论文部分内容阅读
在五华区范围内利用遥感影像人工目视判读463个小班,判读面积2743.83hm~2。以二类调查小班区划数据对目视判读数据进行实地验证,结果显示:判读地类中建设用地、耕地、水域的判读精度最高,个数正判率、面积正判率均达100%;其次是纯林,个数正判率为94.84%,面积正判率为94.73%。造成遥感判读与区划调查差异的原因主要有卫片、人为和自然因素三方面。遥感判读各判读地类的小班个数正判率、面积正判率均≥90%,已满足森林资源二类调查工作中所要求的精度。利用该方法准确勾划出地类界线可取代人工手绘小班边界的传统调查方法,节省大量的人力、物力和财力。
In the area of Wuhua, 463 small classes were visually identified by using remote sensing images, and the interpretation area was 2743.83hm ~ 2. The results showed that the accuracy of construction land, cultivated land and water area of interpretation land was the highest, the number of positive judgment rate and the area correct judgment rate reached 100%. Secondly, Is a pure forest, the number of positive judgment rate was 94.84%, the area of correct judgment rate was 94.73%. The main reasons for the discrepancy between remote sensing interpretation and zoning survey are patchwork, man-made and natural factors. The accuracy of judging the number of small classes and the correct rate of area for each interpretation class by remote sensing was ≥90%, which met the precision required in the second type of investigation of forest resources. Using this method, we can accurately delineate the traditional survey methods that the boundary of land type can replace the boundary of hand-painted small classes, saving a lot of manpower, material and financial resources.