In the last two decades, the automatic classification of marine species based on images has been considered as one of the best approaches. In fact, oceans are complex ecosystems, difficult to access, and often the images obtained are of low quality, in particular deep ocean images. In such a scenario, classification using traditional methods becomes tedious. Therefore, the enhancement or pre-processing techniques to the images before applying classification algorithms is an essential task. In this work, we propose an image enhancement and classification pipeline that allows automated processing of images from benthic moving platforms. Further, we provide the performance accuracy of the proposed deep learning methods using MobileNet for different real time scenarios such as incomplete information, overlapping images, concatenated input images. Lastly the accuracy has been tested for gray scale images that reduces the complexity in the model.