Article Preview
Top1 Introduction
More and more countries have launched a series of remote sensing satellites, and satellite remote sensing technology has developed rapidly. These satellites include land satellites, meteorological satellites, synthetic aperture radar, etc., covering infrared, visible and multispectral bands. Through the long-term continuous Earth Observation of satellites in different return visit periods, a large number of remote sensing earth observation data with multi spectral, multi temporal, multi spatial and multi type resolution have been accumulated for researchers (J. Zhu et al., 2016). Remote sensing data has the characteristics of huge capacity, fast efficiency, diverse types, rich value and difficult identification. Its characteristics are shown in Figure 1, indicating that remote sensing has entered the era of big data.
Figure 1.
Characteristics of remote sensing big data
These remote sensing data come from different data sources and have multiple feature dimensions, which respect the changes of land surface information in spatial, temporal and spectral dimensions. For example, the current landsat series of satellite sensors have accumulated a large number of surface observation data through more than 40 years of long-term observation (phiri & morgenroth, 2017). Modis sensors provide global observation data for about 17 years (xiong et al., 2017). In addition, more and more remote sensing satellites provide huge remote sensing satellite data. Table 1 lists the features of satellite data commonly used by current researchers. Under the background of the internet of things (iot) era, information technology is changing with each passing day, especially the big data processing and analysis technology of the internet of things, which has achieved good results in the methods of distributed tensor sequence decomposition, visual feature recognition, data-driven and edge server quantification, which has achieved well results in the methods of distributed tensor sequence decomposition, visual feature recognition, data-driven and edge server quantification and has been well applied to industrial intelligence, gradually realizing industrial intelligence and constantly changing all aspects of industrial manufacturing (ren et al., 2020; x. Wang, yang, song, et al., 2020; x. Wang, yang, wang, ren and deen, 2020; xu et al., 2020). Inevitably, it also promotes the rapid development of remote sensing big data technology. Great changes have taken place in the research and application of semantic construction of remote sensing big data.