土地利用扫描图件的注记自动识别与入库研究

Identification of Annotation in Scanned Landuse Map and It's Importation to Database

土地利用基础图件建库是国土资源信息化和“数字国土”的重要组成部分。在传统建库过程中,图件上的大量注记主要是人工识别、手工录入到数据库中,影响了数据库建库的进度,并使数据库质量难以保障。为此提出了扫描图件数字注记自动识别与属性自动入库技术流程和方法,设计开发了较为实用的软件。在包头市土地利用详查数据库建库中的应用表明,上述技术方法合理,注记自动识别精度在87%以上,属性的自动入库准确率超过90%,具有较高的实用价值。

Spatial database construction of landuse maps plays an important role in our national land resources information management. Identifying the annotation and importing the attributes to database is a manual work in the traditional landuse database constructing, which costs a lot of time and brings low quality, In this paper, the approach of identifying the numeric annotation in scanned landuse map and importing the attributes to database automatically is discussed, and the practical software is developed. Applied in the landuse database constructing in Baotou, this methods is proved available and efficient.