Production of Typical Terrain Element Interpretation Sample Data Based on FME
The development of intelligent automatic remote sensing image processing technology is relatively lagging behind. Deep learning technology has significantly improved the effect of image feature extraction,but the number and type of deep learning samples are limited. Based on the remote sensing image interpretation sample data,land cover,and national geographical element result data of the national geographical census and monitoring project,and based on the abundant converters in the FME software function library,the work flow is made to produce typical terrain element interpretation sample data such as buildings,water,roads,vegetation,etc. Typical terrain element interpretation sample data is used to expand the high-quality typical terrain element remote sensing interpreta-tion sample database,support multi-sensor,multi-temporal,multi-regional deep learning training,improve the automatic recognition and extraction accuracy of typical terrain elements,and provide sample data support for solving key technical problems in domestic sat-ellite image processing and typical terrain element information extraction.
typical terrain element interpretation sample datalabel datadeep learningFME
万方数据
维普
