Real-time seabed tracking applications play an important role in underwater systems. A lot of them use computer vision for servoing, positioning, navigation, odometry and simultaneous localisation and mapping. They are mostly based on local image features, therefore feature detection, description and matching are crucial for their efficient operations. The aim of this study was to investigate the most popular feature detection and description algorithms such as SIFT, SURF, FAST, STAR, HAR-RIS, ORB, BRISK and FREAK. Additionally, the image correction technique was presented and image enhancement methods were analysed in order to increase efficiency of image features matching. The matching algorithm was based on the homography matrix and random sample consensus technique. Our results indicate that the combination of the histogram equalisation technique and ORB detector and descriptor enables real-time seabed tracking with sufficient efficiency. ARTICLE HISTORY

Autonomous surface vehicles (ASVs) are a critical part of recent progressive marine technologies. Their development demands the capability of optical systems to understand and interpret the surrounding landscape. This capability plays an important role in the navigation of coastal areas a safe distance from land, which demands sophisticated image segmentation algorithms. For this purpose, some solutions, based on traditional image processing and neural networks, have been introduced. However, the solution of traditional image processing methods requires a set of parameters before execution, while the solution of a neural network demands a large database of labelled images. Our new solution, which avoids these drawbacks, is based on adaptive filtering and progressive segmentation. The adaptive filtering is deployed to suppress weak edges in the image, which is convenient for shoreline detection. Progressive segmentation is devoted to distinguishing the sky and land areas, using a probabilistic clustering model to improve performance. To verify the effectiveness of the proposed method, a set of images acquired from the vehicle’s operative camera were utilised. The results demonstrate that the proposed method performs with high accuracy regardless of distance from land or weather conditions.

The paper presents the second part of the final report on all the experiments with biomimetic autonomous underwater vehicle (BAUV) performed within the confines of the project entitled 'Autonomous underwater vehicles with silent undulating propulsion for underwater ISR', financed by Polish National Center of Research and Development. The report includes experiments on the swimming pool as well as in real conditions, that is, both in a lake and in the sea. The tests presented in this part of the final report were focused on navigation and autonomous operation.

Underwater mines pose extreme danger for ships and submarines. Therefore, navies around the world use mine countermeasure (MCM) units to protect against them. One of the measures used by MCM units is mine hunting, which requires searching for all the mines in a suspicious area. It is generally divided into four stages: detection, classification, identification and disposal. The detection and classification steps are usually performed using a sonar mounted on a ship’s hull or on an underwater vehicle. After retrieving the sonar data, military personnel scan the seabed images to detect targets and classify them as mine-like objects (MLOs) or benign objects. To reduce the technical operator’s workload and decrease post-mission analysis time, computer-aided detection (CAD), computer-aided classification (CAC) and automated target recognition (ATR) algorithms have been introduced. This paper reviews mine detection and classification techniques used in the aforementioned systems. The author considered current and previous generation methods starting with classical image processing, and then machine learning followed by deep learning. This review can facilitate future research to introduce improved mine detection and classification algorithms.

In this paper the attempt to make an analysis of distance measurement using a stereo vision system was presented. Main emphasis was placed on the geometric camera calibration. The classical method based on the specially prepared calibration pattern with known dimensions and position in a certain coordinates system was performed. Finally, the metric information obtained from images was presented.