and 4（ i. e. 5°-35°）,according to the correlation between terrain factors（ i. e.elevation and slope） and vegetation distribution. There are frequent human activities,low vegetation coverage,and unstable community growth in this area,which make it vulnerable to interference. Mixed conifer-broadleaved forest is mainly distributed in the elevation higher than 500 meters,with the slope grade between level 2 and 4（ i.e. 5°-35°）. Masson pine forest,broadleaved forest and cupressus funebris forest are mainly distributed in the elevation higher than 500 meters,with the slope grade between level 2 and 5（ i. e. 5°-45°）.