A statistical estimation method for segmentation of sonar range data

In this paper, we describe how to deal with an important sensorial activity that ultrasonic echo-locating systems for mobile robot navigation have often to perform, namely the extraction of straight line segments from range data and the accurate localization of the corresponding planar targets. It is commonplace that range data segmentation starts with using least squares interpolation algorithms for obtaining straight line segments: it is our goal to prove that caution must be called for in order to avoid somewhat misleading results. The case study concerns the use of a linear array formed by three ultrasonic transducers in a 2D specular environment composed of line and point acoustic targets.The segmentation algorithm we propose is subdivided into two functionally distinct modules, namely identification and localization. The identification module is based on a sequential hypothesis testing between alternative hypotheses that explain the sonar range data as originated from line or point targets. With regard to the localization module, we demonstrate that widely used approaches to sensor modeling are, to some extent, deceptively simple: the estimation accuracy for the localization of planar objects may be decreased by the inability of some traditional sonar sensor models to take properly into account the specularity of reflections. A physically based model of acoustic range sensors acting in specular environments allows us to design a localization module which is capable of producing accurate and unbiased estimates of the parameters of a planar geometric feature.The proposed theoretical framework is validated by the results of some experiments carried out with a spatial locating system consisting of a rotating linear array of three ultrasonic transducers.     

新的声纳图像自动分割方法

对水下声纳图像进行目标分割是非常复杂的,它不仅取决于被分割的不同目标,还与海底混响噪声、背景区域有着紧密的联系。通过分析声纳图像的特点,提出了一种新的声纳图像自动分割方法,即利用一种快速的模糊C均值聚类方法来完成初始分割,然后利用初始分割结果对马尔可夫模型的初始参数进行估计,最后,根据马尔可夫理论进行迭代条件估计,得到精确的图像分割结果。最后利用实测数据,验证了此种算法的可行性和有效性。     

基于声纳图像的水平集分割算法研究

针对现有的图像分割方法无法准确地分割声纳图像的问题,提出了一种改进的水平集声纳图像分割方法。介绍了LBF能量模型,借鉴其无重初始化的水平集演化思想。为克服声纳图像中复杂背景带来的负面效应,利用形态学顶帽—底帽变换对声纳图像进行预处理,并在此基础上进行无需初始化的水平集分割。进行仿真对比实验,实验结果显示:与LBF能量模型相比,改进的水平集分割方法更加适应于背景不均匀的声纳图像分割。     

侧扫声纳图像分割中种子区域获取方法

研究了一种基于区域增长的侧扫声纳图像感兴趣区域提取方法。该方法利用边缘检测算子提取侧扫声纳图像的目标影阴区的种子区域,利用阈值分割方法提取目标亮区的种子区域,采用Haralick与Shapiro提出的相似准则进行区域增长,提取感兴趣区域。从处理结果看,该方法对某些类型的侧扫声纳图像的感兴趣区域提取有较好的效果,而且种子区域的引入大大缩短了处理时间。     

Knowledge-based segmentation and labeling of brain structures from MRI images

In this paper, we propose a new knowledge-based method illustrated in the context of segmentation, which labels internal brain structures viewed by magnetic resonance imaging (MRI). In order to improve the accuracy of the labeling, we introduce a fuzzy model of regions of interest (ROI) by analogy with the electrostatic potential distribution, to represent more appropriately the knowledge of distance, shape and relationship of structures. The knowledge is mainly derived from the Talairach stereotaxic atlas. The labeling is achieved by the regionwise labeling using genetic algorithms (GAs), followed by a voxelwise amendment using parallel region growing. The fuzzy model is used both to design the fitness function of GAs, and to guide the region growing. The performance of our proposed method has been quantitatively validated by six indices with respect to manually labeled images.     

结合空间信息的模糊聚类侧扫声纳图像分割

目的 针对侧扫声纳图像具有背景复杂、噪声污染重等特点,提出一种结合空间信息的模糊聚类分割算法(FCM),以提高侧扫声纳图像的分割精度和分割速度。方法 采用结合组合邻域中值滤波的FCM算法,首先选取正十字邻域和斜十字邻域,分别计算两个邻域内的像素灰度中值;然后,结合其中的较小值,引入惩罚项,得到融合灰度信息和空间信息的灰度值;最后,对融合后的灰度图像进行FCM分割。结果 利用该算法对不同尺寸和不同目标的侧扫声纳图像进行分割,并将分割结果与其他5种聚类算法的分割结果进行比较,对比分析每种算法的分割精度和运算时间。结合中值滤波的FCM算法的分割精度和运算时间均优于传统的FCM算法和结合均值滤波的FCM算法,其中结合组合邻域中值滤波的FCM算法的运算速度较快,分割精度略高于结合传统中值滤波的FCM算法。结论 结合组合邻域中值滤波的FCM算法在对侧扫声纳图像进行分割时,具有较强的抗噪性、实时性和较强的边缘保持能力。     

基于知识的车载LiDAR地物自动分类

车载LiDAR已广泛应用于三维数字城市建模、道路信息数据采集等领域。海量点云信息中不同地物目标的自动识别和分类是LiDAR数据后处理的难点之一。根据不同地物目标物理特性、空间拓扑关系及其在点云中的相关特征知识,建立地物分类规则,依据分类知识进行地物自动识别和分类。通过实测数据分类试验,证明该方法可以较好实现建筑物、树木、线杆、行人等不同地物的自动识别和分类。     

Elastic shapes models for improving segmentation of object boundaries in synthetic aperture sonar images

We present a variational framework for naturally incorporating prior shape knowledge in guidance of active contours for boundary extraction in images. This framework is especially suitable for images collected outside the visible spectrum, where boundary estimation is difficult due to low contrast, low resolution, and presence of noise and clutter. Accordingly, we illustrate this approach using the segmentation of various objects in synthetic aperture sonar (SAS) images of underwater terrains. We use elastic shape analysis of planar curves in which the shapes are considered as elements of a quotient space of an infinite dimensional, non-linear Riemannian manifold. Using geodesic paths under the elastic Riemannian metric, one computes sample mean and covariances of training shapes in each classes and derives statistical models for capturing class-specific shape variability. These models are then used as shape priors in a variational setting to solve for Bayesian estimation of desired contours as follows. In traditional active contour models curves are driven towards minimum of an energy composed of image and smoothing terms. We introduce an additional shape term based on shape models of relevant shape classes. The minimization of this total energy, using iterated gradient-based updates of curves, leads to an improved segmentation of object boundaries. This is demonstrated using a number of shape classes in two large SAS image datasets.     

A sonar image segmentation algorithm based on quantum-inspired particle swarm optimization and fuzzy clustering

Neural Computing and Applications - The underwater sonar image has the characteristics of complex background and heavy noise pollution. By using multi-bit quantum system to encode particles, a new...     

Customer data mining for lifestyle segmentation

A good relationship between companies and customers is a crucial factor of competitiveness. Market segmentation is a key issue for companies to develop and maintain loyal relationships with customers as well as to promote the increase of company sales. This paper proposes a method for market segmentation in retailing based on customers’ lifestyle, supported by information extracted from a large transactional database. A set of typical shopping baskets are mined from the database, using a variable clustering algorithm, and these are used to infer customers lifestyle. Customers are assigned to a lifestyle segment based on their purchases history. This study is done in collaboration with an European retailing company.     

Spectral registration of noisy sonar data for underwater 3D mapping

3D mapping is very challenging in the underwater domain, especially due to the lack of high resolution, low noise sensors. A new spectral registration method is presented that can determine the spatial 6 DOF transformation between pairs of very noisy 3D scans with only partial overlap. The approach is hence suited to cope with sonar as the predominant underwater sensor. The spectral registration method is based on Phase Only Matched Filtering (POMF) on non-trivially resampled spectra of the 3D data.     

拖曳线列阵声呐数据录取系统的设计与实现

为了满足拖曳线列阵声呐系统数据传输实时性高、数据量大的要求, 设计了一种基于PCI总线的数据录取系统。硬件方面, 系统采用PCI9054高效率DMA传输技术与外扩大容量FIFO相结合的方案, 实现了高速、连续的声呐数据传输; 软件方面, 系统结合基于双层乒乓结构的内存映射文件和多线程技术实现了数据的实时存储、处理及显示。模拟及现场实验结果表明, 该录取系统能稳定、可靠地实现声呐数据的实时存储、波形回显等功能, 能满足大规模声呐系统数据录取的要求。     

Experiments in curvature-based segmentation of range data

This paper focuses on the experimental evaluation of a range image segmentation system which partitions range data into homogeneous surface patches using estimates of the sign of the mean and Gaussian curvatures. The authors report the results of an extensive testing program aimed at investigating the behavior of important experimental parameters such as the probability of correct classification and the accuracy of curvature estimates, measured over variations of significant segmentation variables. Evaluation methods in computer vision are often unstructured and subjective: this paper contributes a useful example of extensive experimental assessment of surface-based range segmentation     

Map building through pseudo dense scan matching using visual sonar data

This paper presents a novel approach to the vision based grid map building and localization problem that works in a complex indoor environment with a single forward viewing camera. Most existing visual SLAM has been limited to the feature-based method and only a few researchers have proposed visual SLAM methods for building a grid map using a stereo vision system which has not been popular in practical application. In this paper, we estimate the planar depth by applying a simple visual sonar ranging technique to the single camera image and then associating sequential scans through our own pseudo dense adaptive scan matching algorithm reducing the processing time compared to the standard point-to-point correspondence based algorithm and finally produce a grid map. To this end, we construct a Pseudo Dense Scan (PDS) which is an odometry based temporal accumulation of the visual sonar readings emulating omni-directional sensing in order to overcome the sparseness of the visual sonar. Moreover, in order to obtain a much more refined map, we further correct the slight trajectory error incurred in the PDS construction step using Sequential Quadratic Programming (SQP) which is a well-known optimization scheme. Experimental results show that our method can obtain an accurate grid map using a single camera without the need for a high price range sensors or stereo camera.

Classification of functional data: A segmentation approach

We suggest a classification and feature extraction method on functional data where the predictor variables are curves. The method, called functional segment discriminant analysis (FSDA), combines the classical linear discriminant analysis and support vector machine. FSDA is particularly useful for irregular functional data, characterized by spatial heterogeneity and local patterns like spikes. FSDA not only reduces the computation and storage burden by using a fraction of the spectrum, but also identifies important predictors and extracts features. FSDA is highly flexible, easy to incorporate information from other data sources and/or prior knowledge from the investigators. We apply FSDA to two public domain data sets and discuss the understanding developed from the study.     

Image data field for homogeneous region based segmentation

Some innovative methods for image segmentation inspired by physical world are presented in recent years. Aiming to find homogeneous regions and latent semantic information, the paper presents a novel image segmentation method based on image data field. Image data field, developed by simulating the short-range nuclear forces field theory in the physical world, can effectively represent the spatial interactions of neighborhood pixels. Then, the homogeneous regions are characterized by maximum tolerance classes, which induced by homogeneous attraction relation comparing the contributions of potential values in image data field. More specifically, the proposed method mainly focuses on the images with uneven lighting conditions. Compared with the existing relative methods on a variety of images, the experimental results suggest that the presented method is efficient and effective.     

Market segmentation using service levels in data networks

The objective of this paper is to study a communication system based on aM ^(x)/ D/1 queueing system representing a cell-switch network like Asynchronous Transfer Mode (ATM) network. Network structure consists of a single link modeled as a batch arrival markovian queue with non-preemptive head of the line priority service. Network manager (NM) is assumed to be a decision maker at a Management Information System (MIS) department. This paper establishes the incentive compatible pricing which maximizes the net value of the overall corporation, while the delays have to satisfy the Quality of Service (QoS) guarantees. We obtain structural results for the two priority case in the short run. In equilibrium, we find that the network manager maximizes the price spread between the two priority class services. We prove that as the capacity level increases indefinitely, the market is equally divided among the priority classes. In the first part of the paper, we assume that the users do not respond to network manager’s prices. In the second part, we relax this assumption and look at a leader follower game. Users choose their willingness to pay by deciding on how much value they assign to timely transmission of messages after seeing the prices set by the network manager. Our results indicate that unless there is high enough capacity set upex ante, monopoly network provider cannot price discriminate by offering different quality of service via priority classes. This trade-off between ex ante capacity level choice and ex post price discrimination decision is eliminated if the capacity is set high. It is shown in the network literature that best effort services lead to lower quality of service, in general, for a single service. We show that this holds in multiple priority services as well. We prove that when the capacity is also considered as a decision variable, simultaneous capacity and price setting yields the same optimal level with sequential capacity and price choices.     

基于PGA的人体运动捕获数据分割方法

由于人体运动捕获数据的固有非线性,线性方法并不总是能够有效地找到运动捕获数据的内在维度,针对这种情况,提出了基于主测地线分析（PGA）和概率主测地线分析（PPGA）的自动分割方法。这两种方法都将人体运动视为一个有序的姿势序列,并在姿势序列有局部变化处对运动进行分割。基于PGA的分割方法是在运动局部模型的内在维度突然增长处分配一个分割点,基于PPGA的分割方法是在姿势分布发生改变时放置分割点。实验结果表明,该方法都能实现自动分割,且具有较好的分割结果。