Processing of Thermal Satellite Images Using MATLAB

The paper intends to illustrate the usage and combination of different functions designed in a MATLAB-Simulink based GUI to analyse the Infrared images. It combines the use of filters (low pass and high pass), various techniques of deblurring, image registration, image rectification, image stitching, and enhancement. The paper describes the fundamental technical aspects of Digital Image Processing with particular reference to satellite image processing. All the image-processing operations will be grouped into three categories: Image Rectification and Restoration, Enhancement, and Information Extraction. Image Rectification and Restoration deals with the initial processing of raw image data to reduce noise present, while the enhancement procedures are applied to display the image data effectively for its subsequent visual interpretation. It involves techniques such as false colouring and deblurring of the thermal infrared image. A blind deblurring was achieved without the prior knowledge of point spread function using dark channel priors. Data extraction is be done using image segmentation by different methods such as thresholding, morphological segmentation, and region-based segmentation. Multi-class classifiers have been used to classify and label this image data based on different heat regions in the land. This classified data can be used to produce thematic maps of the land cover present in an image. Attempts have also be made to measure the area of each heat region. MATLAB based codes, as well as test results, have been provided throughout the paper. The paper aims to describe the entire process of taking a raw image from the source and processing it to make it interpretable and suitable for data extraction and analysis. A user-friendly MATLAB -Simulink based GUI is also made to do the processing in a smoother, more efficient, and systematic way, allowing more functionality.