This chapter presents a multiscale modeling system by integrating multi-sensor satellite data merging, image reconstruction, and data fusion together in support of machine learning/data mining to automate continuous Earth observations for better environmental management. It is designed to demonstrate the seamless integration of remote sensing, signal processing, statistical analysis, and machine learning techniques in different nexus of “system of systems engineering,” “data science and engineering,” “spatial intelligence,” and “remote sensing science” to collectively elevate the quality of environmental decision making. This integrated software platform is capable of merging and fusing multiple satellite imageries as well as reconstructing missing pixels to provide continuous environmental monitoring throughout a series of image processing, enhancement, pattern recognition, classification, reconstruction, feature extraction, and machine learning/data mining applications. The key algorithms for integration mainly described in this chapter may include but are not limited to:

Spectral Information Adaptation and Synthesis Scheme (SIASS)

SMart Information Reconstruction (SMIR)

Integrated Data Fusion and Machine Learning (IDFM)

The new software platform to be introduced is called “Cross-mission Data merging with Image Reconstruction and fusion in support of Machine learning” (CDIRM). This enabling technology was applied to a water quality management issue in a case study in this chapter for demonstration.