This work presents a dataset with surveillance imagery over the sea captured by a small size UAV. This dataset presents object examples ranging from cargo ships, small boats, life rafts to hydrocarbon slick. This dataset presents scenarios that range from the observation of cargo ships, small boats and life rafts, to a hydrocarbon slick. The video sequences were captured using different types of cameras, at different heights and different perspectives. The dataset also contains thousands of labels with positions of objects of interest. This was only possible to achieve with the labeling tool also described in this work. Additionally, using standard evaluation frameworks, we establish a baseline of results using algorithms developed by the authors which are better adapted to the maritime environment.

Autonomous systems can help firefighting operations by detecting and locating the fire spot from surveillance images and videos. Similar to many other areas of computer vision, Convolutional Neural Networks (CNNs) have achieved state-of-the-art results for fire and smoke detection and segmentation. In practice, input images to a CNN are usually downsized to fit into the network to avoid computational complexities and restricted memory problems. Although in many applications downsizing is not an issue, in the early phases of fire ignitions downsizing may eliminate the fire regions since the incident regions are small. In this paper, we propose a novel method to segment fire and smoke regions in high resolution images based on a multi-resolution iterative quad-tree search algorithm , which manages the application of classification and segmentation CNNs to focus the attention on informative parts of the image. The proposed method is more computationally efficient compared to processing the whole high resolution input, and contains parameters that can be tuned based on the needed scale precision. The results show that the proposed method is capable of detecting and segmenting fire and smoke with higher accuracy and is useful for segmenting small regions of incident in high resolution aerial images in a computationally efficient way.

Fish protein hydrolysates (FPHs) can be obtained from substrates such as fish muscle, skin, and wastes and assign value to these fish by-products. Proteolytic enzymes catalyze the hydrolysis of these fish substrates' peptide bonds resulting in smaller peptides that present several bioactive properties. Hydrolysates' bioactive properties are a function of the fish species used as the substrate, the enzyme selectivity or specificity, pH and temperature applied in the reaction, etc. Furthermore, many pre-treatment methods are being applied to fish protein substrates to improve their enzyme susceptibility and increase the number of smaller bioactive peptides. This review addresses the production of FPHs and the main bioactive properties evaluated recently in the literature and emphasizes the substrate treatments by high-pressure processing, microwave, ultrasound, and thermal treatments to achieve better bioactivity making essential amino acids more available in peptides. The bioactive properties most found in FPHs were antioxidants, antimicrobials, anticancer, and antihypertensive. These bioactivities may vary depending on the conditions of hydrolysis, fish species, and fractionation and isolation of specific peptides.New technologies for the treatment of by-products can reduce process losses and achieve better results by cleavage of proteins. Conversely, encapsulation and film utilization can improve bioactivity, bioavailability, and controlled release when applied to foods, resulting in improved health.

The SEAGULL project aims at the development of intelligent systems to support maritime situation awareness based on unmanned aerial vehicles. It proposes to create an intelligent maritime surveillance system by equipping unmanned aerial vehicles (UAVs) with different types of optical sensors. Optical sensors such as cameras (visible, infrared, multi and hyper spectral) can contribute significantly to the generation of situational awareness of maritime events such as (i) detection and georeferencing of oil spills or hazardous and noxious substances; (ii) tracking systems (e.g. vessels, shipwrecked, lifeboat, debris, etc.); (iii) recognizing behavioral patterns (e.g. vessels rendezvous, high-speed vessels, atypical patterns of navigation, etc.); and (iv) monitoring parameters and indicators of good environmental status. On-board transponders will be used for collision detection and avoidance mechanism (sense and avoid). This paper describes the core of the research and development work done during the first 2 years of the project with particular emphasis on the following topics: system architecture, automatic detection of sea vessels by vision sensors and custom designed computer vision algorithms; and a sense and avoid system developed in the theoretical framework of zero-sum differential games.

The objective of this study was to estimate genetic parameters for survival and weight of Nile tilapia (Oreochromis niloticus), farmed in cages and ponds in Brazil, and to predict genetic gain under different scenarios. Survival was recorded as a binary response (dead or alive), during harvest time in the 2008 grow-out period. Genetic parameters were estimated using a Bayesian mixed linear-threshold animal model via Gibbs sampling. The breeding population consisted of 2,912 individual fish, which were analyzed together with the pedigree of 5,394 fish. The heritabilities estimates, with 95% posterior credible intervals, for tagging weight, harvest weight and survival were 0.17 (0.09-0.27), 0.21 (0.12-0.32) and 0.32 (0.22-0.44), respectively. Credible intervals show a 95% probability that the true genetic correlations were in a favourable direction. The selection for weight has a positive impact on survival. Estimated genetic gain was high when selecting for harvest weight (5.07%), and indirect gain for tagging weight (2.17%) and survival (2.03%) were also considerable.