Designing accurate and automatic multi-target detection is a challenging problem for autonomous vehicles. To address this problem, we propose a late multi-modal fusion framework in this paper. The framework provides complimentary information from RGB and thermal infrared cameras in order to improve the detection performance. For this purpose, it first employs RetinaNet as a dense simple deep model for each input image separately to extract possible candidate proposals which likely contain the targets of interest. Then, all proposals are generated by concatenating the obtained proposals from two modalities. Finally, redundant proposals are removed by Non-Maximum Suppression (NMS). We evaluate the proposed framework on a real marine dataset which is collected by a sensor system onboard a vessel in the Finnish archipelago. This system is used for developing autonomous vessels, and records data in a range of operation and climatic conditions. The experimental results show that our late fusion framework can get more detection accuracy compared with middle fusion and uni-modal frameworks.