Transmission Binary Mapping Algorithm with Deep Learning for Underwater Scene Restoration

Deep learning based image resolution reconstruction has gained popularity recently. Using sensors, underwater pictures are captured. Due to the water's exponential attenuation of the transmitted light, underwater photographs have poor visibility, including low contrast and color blurring. The identification of targets and further evaluation are made more difficult by the underwater images' degradation. This paper offers a hybrid underwater image processing method called deep learning based transmission binary mapping algorithm that combines a color balance algorithm with an underwater picture restoration technique to solve these issues. This research work proposes an enhanced picture super-resolution reconstruction approach. It is based on transmission based binary mapping algorithm for enhancement in the efficiency of machine learning concepts. To calculate transmission maps, RGB attenuation for each channels are decomposed. These suggested approach can efficiently determine light at the background and combined transmission binary mapping at underwater images, which can then be used to recover these image. In addition, color attenuation balancing technique helps to provide upgraded scenes that seems realistic. Finally, thorough testing are put into place to evaluate how well our method works. The suggested method surpasses a number of popular underwater image enhancement and restoration techniques, according to tests on photographs taken in various underwater situations.