ABSTRACT
Data merging, introduced in Chapter 11, can improve data availability and reduce cloud impact with the involvement of multiple satellites. However, data merging can rarely remove all cloudy pixels. The main objective of this chapter is to discuss cloud contamination issues in optical remote sensing imageries during Earth observations of land- or water-based objects. It aims to introduce three types of philosophical thinking applied for cloudy pixel removal, in which one type of image reconstruction with the long-term memory effect technique is highlighted. In general, the image processing techniques described in this chapter include:
basic image pre-processing techniques for optical and infrared sensors to reduce the cloud impact with short-term memory effect with synergistic time series opportunities via various algorithms, and
advanced image processing techniques for optical and infrared sensors with the aid of the integrated pattern recognition, big data analytics, and machine learning technique based on long-term memory effect for cloudy pixel reconstruction with the aid of the concept of cognitive science.
