D.M. Lane

The AMADEUS dextrous subsea hand: design, modeling, and sensor processing

This paper describes the mechanical design, finger modeling, and sensor signal processing for a dextrous subsea robot hand incorporating force and slip contact sensing. The hand uses a fluid-filled tentacle for each finger, which has inherent passive compliance, and no moving parts. Force sensing uses strain gauges mounted in the fingertip, potted within a silicon elastomer. Slip sensing uses a piezoelectric strip to detect vibration, embedded 1 mm below the elastomer surface. Static models of finger motion are presented and validated based on bending moments and hydraulic pressure. The design of a stochastic estimator is also described for sensor fusion of contact force magnitude and direction data, obtained using redundant strain gauges in the fingertip. Finally, linear dynamic models of the finger dynamics in contact with a rigid surface are obtained using least squares and recursive least squares parameter estimation, as a precursor to closed-loop force control during grasping.

Texture analysis for seabed classification: co-occurrence matrices vs. self-organizing maps

Considers two well-known pattern recognition techniques using texture analysis. The first is the co-occurrence matrix method which relies on statistics and the second is the Kohonen map which comes from the artificial neural networks domain. Both methods are used as feature extraction methods. The extracted feature vectors are fed to a second Kohonen map used as classifier. The authors report briefly some results of their experimental assessment of the merit of each technique when applied to the problem of classifying the seabed from sequences of real images.

Superellipse Fitting for the Recovery and Classification of Mine-Like Shapes in Sidescan Sonar Images

Mine-like object classification from sidescan sonar images is of great interest for mine counter measure (MCM) operations. Because the shadow cast by an object is often the most distinct feature of a sidescan image, a standard procedure is to perform classification based on features extracted from the shadow. The classification can then be performed by extracting features from the shadow and comparing this to training data to determine the object. In this paper, a superellipse fitting approach to classifying mine-like objects in sidescan sonar images is presented. Superellipses provide a compact and efficient way of representing different mine-like shapes. Through variation of a simple parameter of the superellipse function different shapes such as ellipses, rhomboids, and rectangles can be easily generated. This paper proposes a classification of the shape based directly on a parameter of the superellipse known as the squareness parameter. The first step in this procedure extracts the contour of the shadow given by an unsupervised Markovian segmentation algorithm. Afterwards, a superellipse is fitted by minimizing the Euclidean distance between points on the shadow contour and the superellipse. As the term being minimized is nonlinear, a closed-form solution is not available. Hence, the parameters of the superellipse are estimated by the Nelder-Mead simplex technique. The method was then applied to sidescan data to assess its ability to recover and classify objects. This resulted in a recovery rate of 70% (34 of the 48 mine-like objects) and a classification rate of better than 80% (39 of the 48 mine-like objects).

Real-time automatic sea-floor change detection from video

It is often the case that only sparse sequences of videos from scientific underwater surveys actually contain important information for the expert. A system automatically detecting those critical parts, particularly during the post-mission tape analysis, would alleviate the expert work load and improve data exploitation. The authors present a novel set of algorithms to detect in real time significant context changes in benthic videos. The detectors presented rely on an unsupervised image classification scheme: the time changes in the image contents are analyzed in the feature space. The algorithms are explained in detail, and experimental results with real underwater images reported. Various issues related to the complexity of the problem of automatically analysing underwater videos are also discussed.

Superellipse fitting for the classification of mine-like shapes in side-scan sonar images

The problem of mine-like object classification is of great interest for military applications. A standard procedure is to perform classification based on features extracted from the shadow of the image. However the classification depends up to a certain extent on the accuracy of the features, Alternative approaches such as template matching or working directly on the image have also been studied. However this may not be feasible as they are computationally expensive. In this paper an original method using a superellipse detection procedure to classify mine-like objects in side-scan sonar images is presented. Superellipses provide a compact and efficient way of representing different mine-like shapes. By simply varying the squareness of the function different shapes such as spheres, rhomboids and rectangles can be easily generated. Hence we propose a classification of the shape based on the squareness parameter. The first step in this procedure extracts the contour of the shadow given by an unsupervised Markovian segmentation algorithm. Afterwards a superellipse is fitted by minimising an appropriate metric. As the term being minimised is non-linear a closed form solution is not available. Hence the parameters of the superellipse are estimated by the Nelder-Mead simplex technique. The method has been tested and assessed on real side-scan sonar images, providing satisfactory results. We conclude this work discussing the feasibility of the superellipse fitting model for mine classification and other applications such as pipe modelling. Also further extensions of this work are outlined.

Aspects of the design and development of a subsea dextrous grasping system

This paper considers aspects of a 3 finger dextrous grasping system. Its aim is to provide low force grasping and manipulation of compliant or rigid objects in the ocean up to 150 mm diameter and 5 kg mass. The gripper fingers are constructed using a robust flexible-motion generation mechanism that allows passively compliant, omni-directional finger movements with no joints. The fingertips are equipped with a tactile sensor measuring magnitude and direction of applied force, independent of ambient pressure and temperature. Supervisory control will be used as an operator assistant, to achieve automated grasping with error recovery. Results are described for open loop manipulation and assembly and for calibration of the force sensor. The architecture of the automated grasp planning subsystem is discussed, with examples of its operation.<<ETX>>

An expectation-maximization framework for the estimation of bathymetry from side-scan sonar images

In this paper a new procedure for the computation of seabed altitude information from side-scan sonar data is presented. Although side-scan sensors do not provide direct measures of seabed elevation, their images are directly related to seabed topography. Using a mathematical model for the sonar ensonification process, approximations to the seabed characteristics can be inferred from the sonar image. The problem is however severely under-constrained, in the sense that not all the parameters involved in the image formation process can be directly determined from the side-scan image. To overcome this difficulty, we propose the utilization of a multiresolution expectation-maximization framework to select the most probable parameters from the solution space. At every iteration the estimated solution is used to simulate a side-scan image of the observed scene, which is then be compared to the side-scan image actually observed; solution parameters are then refined using gradient-descent optimization. The process is repeated until convergence is achieved.

Bathymetric side-scan backscatter map restoration based on data fusion

Backscatter image mosaics frequently suffer from noticeable visual artifacts and missing regions, which reduce the performance of subsequent feature extraction or classification processes. In this paper, a new fusion-based method for the restoration of bathymetric side-scan backscatter images is presented. First, the values of missing backscatter pixels are estimated using contextual information provided by the underlying bathymetry mesh. As a final step, a histogram regularization process is performed to remove intensity artifacts caused by inaccurate compensation of sensor gain. Restoration results are presented for Simrad EM12S images.