Taking Baoding in Hebei Province for example, considering the high resolution WorldView images for data source, and based on eCognition software, this paper uses the "bottom-up" multi-scale segmentation method proposed by object-oriented classification techniques to monitor the changes in land use in Baoding from April to July of 2013. Through the calculation, the correct rate of land use monitoring change map is 91.40%. Findings show that: compared to the previous method of visual interpretation, object-oriented automatic classification method improves the operational efficiency, and avoids the omissions of artificial visual interpretation and ensures the correctness and integrity of the data updating and collection in land use monitoring.%本文以河北省保定市为例，以高分辨率WorldView影像为数据源，基于eCognition软件，采用面向对象分类技术提出的“自下而上”的多尺度分割方法，对保定市市区2013年4月至7月间土地利用变化情况进行监测，经计算，土地利用监测变化图斑核查的正确率为91.40%。研究结论表明：相比于以往的人工目视解译方法，面向对象自动分类方法提高了作业效率，同时可以避免人工目视解译中产生的遗漏，保证了土地利用监测中数据更新采集的正确性以及完整性。
展开▼
