Time series classification for new-generation Earth observation satellites

该项目旨在开发卫星图像的时间序列分类方法，以生成准确的时间土地覆盖图。最新一代卫星刚刚开始频繁、完全、高分辨率地对地球成像，而且对最终用户免费--这是一个前所未有的机会，可以监测我们星球系统的流量。然而，时间序列分类技术不能按比例处理如此丰富的数据。该项目预计其时间序列技术将适用于农业规划、防火和灾害测绘，并预计对地球观测方案的重大投资可以产生更大的价值。

This project aims to develop time series classification methods for satellite images, to produce accurate temporal land-cover maps. Latest generation satellites have just begun imaging Earth frequently, completely, in high-resolution, and at no charge to end-users – an unprecedented opportunity to monitor the flux of our planet's systems. However, time series classification techniques do not scale to handle such wealth of data. The project anticipates its time series technologies will be applicable in agriculture planning, fire prevention, and disaster mapping, and that substantially greater value can be derived from significant investments into Earth Observation programmes.