摘要
提出了一种从农业专题图中自动获取农业信息的方法。采用计算机视觉和图像距离变换的方法提取专题图中符号的颜色特征和形态特征,根据图例建立符号模板库,然后把符号从扫描专题图中分割出来,最后由符号骨架线的内距值中误差实现符号信息的自动获取。试验结果表明,系统对农业专题图点状符号的正确分割率和正确识别率分别达到97%和93%,线状、面状符号的正确分割率和正确识别率为99%。
In order to speed up the construction of agricultural information system, a method of automatic acquiring agricultural information from thematic maps was presented. The color feature and morphological feature of symbols were extracted using computer vision and image distance transformation. In this system, the symbol template library was built according to legend firstly, and then the symbols were extracted from the thematic map. Finally, the attribute of symbols were automatic acquired by the in-distance error of symbol skeleton. The experimental results showed that the extraction accuracy and recognition accuracy of point symbols achieved 97 % and 93 %, and of linear symbols and area symbols achieved 99 %.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2008年第5期99-103,共5页
Transactions of the Chinese Society for Agricultural Machinery
关键词
专题图
信息
自动获取
Thematic map, Information, Automatic acquisition
