António José Oliveira

Abstract
Blob features are particularly common in acoustic imagery, as isolated objects (e.g., moorings, mines, rocks) appear as blobs in the acquired images. This work focuses the application of the SIFT, SURF, KAZE and U-SURF feature extraction algorithms for blob feature tracking towards Simultaneous Localization and Mapping applications. We introduce a modified feature extraction and matching pipeline intended to improve feature detection and matching precision, tackling performance deterioration caused by the differences between optical and acoustic imagery. Experimental evaluation was undertaken resorting to datasets collected from a water tank structure.

Abstract
In-situ calibration of marine sensors requires close-range positioning. In turn, localization relative to a given object of interest is necessary. This paper deals with the detection of a vertical cable hanging from a marine observatory implemented by means of a moored buoy. An algorithm composed of sequential image filtering, segmentation and template matching is proposed. Two approaches for generating the cable's acoustic image template are introduced. The performance of the approaches, obtained by comparison with ground-truth measurements, are illustrated over challenging cluttered acoustic images collected in a test tank. The results indicate a performance better than 74% of the best candidate to match the actual cable.

Abstract
Wall and other planar structures are common in environments as harbors, marinas, or dams. In this paper we introduce an algorithm aimed at the identification of these structures through acoustic images retrieved from an imaging sonar, building on an application of the Hough Transform algorithm. We employ a polar-based line parametric model for improved computational efficiency and further adapt the core Hough Transform blocks to the characteristics of acoustic imaging. The developed solution was subjected to experimental tests employing acoustic data acquired in a water tank, from different viewpoints and under different sonar gain configurations.

Abstract
In underwater navigation, sonars are useful sensing devices for operation in confined or structured environments, enabling the detection and identification of underwater environmental features through the acquisition of acoustic images. Nonetheless, in these environments, several problems affect their performance, such as background noise and multiple secondary echoes. In recent years, research has been conducted regarding the application of feature extraction algorithms to underwater acoustic images, with the purpose of achieving a robust solution for the detection and matching of environmental features. However, since these algorithms were originally developed for optical image analysis, conclusions in the literature diverge regarding their suitability to acoustic imaging. This article presents a detailed comparison between the SURF (Speeded-Up Robust Features), ORB (Oriented FAST and Rotated BRIEF), BRISK (Binary Robust Invariant Scalable Keypoints), and SURF-Harris algorithms, based on the performance of their feature detection and description procedures, when applied to acoustic data collected by an autonomous underwater vehicle. Several characteristics of the studied algorithms were taken into account, such as feature point distribution, feature detection accuracy, and feature description robustness. A possible adaptation of feature extraction procedures to acoustic imaging is further explored through the implementation of a feature selection module. The performed comparison has also provided evidence that further development of the current feature description methodologies might be required for underwater acoustic image analysis.

Abstract