基于深度学习的水面无人艇目标检测算法综述

Review on Object Detection Algorithms for Unmanned Surface Vehicles Based on Deep Learning

摘要:

随着人工智能的发展,水面无人艇可代替人工进行危险任务作业,目标检测算法是其完成自主探测的核心技术。深度学习技术克服了手工特征精度低、通用性差等局限性,已成为图像处理的主流方法。首先对当前基于深度学习的目标检测算法的发展现状进行了全面的总结,对算法分类进行了详细的定义,并指出了不同类型算法的优缺点及适用场景。然后调研了无人艇水面目标检测技术的研究现状,指出了各类深度学习工作的贡献、优势和局限性。最后总结了面向水面无人艇的深度学习目标检测算法中亟需解决的关键科学问题,并对可行的方案以及该应用研究领域的未来发展做了进一步的展望。

Abstract:

With the development of artificial intelligence, unmanned surface vehicles can replace manual operations for dangerous tasks, and object detection algorithms are the core technology for autonomous detection. Deep learning technology overcomes the limitations of low accuracy and poor versatility of manual features, and has become the mainstream method of image processing. Firstly, a comprehensive summary of the current development status of deep learning-based object detection algorithms is carried out, the classification of algorithms is defined in detail, and the advantages, disadvantages and applicable scenarios of different types of algorithms are pointed out. Then, the research status of unmanned surface vehicles object detection technology is investigated, and the contributions, advantages and limitations of various deep learning works are pointed out. Finally, the key scientific problems that need to be solved urgently in the deep learning object detection algorithm for unmanned surface vehicles. Meanwhile, the feasible solutions and the future development of this application research field are further prospected.