A Real-Time FPGA Accelerator Based on Winograd Algorithm for Underwater Object Detection

Real-time object detection is a challenging but crucial task for autonomous underwater vehicles because of the complex underwater imaging environment. Resulted by suspended particles scattering and wavelength-dependent light attenuation, underwater images are always hazy and color-distorted. To overcome the difficulties caused by these problems to underwater object detection, an end-to-end CNN network combined U-Net and MobileNetV3-SSDLite is proposed. Furthermore, the FPGA implementation of various convolution in the proposed network is optimized based on the Winograd algorithm. An efficient upsampling engine is presented, and the FPGA implementation of squeeze-and-excitation module in MobileNetV3 is optimized. The accelerator is implemented on a Zynq XC7Z045 device running at 150 MHz and achieves 23.68 frames per second (fps) and 33.14 fps when using MobileNetV3-Large and MobileNetV3-Small as the feature extractor. Compared to CPU, our accelerator achieves 7.5×–8.7× speedup and 52×–60× energy efficiency.

Monitoring of Turbidity Variation in the Ukai Reservoir, Gujarat, INDIA, during 1993-2018 using Landsat Series of Dataset

Turbidity is one of the important water quality parameters, which is required to understand the eco-hydrological process such as a trophic state of water, soil erosion into the river system, mixing of other water sources, runoff, discharge etc. An algorithm has been developed to estimate the turbidity (in NTU: Nephelometric Turbidity Unit) over inland waters using Red band of optical multispectral dataset. Field measurements were carried out over Ukai reservoir for 27-28th March 2018 for pre monsoon and 27-30th September 2018 for post monsoon seasons, sampling sites ranging from turbid to clear water. Where in situ water leaving reflectance and turbidity were measured. Model was derived between in situ measured turbidity and spectral reflectance of Red band of Landsat series of datasets includes Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+) and Landsat 8 Operational Land Imager (OLI) data from 1993-2018. The model was applied to derive the turbidity maps of Ukai reservoir for pre-monsoon (March, April and May months) season and post monsoon (September, October and November months) seasons. Overall turbidity was in the range of 1.47-25 NTU during the field data collection for both pre and post monsoon seasons. To investigate the results in detail, the reservoir was divided into three parts, i.e. Down (A), Middle (B) and Up Streams (C). The water was relatively clear in the downstream portion with average turbidity less than 5 NTU over the study period. While maximum turbidity was observed in the upstream portion with values more than 20 NTU. In the middle portion, the turbidity values were fluctuating within the range 4-13 NTU with an average value of 6 NTU. These turbidity maps can be used to determine underwater light attenuation that has importance in ecosystem modelling.

Differentiation between normal and tumor mammary glands with depth-resolved attenuation coefficient from optical coherence tomography

Optical coherence tomography (OCT) is a well-established imaging technology for high-resolution, cross-sectional imaging of biological tissues. Imaging processing and light attenuation coefficient estimation allows to further improve the OCT diagnostic capability. In this paper we use a commercial OCT system, Telesto II-1325LR from Thorlabs, and demonstrate its ability to differentiate normal and tumor mammary mouse glands with the OCT attenuation coefficient. Using several OCT images of normal and tumor mammary mouse glands (n=26), a statistical analysis was performed. The attenuation coefficient was calculated in depth, considering a slope of 0.5 mm. The normal glands present a median attenuation coefficient of 0.403 mm-1, comparatively to 0.561 mm-1 obtained for tumor mammary glands. This translates in an attenuation coefficient approximately 39 % higher for tumor mammary glands when compared to normal mammary glands. The OCT attenuation coefficient estimation eliminates the subjective analysis provided by direct visualization of the OCT images.

Deep Learning Approach for Submerged Image Enhancement

Because of underwater pictures application in ocean engineering, ocean research, marine biology, and marine archaeology to name a few, underwater picture enhancement was widely publicized in the last several years. Underwater photos frequently upshot in low contrast, blurred, color distortion, hazy, poor visible images. This is because of light attenuation, absorption, scattering (forward scattering and backward scattering), turbidity, floating particles. As a result, effective underwater picture solution must be developedin order to improve visibility, contrast, and color qualities for greater visual quality and optical attractiveness. Many underwater picture enhancing approaches have been proposed to overcome these challenges; however they all failed to produce accurate results. Hence for this we first undertook a large scale underwater image dataset which is trained by convolution neural network (CNN) and then we have studied and implemented a deep learning approach called very deep super resolution (VDSR) model for improving the color, contrast, and brightness of underwater photos by using different algorithms such as white balance, histogram equalization, and gamma correction respectively. Moreover, our method is compared with the existing method which reveals that our method surpassesthe existing methods Keywords: CNN, gamma correction, histogram equalization, underwater image enhancement, VDSR, white balance

Hatching Under Brownification: DOC-Mediated Changes in Physical, but Not Chemical Properties of Water Affect Hatching Patterns of Cladocera Resting Eggs

Dissolved organic carbon (DOC) is often related to the brownification of water in continental aquatic systems and to changes in the physiology of zooplankton organisms. Zooplankton resting eggs are particularly sensitive to changes in light and chemical characteristics of water, but the physical and chemical effects associated to DOC on dormant stages have never been tested before. Herein we tested how DOC affects hatching rates and time to hatching of Cladocera (Diaphanosoma birgei) resting eggs. In order to analyze the chemical (i.e. toxic) and physical (i.e. light attenuation) effects of DOC on hatching patterns, resting eggs were exposed to different concentrations of DOC in an experimental design which isolated chemical from physical effects. We observed higher hatching rates and lower time to hatching at intermediate DOC concentrations. DOC effects on hatching mainly relied on light attenuation, while chemical effects were likely of minor importance. We conclude that DOC may change Cladocera emergence patterns mainly through light attenuation in the water column.

Blue-Light Transmittance In Sunglasses Over Long-Term Irradiation Within A Solar Simulator

Blue-light transmittance in sunglasses plays an important role for the consumer regarding eye health. Current standards do not have specific requirements for blue-light protection, as in the past. However, the literature has warned about the potential harms of blue light for the eye. The limits imposed in the past state that the average transmission in the 380 nm to 500 nm region should not exceed 1.2X the transmittance in the visible range (380 nm – 780 nm). This work investigates, besides the transmittance of blue light in sunglasses, whether those limits imposed aiming to eye health are respected. Additionally, this study examines the blue-light transmittances pre- and post-artificially aging the material in a solar simulator. Twelve samples of sunglasses were tested for compliance with a former standard and submitted to the aging process up to 2500 h within the solar simulator. The results showed relevant changes in the lenses over time, that is, they considerably lost their blue-light attenuation capabilities. The results suggest that the aging test should be carried out on sunglasses not only for ultraviolet radiation, as required by most standards, but also for blue light. Furthermore, the standards should comprise some constraints concerning the blue-light attenuation.

Inherent optical properties and optical characteristics of dissolved organic and particulate matter in an Arctic fjord (Storfjorden, Svalbard) in early summer

There have been considerable efforts to understand the hydrography of the Storfjorden fjord (Svalbard). A recurring winter polynya with large sea ice production makes it an important region of dense water formation at the scale of the Arctic Ocean. In addition, this fjord is seasonally influenced by freshwater inputs from sea-ice melt and the surrounding islands of the Svalbard archipelago which impacts the hydrography. However, the understanding of factors controlling the optical properties of the waters in Storfjorden are lacking and are crucial for development of more accurate regional bio-optical models. Here, we present results from the first detailed optical field survey of Storfjorden conducted in early summer of 2020. In addition to the expected seasonal contribution from phytoplankton, we find that in early summer waters in Storfjorden are optically complex with a significant contribution from coloured dissolved organic matter (33–64 % of the non-water absorption at 443 nm) despite relatively low CDOM concentrations, and in the nearshore or near seabed from non-algal particles (up to 61 % of the non-water absorption at 550 nm). In surface waters, the spatial variability of light attenuation was mainly controlled by inorganic suspended matter originating from river runoff. A distinct subsurface maximum of light attenuation was largely driven by a subsurface phytoplankton bloom, controlled by stratification resulting from sea-ice melt. Lastly, the cold dense bottom waters of Storfjorden, from winter sea ice production, which periodically overflows into the Fram Strait, was found to contain elevated levels of both non-algal particles and dissolved organic matter, which is likely caused by the dense flows of the nepheloid layer interacting with the sea bed.

Intraoperative Control of Hemoglobin Oxygen Saturation in the Intestinal Wall during Anastomosis Surgery

During surgery for colon cancer, monitoring of the oxygen saturation of hemoglobin in the tissues under study makes it possible to assess the degree of blood supply to the anastomosis areas of the colon. Adequate blood supply in this area is decisive in terms of the consistency of the anastomosis and can significantly reduce the risk leakage of anastomosis. In this work, we propose a new approach to assessing the hemoglobin oxygen saturation based on measuring both the diffuse reflectance and transmittance spectra of the colon wall tissues. The proposed method is based on the use of two fiber-optic tools for irradiation from both sides—the intestinal lumen and the outside of the intestinal wall. The spectra are recorded from the external side. To determine the degree of hemoglobin saturation, two algorithms, both based on the Taylor series expansion of the coefficient of light attenuation by tissues, are proposed. The results of a clinical study of the proposed approach on volunteers were obtained, allowing to draw a conclusion about the applicability of the approach in a clinical setting.

Cyclical Patterns and a Regime Shift in the Character of Phytoplankton Blooms in a Restricted Sub-Tropical Lagoon, Indian River Lagoon, Florida, United States

This paper examines the character of phytoplankton blooms in a restricted sub-tropical lagoon along the Atlantic coast of central Florida. The results of the 23-year study (1997–2020) provide evidence for multiple types of variability in bloom activity, including cyclical patterns, stochastic events, and most prominently a regime shift in composition and intensity. Cyclical patterns (e.g., El Niño/La Niña periods) and stochastic events (e.g., tropical storms) influenced rainfall levels, which in turn impacted nutrient concentrations in the water column and the timing and intensity of blooms. In 2011, a major change occurred in the character of blooms, with a dramatic increase in peak biomass levels of blooms and the appearance of new dominant taxa, including the brown tide species Aureoumbra lagunensis and other nanoplanktonic species. Results of quantitative analyses reveal system behavior indicative of a regime shift. The shift coincided with widespread losses of seagrass community and reduced drift algae biomass. A combination of exceptionally low water temperatures in the winters of 2009/2010 and 2010/2011, hypersaline conditions associated with drought conditions, and high light attenuation caused by blooms appear to have contributed to the widespread and protracted decline in seagrass and drift macroalgal communities in the lagoon, leading to shifts in distribution of internal and external nutrient sources toward phytoplankton.