摘要
在湖南省按照不同地貌类型和地类破碎程度抽取4 km×4 km大样地10个,采用不同判读方式(充分参考林业档案信息与不参考林业档案信息)和不同判读人员(有5年以上遥感判读经验人员与无任何遥感判读经验人员)对10个大样地进行遥感判读区划,并对判读区划结果进行实地调查验证。分析表明:各地类正判率与其面积成正相关,其它调查因子判读难度较大;充分参考林业档案信息能够显著提高正判率,但优势树种、起源、龄组、郁闭度等因子综合正判率仍然偏低;不同判读工作经验人员对遥感判读有一定的影响,无判读工作经验人员应经过一定程度培训后才可参与判读工作。
In this paper, according to the different types of geomorphology and crushing degree of land type, we chose 10 samples with area of 4 km ×4 km in Hunan province. Remote sensing images of 10 samples were interpreted by different interpretation ways（ with full reference forestry archives information or without it） and different interpretation personnel（ with more than 5 years remote sensing interpretation experience or without any remote sensing interpretation experience）. Then, we conducted field investigation and verification on the interpretation regionalization results. The results showed that, 1 ） the accuracy of remote sensing were positively related with area of each land type; other inventory factors were difficult to interpreted ;2） full reference forestry archives information could significantly improve the rate of correct discrimination, but the correct interpretation rates of other inventory factors were still low ; 3 ） different interpretation work experience of remote sensing interpretation had certain influence; personnels without interpretation work experience should be trained to do interpretation work.
出处
《中南林业调查规划》
2015年第2期39-43,共5页
Central South Forest Inventory and Planning
关键词
遥感判读
实地验证
正判率
remote sensing interpretation
field investigation and verification
correct interpretation rate
