基于形状特征的XBT探头检测与识别研究

为改进国产投弃式温深仪(XBT)的下降速率公式,拟采用基于图像处理的方法测量其下落过程中的运动参数并分析,为此要实现XBT探头的自动提取。为实现视频图像中XBT探头的自动检测识别,提出了一种基于形状特征的目标检测识别算法。首先通过背景差分法去除背景噪声并阈值分割,采用连通区域的标记算子将二值图像标记为互不连通的待识别区域。然后提取连通区域的偏心率、致密性以及标记图构造描述目标的特征向量,通过计算各连通区域的特征评价函数,利用模糊综合评价方法识别XBT探头。最后对研究方法进行了物理实验验证。实验结果表明:用模糊综合评价方法能较好地在复杂背景下实现目标的识别,能够有效地在序列图像中准确的检测识别出XBT探头目标,为XBT运动参数的测量打下了基础。     

基于视觉的水雷目标识别方法研究

水下目标识别一直是目标识别领域的研究热点,为此,以水雷目标识别为需求背景提出了一种基于视觉的水雷目标识别方法。该方法从水雷目标形状的本质特征出发,根据几何矩的物理意义,结合区域特征和边界特征构造了三个适合于水雷目标形状描述子,采用阈值判决的方法,实现水雷目标的识别。实验结果表明,该方法比基于不变矩的方法识别率更高,具有较好的稳定性,尤其适用于水下特定形状的目标识别和目标受到部分遮挡的情况。其中水下图像处理和特征的定义方法对水下目标识别具有指导和借鉴意义。     

基于SVM的海底声纳图像底质识别*

对海底底质声纳图像分割,提取单元特征向量进行主成份分析,选取均值、标准差、对比度、相关系数、能量及同质性作为训练特征向量,优化支持向量机(SVM)参数.利用多分类 SVM 对砂、礁石、泥进行训练,有效、快速地分离出了3种底质类型,测试精度高于80%,表明该方法可有效用于海底底质声纳图像的识别和分类.     

图像自动识别技术在海洋浮游生物分析中的应用

介绍了一种利用数学形态学特征和Gabor纹理特征,结合主成分分析与支持向量机对胶州湾沿岸7种浮游生物的活体图像进行自动识别的方法。实验结果表明,基于主成分分析的降维识别模式可以提高系统识别性能,其平均识别正确率达78.5%,通过对图像采集、图像处理、特征的选取等方面做进一步的改进和提高,基于计算机数字图像的海洋浮游生物自动识别方法将为海洋生态环境监测提供新的实时、快速、高效检测平台。     

基于深度学习的水下目标声学识别与定位技术研究

为实现自主水下潜器(Autonomous Underwater Vehicle,简称 AUV)的自主目标探测识别与定位任务,以侧扫声呐数据为依据,考虑到扫描式声呐成像的特点,针对金属球类目标,基于 Darknet 框架设计了一种轻量化深度学习目标识别模型,并结合人工特征进行目标特性分析。同时对声呐图像设计了有效的图像增强方法。实验表明：上述目标识别方法在保证目标识别准确率的同时,具有较高的目标识别速率,适于低功耗嵌入式平台部署。     

基于融合GMM与SVM的船载非侵入式负荷识别方法

为保障安全的船载用电环境,实现合理的用电配置与管理,在船载电力布局中引入非侵入式电力负荷监测。提出了一种融合高斯混合模型(GMM)与支持向量机(SVM)的电器类型识别算法。该方法利用暂态事件检测所提取的有效负荷特征,建立具有较好统计分布能力的GMM模型和具有较好泛化能力的SVM模型。对两种算法的概率分布进行融合生成最终识别结果。实验结果表明,相对单独应用SVM模型,本文所用方法在准确率和稳定性方面均有一定程度的提升,且实现复杂度低,具有良好的实用价值。     

二进制粒子群支持向量机算法在SAR图像海面溢油特征选择的应用

对合成孔径雷达(synthetic Aperture Radar,SAR)图像提取得到的几何、灰度、纹理特征共66个特征量,采用封装模式算法进行特征选择,降低特征维度并提高对溢油及疑似溢油样本的识别率。特征选择采用二进制离散粒子群优化(binary particle swarm optimization,BPSO)和支持向量机(support vector machine method,SVM)的封装模式算法(BPSO-SVM)进行,该方法在特征选择的同时可对支持向量机模型中的参数进行优化。论文采用BPSO-SVM算法和序列前向搜索(sequential forward selection,SFS)算法、序列后向搜索(sequential backward selection,SBS)算法与SVM算法相结合特征优化算法(SFS-SVM和SBS-SVM算法)进行实验。并将BPSO-SVM算法、SFS-SVM算法、SBS-SVM算法和直接使用SVM算法的分类识别结果进行比较。实验结果表明,BPSO-SVM算法在SAR图像上溢油特征量筛选与识别效率方面行之有效。     

基于弹性网回归的水下目标无监督特征选择算法

针对水声数据的特征中含有大量冗余、不相关和噪声特征，导致水声目标识别正确率降低的问题，提出了一种新的水下目标特征选择方法——基于弹性网回归的无监督特征选择算法(Unsupervised fea- ture selection algorithm based on elastic-net regression，UFSER)。 该算法利用谱回归得到高维水声数据和其低维表示之间的回归系数矩阵，并且在回归框架中加入弹性网惩罚项优化求解回归系数矩阵；最后，对回归系数矩阵进行稀疏化从而对特征的分类性能进行评价。 使用实测水声数据集和 UCI 声呐数据集进行特征选择和 SVM 分类实验。 实验结果表明：在特征数目分别减少 60.6%和 60%的情况下，分类识别正确率较特征选择前提升了 1.05%和 6.6%。     

傅里叶识别分析在赤潮生物自动识别中的应用研究

本文研究了赤潮优势种识别技术,并实际运用于厦门大学流式细胞技术的赤潮实时监控系统FCAM.使用傅里叶形状特征描述子分析结合SVM算法作分类计算,通过对2009年4～10月厦门海域6种最常见的赤潮优势种的3000个样本为专门研究,并结合另外2种描述子算法将实验结果的识别精度提高到95.8％,具备较好的代表性.     

面向ARV的视觉辅助水下对接方法研究

针对水下机器人作业过程中的近距离引导水下对接问题,以自治缆控水下机器人(ARV)为研究对象,为实现快速、高精度的水下对接,设计了基于反射光源识别的单目视觉辅助水下对接方法。通过在对接口布置反光带并将其作为目标图像,设计了图像处理和特征点提取的算法,经过图像特征信息的分析处理,优化算法的时间复杂度,提高了特征点提取的准确率和识别效率。最后设计了传统的引导灯方案与反光带方案的对比试验,验证算法的可行性,证明在水下环境下识别反射光源的方法极大提升了位置估计的准确率,同时在计算速度上也有显著提高,弥补了传统水下对接方法中精度不足或计算量大的缺陷,更好地满足了ARV水下对接的需求,能为潜水器实现水下自主对接提供参考依据。     

Semisupervised heterogeneous ensemble for ship target discrimination in synthetic aperture radar images

Ship detection using synthetic aperture radar (SAR) plays an important role in marine applications. The existing methods are capable of quickly obtaining many candidate targets, but numerous non-ship objects may be wrongly detected in complex backgrounds. These non-ship false alarms can be excluded by training discriminators, and the desired accuracy is obtained with enough verified samples. However, the reliable verification of targets in large-scene SAR images still inevitably requires manual interpretation, which is difficult and time consuming. To address this issue, a semisupervised heterogeneous ensemble ship target discrimination method based on a tri-training scheme is proposed to take advantage of the plentiful candidate targets. Specifically, various features commonly used in SAR image target discrimination are extracted, and several acknowledged classification models and their classic variants are investigated. Multiple discriminators are constructed by dividing these features into different groups and pairing them with each model. Then, the performance of all the discriminators is tested, and better discriminators are selected for implementing the semisupervised training process. These strategies enhance the diversity and reliability of the discriminators, and their heterogeneous ensemble makes more correct judgments on candidate targets, which facilitates further positive training. Experimental results demonstrate that the proposed method outperforms traditional tri-training.     

Zero-G Class Underwater Robots: Unrestricted Attitude Control Using Control Moment Gyros

The "Zero-G" is designated as a new class of underwater robot that is capable of unrestricted attitude control. A novel control scheme based on internal actuation using control moment gyros (CMGs) is developed to provide Zero-G class autonomous underwater vehicles (AUVs) with this unique freedom in control. This is implemented in the CMG-actuated Zero-G class internal kinematic underwater robot actuation (IKURA) system that was developed as part of this research. A series of experiments are performed to demonstrate the practical application of CMGs and verify the associated theoretical developments. The ability to actively stabilize the translational dynamics of the robot is assessed and unrestricted attitude control is demonstrated in an experiment that involves vertically pitched diving and surfacing in surge. Finally, potential applications for Zero-G class AUVs are discussed.     

LIC color texture enhancement algorithm for ocean vector field data based on HSV color mapping and cumulative distribution function

Texture-based visualization method is a common method in the visualization of vector field data. Aiming at adding color mapping to the texture of ocean vector field and solving the ambiguity of vector direction in texture image, a new color texture enhancement algorithm based on the Line Integral Convolution (LIC) for the vector field data is proposed, which combines the HSV color mapping and cumulative distribution function calculation of vector field data. This algorithm can be summarized as follows: firstly, the vector field data is convoluted twice by line integration to get the gray texture image. Secondly, the method of mapping vector data to each component of the HSV color space is established. And then, the vector field data is mapped into HSV color space and converted from HSV to RGB values to get the color image. Thirdly, the cumulative distribution function of the RGB color components of the gray texture image and the color image is constructed to enhance the gray texture and RGB color values. Finally, both the gray texture image and the color image are fused to get the color texture. The experimental results show that the proposed LIC color texture enhancement algorithm is capable of generating a better display of vector field data. Furthermore, the ambiguity of vector direction in the texture images is solved and the direction information of the vector field is expressed more accurately.     

基于换位原理的对手建模模型

对抗性多机器人系统是 1个典型的多智能体系统 (MAS)。在 MAS中 ,建立对手模型 ,并进行意图和规划识别是对抗活动的基本前提。BDI是描述思维的有效手段 ,可用来建立通用的对手思维状态模型。本文在 BDI模型的基础上结合换位原理 ,提出了 1种对抗性多机器人系统中的对手建模模型 ,并将此方法应用到足球机器人系统 ,获得了较好的效果。     

粤西海岛海岸带典型地类SPOT-5影像解译标志

基于SPOT-5假彩色合成影像,在现场踏勘的基础上,对粤西海岛海岸带地物进行了全面的分析,总结了地物的遥感影像特征,建立了耕地、园地、林地、居民地及工矿仓储用地、水域和滩涂6类典型地物的遥感解译标志.采用典型样区与样点相结合的方法,对基于该解译标志提取的结果进行了踏勘验证,经检验,解译精度在90%以上.粤西海岛海岸带卫...     

基于频域典型相关分析的水下目标回波识别

水下目标回波的特征提取与分类识别是当前主动声纳关键技术之一。采用基于回波频域特性的典型相关分析算法(CCA:Canonical Correlation Analysis)提取回波的特征,这些特征集中体现了不同目标回波的综合相关特性。设计合适的支持向量机分类器,并获得识别结果。利用这一方法对湖试中的不同目标回波进行分类识别,分析了不同接收信噪比条件下的性能,获得了理想的结果。     

Survey of underwater robot positioning navigation

Underwater robot positioning and navigation achieve autonomous underwater robot movement based on the premise that positioning obtains the coordinates of the relative position of the underwater robot using a sensor, and navigation yields a known location to the destination path planning. Due to the location of an underwater robot and the complex and changing environment in which it operates, it is difficult to achieve precise positioning using the traditional positioning method. This paper systematically analyzes and summarizes several typical localization and navigation methods of underwater robots, such as multisensor information fusion technology, underwater acoustic localization and navigation methods, GPS buoy, underwater vision, SLAM and coordinate localization and navigation of multiple underwater robots. Multisensory information fusion technology integrates the advantages of the above methods, enhances the system stability and robustness, overcomes the disadvantages of traditional positioning and navigation, and enables the autonomous navigation and positioning of underwater robots. Underwater acoustics enable flexible and convenient positioning, whereas GPS can achieve high-precision and high-positioning navigation information, and visual positioning effectively overcomes the problem of error accumulation. Multirobot cooperative positioning resolves the problem of positioning failure caused by the collapse of a single system and completes complex tasks that cannot be completed by a single robot, thus enhancing the stability and robustness of the system. This paper systematically describes the realization of these methods, presents an actual analysis of their respective advantages and problems, and discusses the development of the field of research prospects and application prospects.     

基于海色融合数据的东海黑潮锋面检测研究

海洋锋是典型的海洋中尺度现象之一。目前卫星遥感主要利用海表温度数据分析海洋锋,但由于西北太平洋海域夏季海表温度的趋同特性,不能进行有效的锋面监测;而不同水团所具有的生物光学特性往往是不同的,且不具有太阳辐射引起的显著性季节变化,因此海色资料也成为检测海洋锋的有效数据源。文中以东海黑潮为例,详细说明了基于叶绿素a浓度融合数据,采用梯度法进行海洋锋面检测的过程,通过比较不同季节不同梯度阈值得到的东海黑潮锋结果,从保持锋面的完整性及对零碎锋区的剔除效应方面,选取了不同季节较优的梯度阈值。总体来说,文中检测出的东海黑潮区域海色锋与海流黑潮强流区较吻合,12月至4月东海黑潮海色锋检测结果不如海温锋,而5-11月东海黑潮海色锋检测结果优于海温锋,特别是台湾以东黑潮区域,不论什么季节海温锋都没有体现,而海色锋始终很明显。利用文中提出的海洋锋检测算法、分析方法及选择的梯度阈值可以有效地检测东海黑潮区域的海洋锋面,结合海色锋和海温锋,可以监测分析东海黑潮强流区的时空变化。     

一种基于四元数的彩色遥感影像水域提取算法

通过研究四元数在彩色影像上应用的优势,提出了一种基于四元数的彩色遥感影像水域提取的算法。实验表明,该算法具有明显的优越性和适应性,对卫星影像海岸水沿线的提取具有指导意义。     

An Integrated Hydrodynamics and Control Model of A Tethered Underwater Robot

An integrated hydrodynamics and control model to simulate tethered underwater robot system is proposed. The governing equation of the umbilical cable is based on a finite difference method, the hydrodynamic behaviors of the underwater robot are described by the six-degrees-of-freedom equations of motion for submarine simulations, and a controller based on the fuzzy sliding mode control (FSMC) algorithm is also incorporated. Fluid motion around the main body of moving robot with running control ducted propellers is governed by the Navier–Stokes equations and these nonlinear differential equations are solved numerically via computational fluid dynamics (CFD) technique. The hydrodynamics and control behaviors of the tethered underwater robot under certain designated trajectory and attitude control manipulation are then investigated based on the established hydrodynamics and control model. The results indicate that satisfactory control effect can be achieved and hydrodynamic behavior under the control operation can be observed with the model; much kinematic and dynamic information about tethered underwater robot system can be forecasted, including translational and angular motions of the robot, hydrodynamic loading on the robot, manipulation actions produced by the control propellers, the kinematic and dynamic behaviors of the umbilical cable. Since these hydrodynamic effects are fed into the proposed coupled model, the mutual hydrodynamic influences of different portions of the robot system as well as the hydrological factors of the undersea environment for the robot operation are incorporated in the model.