复杂水域主动式智能感知方法研究

对复杂水域进行感知是实现智能化海事监管系统的必要条件。传统的被动式监控感知系统一般基于固定传感器，其成像范围有限，数据分辨率低。针对目前主动式感知系统存在的传感器最佳成像参数实时计算、内外参数实时标定以及多传感器协作等难题，本项目提出基于Cell的主动式感知方法。首先研究基于综合性约束条件的复杂水域最优Cell分割问题，将感知水域分割成一个个Cell单元。在此基础上，研究通过离线的方法计算每个传感器单元针对每个Cell单元的最优成像参数，进而对每个传感器单元的内外参数进行精确标定。通过引入Cell概念，系统采用离线的方法解决传感器最优成像参数在线实时计算以及内外参数在线实时标定等难题。此外，通过Cell单元与传感器单元的状态关系，研究多传感器交互与协作算法，完成对水上目标的主动式检测、识别与跟踪，实现对复杂水域的主动式智能感知。研究工作将有助于推进内河海事监管系统的智能化与信息化水平。

Intelligent surveillance and sensing of complex waterway areas are crucial to develop intelligent maritime management systems. Traditional surveillance or sensing systems usually work in a passive way by using fixed sensors with limited surveillance areas and low data resolutions. Hence, these systems are not applicable for large-scale waterway areas. Existing active surveillance or sensing systems suffer from the problems of online configuration of sensor parameters, sensor online calibration, and the interaction and cooperation between multiple sensors. This project proposes a cell-based active sensing system for complex waterway areas. In this system, optimized partition of the waterway area with cells is the first problem that we need to address. Based on cell partition, we can offline compute the optimal parameters of each sensor for each cell unit. Afterwards, we can accurately calibrate both intrinsic and extrinsic parameters of the sensors to different cell units offline. By introducing cell partition conception, we can successfully solve the problem of online optimal configuration of sensor parameters and online sensor calibration by using an offline approach. In addition, we can develop a mechanism for interaction and cooperation of multiple sensors based on the states of cell and sensor units. Hence, the developed system can actively detect, recognize, and tracking waterway objects for active and intelligent sensing of complex waterway areas. The researches of this project are expected to help promote the intelligent maritime management systems.