High-resolution images from satellites and airplanes have become ubiquitous in the current digital landscape, and readily available to the public. In recent years, deep learning approaches, and in particular deep convolutional neural networks have revolutionized computer vision. Such deep learning models thrive with an abundance of data, creating enormous potential at the nexus of computer vision and satellite imaging. To this end, we propose to develop DeepSpace-AI, a robust platform to automate processing and object recognition in satellite imaging for a range of applications in marine environmental monitoring and beyond. DeepSpace-AI will serve as a platform for the annotation of satellite image datasets, training of deep learning models, automated object recognition in real-time, and a dashboard for analyzing and interpreting results. Providing a single platform for multiple object recognition tasks will provide unprecedented opportunities for observing behaviors and trends involving combinations of visible phenomena. Finally, the proposed system will automate the import of publicly available aerial and satellite data as it becomes available, enabling pseudo real-time monitoring and rapid detection of global environmental events.