Recirculating aquaculture systems (RASs) are intensive aquaculture facilities models that depend on diverse water treatment equipment to maintain good water quality and produce safe and healthy high-quality aquatic products. This article combines the main farming-mode of water purification recirculating processes with recent cultivation-mode scientific research and the current development of the recirculating aquaculture industry. Harmful substances are present in aquaculture wastewater due to large quantities of residual particulate matter such as residual feed, faeces and small suspended solid particles, as well as ammonia, nitrite, bacteria and carbon dioxide (CO 2 ), in the water. These harmful substances seriously affect the quality of aquatic products, so water treatment equipment is needed to remove these substances, add oxygen (O 2 ) to the water and adjust the temperature of the water to ensure a high-quality environment for fish survival. This article reviews the equipment for physical filtration (e.g. solid-liquid separation equipment, microscreen drum filter and foam fractionator) that could remove suspended solids during the water treatment of RASs and the equipment for biological filtration (e.g. fluidized sand biofilter (FSB), moving-bed biofilm reactor (MBBR) and rotating biological contactor (RBC)) that could remove ammonia nitrogen, nitrite and other hazardous substances from wastewater, as well as equipment for water disinfection and sterilization, O 2 addition, CO 2 removal and temperature control. Comprehensive analysis and discussion of water treatment efficiency are provided for reference to create efficient high-end recirculation aquaculture models and increase the precision and intelligence degree of recirculating water treatment technologies in the future.

Despite the recent success of stereo matching with convolutional neural networks (CNNs), it remains arduous to generalize a pre-trained deep stereo model to a novel domain. A major difficulty is to collect accurate groundtruth disparities for stereo pairs in the target domain. In this work, we propose a self-adaptation approach for CNN training, utilizing both synthetic training data (with groundtruth disparities) and stereo pairs in the new domain (without ground-truths). Our method is driven by two empirical observations. By feeding real stereo pairs of different domains to stereo models pre-trained with synthetic data, we see that: i) a pre-trained model does not generalize well to the new domain, producing artifacts at boundaries and illposed regions; however, ii) feeding an up-sampled stereo pair leads to a disparity map with extra details. To avoid i) while exploiting ii), we formulate an iterative optimization problem with graph Laplacian regularization. At each iteration, the CNN adapts itself better to the new domain: we let the CNN learn its own higher-resolution output; at the meanwhile, a graph Laplacian regularization is imposed to discriminatively keep the desired edges while smoothing out the artifacts. We demonstrate the effectiveness of our method in two domains: daily scenes collected by smartphone cameras, and street views captured in a driving car.

No abstract

With the developments of dual-lens camera modules, depth information representing the third dimension of the captured scenes becomes available for smartphones. It is estimated by stereo matching algorithms, taking as input the two views captured by dual-lens cameras at slightly different viewpoints. Depth-of-field rendering (also be referred to as synthetic defocus or bokeh) is one of the trending depthbased applications. However, to achieve fast depth estimation on smartphones, the stereo pairs need to be rectified in the first place. In this paper, we propose a cost-effective solution to perform stereo rectification for dual-lens cameras called direct self-rectification, short for DSR 1 . It removes the need of individual offline calibration for every pair of dual-lens cameras. In addition, the proposed solution is robust to the slight movements, e.g., due to collisions, of the dual-lens cameras after fabrication. Different with existing self-rectification approaches, our approach computes the homography in a novel way with zero geometric distortions introduced to the master image. It is achieved by directly minimizing the vertical displacements of corresponding points between the original master image and the transformed slave image. Our method is evaluated on both realistic and synthetic stereo image pairs, and produces superior results compared to the calibrated rectification or other self-rectification approaches.

No abstract

No abstract