Satellite Image Compression by Random Forest Optimization Techniques and Performance Comparison Using a Multispectral Image Compression Method
Srikanth Bethu, Sanjana Vasireddy, D. Ushasree, Asrar Ahmed, P. Vara Prasad · Apple Academic Press eBooks · 2023
Multispectral image compression is a current dominating challenge topic in research attention. Satellite communications, radars, and sensing area technologies are continuously monitoring the earth, space, and environment. In competitive world sources, like power consumption, storage is available with more economy. Also, performance capability remains limited. In this process, multispectral image processing techniques and its usage requirements are compulsory. The related geographical information, optical information, disaster monitoring water wells, etc., are monitored by satellite 68 cameras. So, image quality compression, attacks, histogram equalization, machine learning statistical parameters need to be improved. Existing methods are mainly based on matrix-based modeling, discrete wavelet transform techniques segmentation, low-rank tensor decomposition, but they fail to discover the different strip components and have more limitations. Machine learning also can solve the problems of spectral redundancy, subbands removing models. In this research, we are using the natural random forest machine learning model (NRFML). This model compresses and trains the multispectral image for various application usage, at final comparison the parameters like MSE, peak signal-to-noise ratio, NCC, and structural similitude with existing methods conclude that the proposed NRFML is competing with previous methods.