基于多模态感知的水下目标检测应用构想

Research on Application of Underwater Object Detection Based on Multimodal Perception

摘要:

水下目标检测在海洋生物研究、考古探索、军事防御等多领域广泛应用，随着人工智能快速发展，水下目标检测也朝着无人化、智能化发展。深度学习采用神经网络挖掘信息特征，在速度和精度上均表现出优异的性能，成为了计算机视觉技术的主流算法，然而水下环境复杂，将其应用于水下图像目标检测仍存在较大的挑战。水下目标各模态信息互补，特征丰富，有利于目标检测识别，因此结合应用场景调研现有技术，然后设计基于深度学习的多模态水下目标检测系统，同时对比分析了现有关键技术的优缺点，最后对多模态目标检测系统未来发展进行总结与展望，具有重要意义。

Abstract:

Underwater object detection has been widely applied in various fields，such as marine biology research，archaeological exploration，and military defense. With the rapid development of artificial intelligence， underwater object detection has also become unmanned and intelligent. Deep learning uses neural networks to mine information features，demonstrating excellent performance in both speed and accuracy，and has become the mainstream algorithm in computer vision technology. However，in complex underwater environments，there are still significant challenges in applying it to underwater image object detection. The complementary information and rich features of various modalities of underwater targets are beneficial for target detection and recognition. Therefore，this article combines application scenarios to investigate existing technologies，and then designs a multimodal underwater target detection system based on deep learning. At the same time，the advantages and disadvantages of existing core technologies are compared and analyzed. Finally，a summary and outlook on the future development of multimodal object detection systems are carried out，which is of great significance.