量子衍生自适应协同演化的水下声纳图像目标检测关键技术研究

项目名称： 量子衍生自适应协同演化的水下声纳图像目标检测关键技术研究

项目编号： No.41306086

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 王兴梅

作者单位： 哈尔滨工程大学

项目金额： 26万元

中文摘要： 随着水下目标识别技术在海洋开发领域的应用越来越广泛，其关键环节的水下声纳图像目标检测技术研究，已成为军民领域具有重要理论价值和现实意义的课题。本课题在前期研究工作基础上，通过量子衍生自适应协同演化对水下声纳图像目标检测关键技术进行研究。课题开展基于纹理信息描述指标的流形距离二阶段核聚类方法研究，构建水下声纳图像聚类模型，实现精确的初始分类种群；在该模型下开展量子衍生混合蛙跳自适应协同演化检测分析方法研究，以获得高精度的水下目标检测演化结果；开展区域分布密度综合指标的检测算法有效性评价模型的研究，提高检测有效性评价工作的准确度。本课题通过以上研究，期望实现对水下声纳图像这一复杂数据形式进行高效精确的目标检测，并通过量子衍生自适应协同演化为水下声纳图像目标检测研究提供新的思路和探索。

中文关键词： 量子衍生；混合蛙跳；自适应协同演化；水下声纳图像；目标检测

英文摘要： With wide application of underwater objects recognition technologies in ocean development, underwater sonar image objects detection which is its key link has become a project with important theoretical value and practical significance in military and civil field. The key technologies of underwater sonar image objects detection through quantum-inspired self-adaptive co-evolution in this project is conducted on basis of previous work. Manifold distance of two-phase kernel clustering method based on the texture information description indexes will be proposed to construct underwater sonar image clustering model, accurate initial classification of populations will be achieved. Then the detection analysis method of the self-adaptive co-evolution based on the quantum-inspired shuffled frog leaping will be presented under this model, which will obtain high precision evolution results of underwater objects detection. At the same time, the validity evaluation model of detection method will be researched by the comprehensive indexes of density of regional distribution, the nicety of detection validity evaluation will be improved. All of research contents of this project are developed to do the efficient and accurate objects detection on underwater sonar image as complex data form, and supply new thought and exploration fo

英文关键词： Quantum-inspired；Self-adaptive Co-evolution；Shuffled frog leaping；Underwater Sonar Image；Objects Detection