Developing an Autonomous Hovercraft for Benthic Surveying in Very Shallow Waters

2020 ◽  
Author(s):  
Meghan Troup ◽  
David Barclay ◽  
Matthew Hatcher
Keyword(s):  
Remote Sensing ◽  
Large Scale ◽  
Water Clarity ◽  
Research Area ◽  
Optical Methods ◽  
Small Scale ◽  
Sediment Type ◽  
Side Scan Sonar ◽  
Single Beam ◽  
Remote Sensing Methods

<p>Benthic surveys in very shallow water (< 1 meter) are often carried out by remote sensing methods such as LiDAR, satellite imagery, and aerial photography, or by written observations paired with GPS point measurements and underwater video. Remote sensing can be helpful for large scale mapping endeavors, but the optical methods commonly used are limited in their effectiveness by cloud cover and water clarity. In situ surveys are often carried out manually and can therefore be quite inefficient. A proposed alternative method of small scale, high resolution mapping is an autonomous, amphibious hovercraft, fitted with high frequency single-beam and side-scan sonar instruments. A hovercraft can move seamlessly from land to water which allows for convenient and simple deployment. The sonar instruments are attached to a boat-shaped outrigger hull that can be raised and lowered automatically, enabling data collection in water as shallow as 10 cm. These data are used to extract seafloor characteristics in order to create detailed maps of the research area that include information such as sediment type, presence and extent of flora and fauna, and small-scale bathymetry.</p>

Research of Remote Sensing Image Segmentation Based on Mean Shift and Region Merging

2011 ◽  
Vol 90-93 ◽  
pp. 2836-2839 ◽  
Author(s):  
Jian Cui ◽  
Dong Ling Ma ◽  
Ming Yang Yu ◽  
Ying Zhou
Keyword(s):  
Remote Sensing ◽  
Image Segmentation ◽  
Large Scale ◽  
Mean Shift ◽  
Image Features ◽  
Small Scale ◽  
Region Merging ◽  
Segmentation Method ◽  
Shift Method ◽  
Scale Parameters

In order to extract ground information more accurately, it is important to find an image segmentation method to make the segmented features match the ground objects. We proposed an image segmentation method based on mean shift and region merging. With this method, we first segmented the image by using mean shift method and small-scale parameters. According to the region merging homogeneity rule, image features were merged and large-scale image layers were generated. What’s more, Multi-level image object layers were created through scaling method. The test of segmenting remote sensing images showed that the method was effective and feasible, which laid a foundation for object-oriented information extraction.

scholarly journals Multi-Scale Remote Sensing Semantic Analysis Based on a Global Perspective

2019 ◽  
Vol 8 (9) ◽  
pp. 417 ◽  
Author(s):  
Wei Cui ◽  
Dongyou Zhang ◽  
Xin He ◽  
Meng Yao ◽  
Ziwei Wang ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Receptive Field ◽  
Large Scale ◽  
Semantic Information ◽  
Short Term Memory ◽  
Global Perspective ◽  
Small Scale ◽  
Multi Scale ◽  
Memory Network ◽  
Long Short Term Memory

Remote sensing image captioning involves remote sensing objects and their spatial relationships. However, it is still difficult to determine the spatial extent of a remote sensing object and the size of a sample patch. If the patch size is too large, it will include too many remote sensing objects and their complex spatial relationships. This will increase the computational burden of the image captioning network and reduce its precision. If the patch size is too small, it often fails to provide enough environmental and contextual information, which makes the remote sensing object difficult to describe. To address this problem, we propose a multi-scale semantic long short-term memory network (MS-LSTM). The remote sensing images are paired into image patches with different spatial scales. First, the large-scale patches have larger sizes. We use a Visual Geometry Group (VGG) network to extract the features from the large-scale patches and input them into the improved MS-LSTM network as the semantic information, which provides a larger receptive field and more contextual semantic information for small-scale image caption so as to play the role of global perspective, thereby enabling the accurate identification of small-scale samples with the same features. Second, a small-scale patch is used to highlight remote sensing objects and simplify their spatial relations. In addition, the multi-receptive field provides perspectives from local to global. The experimental results demonstrated that compared with the original long short-term memory network (LSTM), the MS-LSTM’s Bilingual Evaluation Understudy (BLEU) has been increased by 5.6% to 0.859, thereby reflecting that the MS-LSTM has a more comprehensive receptive field, which provides more abundant semantic information and enhances the remote sensing image captions.

scholarly journals Cartographic Remote-Sensing Monitoring of Glaciological Systems (Example, Mount El‛ Brus, U.S.S.R.) (Abstract)

1987 ◽  
Vol 9 ◽  
pp. 247-248
Author(s):  
Yu. F. Knizhnikov ◽  
V.I. Kravtsova ◽  
I.A. Labutina
Keyword(s):  
Remote Sensing ◽  
High Resolution ◽  
Large Scale ◽  
Quantitative Information ◽  
Snow Accumulation ◽  
Aerial Photographs ◽  
Small Scale ◽  
Medium Scale ◽  
Remote Sensing Monitoring ◽  
Periglacial Zone

Remote-sensing methods in monitoring the glacierization of Mount EI‛ brus are used to produce base and dynamic maps, and to obtain quantitative information (dynamic indices) about the rate, intensity, and variations of the process. The monitoring system is divided, according to scope and territory covered, into small-scale for total glacierization and the periglacial zone, medium-scale for separate glaciers, and large-scale (detailed) for part of the glaciers or sectors of the adjoining slopes. The approximate relationship of even scales is 1 : 4. Small-scale monitoring remote-sensing systems are important for making maps showing the complex characteristics of the glaciological system. A series of maps was produced including geographical, those of high-altitude zones, slope and exposure angles, geological, glaciomorphological, climatic (temperature, precipitation, and winds), distribution of direct solar radiation, hydrological (source of streams), seats of avalanches, and landslides. All these data serve as a cartographical basis in monitoring the glacierization of Mount EI‛ brus. They are compiled from remotely sensed and Earth-based data. Current monitoring on a small scale includes observations of the conditions which determine the existence of the glacial system - this includes data on winter snowfall and the period of snow cover. These observations were obtained from meteorological and resource satellites, and from scanner data of medium and high resolution. Also important are observations of changes in the outline of glaciers, times of snowfall and character of the distribution of snow, and its redistribution due to avalanches and snowstorms. High-resolution space photographs, small-scale aerial photographs, and aerovisual observations provide the data for these observations. It has been determined that the area of the glaciers of Mount El‛ brus has been reduced by 1 % in the last 25 years, i.e. the rate of its deglacierization dropped sharply as compared to preceding decades. The role of quantitative information gains importance in the medium-scale level of monitoring. Topographical maps of separate glaciers compiled from aerial photographs or data from ground stereo-photogrammetric surveys constitute the base maps at this level. The main method used in monitoring were large-scale surveys from aircraft, perspective surveys from helicopters, and phototheodolite surveys. Multi-date surveys of the glaciers provide data about the changes in their outlines and height, the character of their relief, their moraines, the amount of snow accumulation and ablation in separate years, the surface rates of ice flow and their fluctuations. The techniques by which quantitative information is obtained about changes in the glaciers are derived from processing the data of multi-date surveys. The organization and techniques of phototheodolite surveys have been improved. A theory evolved for determining the surface-ice movement by stereo-photogrammetric means and the technique for it has also improved; algorithms and programs for machine processing of the data of multi-date surveys (ground and from aircraft) have been produced At this level of monitoring, it has been found that the retreat rate of most glaciers has slowed down and several glaciers are now in equilibrium. Several glaciers became active at the beginning of the 1970s and 1980s; this was accompanied by an increase in their height and forward movement. For example, activation of Kyukyurtlyu Glacier has been recorded (higher surface and increasing flow rate) which has caused the glacier to move forward 100 m. Surveys at an interval of 2 years recorded the beginning of the process of retreat of this glacier. Detailed monitoring is used to detect the mechanism of the dynamic processes and to study it on local representative sectors. On a glacier it may take the form of annual surveys of its tongue, which makes it possible to observe the processes of formation of moraines and glacio-fluvial relief. Studies may also be made of the mechanism of the movement of avalanches and landslides, deducing their quantitative characteristics and appraising the results of avalanches and landslides. Multi-date surveys of sectors of the slopes provide information about processes in the periglacial zone. At this level, regularly repeated ground stereo-photogrammetric surveys are the main means of observation. Glaciological remote-sensing monitoring provides a wealth of data for theoretical development in the field of glaciology. It makes it possible to forecast and produce warnings about hazardous processes and phenomena.

scholarly journals New Statistical Model for Variability of Aerosol Optical Thickness: Theory and Application to MODIS Data over Ocean*

2016 ◽  
Vol 73 (2) ◽  
pp. 821-837 ◽  
Author(s):  
Mikhail D. Alexandrov ◽  
Igor V. Geogdzhayev ◽  
Kostas Tsigaridis ◽  
Alexander Marshak ◽  
Robert Levy ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Optical Thickness ◽  
Large Scale ◽  
Climate Models ◽  
Marine Boundary Layer ◽  
Structure Functions ◽  
Aerosol Optical Thickness ◽  
Small Scale ◽  
Scale Invariant ◽  
Modis Data

Abstract A novel model for the variability in aerosol optical thickness (AOT) is presented. This model is based on the consideration of AOT fields as realizations of a stochastic process that is the exponent of an underlying Gaussian process with a specific autocorrelation function. In this approach, AOT fields have lognormal PDFs and structure functions with the correct asymptotic behavior at large scales. The latter is an advantage compared with fractal (scale invariant) approaches. The simple analytical form of the structure function in the proposed model facilitates its use for the parameterization of AOT statistics derived from remote sensing data. The new approach is illustrated using a 1-yr-long global MODIS AOT dataset (over ocean) with 10-km resolution. It was used to compute AOT statistics for sample cells forming a grid with 5° spacing. The observed shapes of the structure functions indicated that, in a large number of cases, the AOT variability is split into two regimes that exhibit different patterns of behavior: small-scale stationary processes and trends reflecting variations at larger scales. The small-scale patterns are suggested to be generated by local aerosols within the marine boundary layer, while the large-scale trends are indicative of elevated aerosols transported from remote continental sources. This assumption is evaluated by comparison of the geographical distributions of these patterns derived from MODIS data with those obtained from the GISS GCM. This study shows considerable potential to enhance comparisons between remote sensing datasets and climate models beyond regional mean AOTs.

scholarly journals AUTOMATIC LARGE-SCALE 3D BUILDING SHAPE REFINEMENT USING CONDITIONAL GENERATIVE ADVERSARIAL NETWORKS

2018 ◽  
Vol XLII-2 ◽  
pp. 103-108 ◽  
Author(s):  
K. Bittner ◽  
P. d’Angelo ◽  
M. Körner ◽  
P. Reinartz
Keyword(s):  
Remote Sensing ◽  
Urban Areas ◽  
Large Scale ◽  
Point Clouds ◽  
Research Area ◽  
Sar Interferometry ◽  
Generative Adversarial Networks ◽  
Generative Adversarial Network ◽  
Building Shape ◽  
Object Shapes

<p><strong>Abstract.</strong> Three-dimensional building reconstruction from remote sensing imagery is one of the most difficult and important 3D modeling problems for complex urban environments. The main data sources provided the digital representation of the Earths surface and related natural, cultural, and man-made objects of the urban areas in remote sensing are the <i>digital surface models (DSMs)</i>. The DSMs can be obtained either by <i>light detection and ranging (LIDAR)</i>, SAR interferometry or from stereo images. Our approach relies on automatic global 3D building shape refinement from stereo DSMs using deep learning techniques. This refinement is necessary as the DSMs, which are extracted from image matching point clouds, suffer from occlusions, outliers, and noise. Though most previous works have shown promising results for building modeling, this topic remains an open research area. We present a new methodology which not only generates images with continuous values representing the elevation models but, at the same time, enhances the 3D object shapes, buildings in our case. Mainly, we train a <i>conditional generative adversarial network (cGAN)</i> to generate accurate LIDAR-like DSM height images from the noisy stereo DSM input. The obtained results demonstrate the strong potential of creating large areas remote sensing depth images where the buildings exhibit better-quality shapes and roof forms.</p>

scholarly journals Acoustic Approach to Determining Seabed Substrates Distribution at Mandi Darah Island, Sabah

2021 ◽  
Vol 10 (2) ◽  
pp. 374-382
Author(s):  
Mohd Shukri Mohd Yusop ◽  
Mohd Norsyarizad Razali ◽  
Nazirah Md Tarmizi ◽  
Mohd Najib Abdul Ghani Yolhamid ◽  
M.N. Azzeri ◽  
...  
Keyword(s):  
Direct Observation ◽  
Large Scale ◽  
Life Support ◽  
Natural Habitat ◽  
Marine Ecosystem ◽  
Water Clarity ◽  
Habitat Mapping ◽  
Ecological Processes ◽  
Single Beam ◽  
Management And Conservation

Marine ecosystems and natural habitat play the important role of the Earth’s life support system. They significantly contribute to economies and food safety and help preserve ecological processes. However, the devastation of the marine ecosystem in Malaysia due to the human factor and climate change is quite alarming. Therefore, spatial marine information, especially on the distribution of seabed substrates and habitat mapping, are of utmost importance for marine ecosystem management and conservation. Traditionally, seabed substrate and habitat mapping were classified based on direct observation techniques such as photography, video, sampling, coring and scuba diving. These techniques are often limited due to water clarity and weather conditions and only suitable for smaller scale surveys. In this study, we employed an acoustic approach using the RoxAnn Acoustic Ground Discrimination System (AGDS) with a high-frequency single-beam echo sounder to examine the distribution of seabed substrate at the Mandi Darah Island, Sabah. The acoustic signals recorded by AGDS are translated into hardness and roughness indices which are then used to identify the unique characteristics of the recorded seabed types. The analysis has shown that fifteen types of substrates, ranging from silt to rough/some seagrass, have been identified and classified. The findings demonstrated that the acoustic method was a better alternative for seabed substrate determination than the conventional direct observation techniques in terms of cost and time spent, especially in large scale surveys. The seabed substrate dataset from this study could be used as baseline information for the better management and conservation of the marine ecosystem.

scholarly journals Classification of the Complex Agricultural Planting Structure with a Semi-Supervised Extreme Learning Machine Framework

2020 ◽  
Vol 12 (22) ◽  
pp. 3708 ◽  
Author(s):  
Ziyi Feng ◽  
Guanhua Huang ◽  
Daocai Chi
Keyword(s):  
Remote Sensing ◽  
Support Vector Machine ◽  
Computational Efficiency ◽  
Large Scale ◽  
Statistical Data ◽  
Support Vector ◽  
Small Scale ◽  
Remote Sensing Images ◽  
Learning Machine

Many approaches have been developed to analyze remote sensing images. However, for the classification of large-scale problems, most algorithms showed low computational efficiency and low accuracy. In this paper, the newly developed semi-supervised extreme learning machine (SS-ELM) framework with k-means clustering algorithm for image segmentation and co-training algorithm to enlarge the sample sets was used to classify the agricultural planting structure at large-scale areas. Data sets collected from a small-scale area within the Hetao Irrigation District (HID) at the upper reaches of the Yellow River basin were used to evaluate the SS-ELM framework. The results of the SS-ELM algorithm were compared with those of the random forest (RF), ELM, support vector machine (SVM) and semi-supervised support vector machine (S-SVM) algorithms. Then the SS-ELM algorithm was applied to analyze the complex planting structure of HID in 1986–2010 by comparing the remote sensing estimated results with the statistical data. In the small-scale case, the SS-ELM algorithm performed better than the RF, ELM, SVM, and S-SVM algorithms. For the SS-ELM algorithm, the average overall accuracy (OA) was in a range of 83.00–92.17%. On the contrary, for the other four algorithms, their average OA values ranged from 56.97% to 92.84%. Whereas, in the classification of planting structure in HID, the SS-ELM algorithm had an excellent performance in classification accuracy and computational efficiency for three major planting crops including maize, wheat, and sunflowers. The estimated areas by using the SS-ELM algorithm based on the remote sensing images were consistent with the statistical data, and their difference was within a range of 3–25%. This implied that the SS-ELM framework could be served as an effective method for the classification of complex planting structures with relatively fast training, good generalization, universal approximation capability, and reasonable learning accuracy.

Using remote sensing to monitor peatland fire occurrence and recovery

2020 ◽  
Author(s):  
Kirsten Lees ◽  
Josh Buxton ◽  
Chris Boulton ◽  
Tim Lenton
Keyword(s):  
Remote Sensing ◽  
Great Britain ◽  
Management Practice ◽  
Google Earth ◽  
Small Scale ◽  
Fire Occurrence ◽  
Wildfire Occurrence ◽  
Visual Record ◽  
Remote Sensing Methods ◽  
The Impact

&lt;p&gt;Many peatland areas in Great Britain are managed as grouse moors, with regular burns as part of management practice to encourage heather growth. Remote sensing has the potential to monitor the size, location, and impact of these burns using new fine resolution satellites such as Sentinel-2. Google Earth Engine allows large areas to be analysed at small scale over several years, building up a visual record of fire occurrence. This study uses satellite data to map managed burns on several areas of moorland around Great Britain, and uses remote sensing methods to assess the impact of this management strategy on vegetation cover. The project also considers how areas subject to managed burns react to wildfire occurrence, with the 2018 Saddleworth wildfire as a case study.&lt;/p&gt;

scholarly journals Validation of the Accuracy of Different Precipitation Datasets over Tianshan Mountainous Area

2015 ◽  
Vol 2015 ◽  
pp. 1-10 ◽  
Author(s):  
Chuancheng Zhao ◽  
Shuxia Yao ◽  
Shiqiang Zhang ◽  
Haidong Han ◽  
Qiudong Zhao ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Satellite Remote Sensing ◽  
Large Scale ◽  
Regional Scale ◽  
Precipitation Amount ◽  
Rain Gauge ◽  
Small Scale ◽  
Accurate Information ◽  
Mountainous Area ◽  
Basin Scale

Precipitation is one of the important water supplies in the arid and semiarid regions of northwestern China, playing a vital role in maintaining the fragile ecosystem. In remote mountainous area, it is difficult to obtain an accurate and reliable spatialization of the precipitation amount at the regional scale due to the inaccessibility, the sparsity of observation stations, and the complexity of relationships between precipitation and topography. Furthermore, accurate precipitation is important driven data for hydrological models to assess the water balance and water resource for hydrologists. Therefore, the use of satellite remote sensing becomes an important means over mountainous area. Precipitation datasets based on station data or pure satellite data have been increasingly available in spite of several weaknesses. This paper evaluates the usefulness of three precipitation datasets including TRMM 3B43_V6, 3B43_V7, and Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation with rain gauge data over Tianshan mountainous area where precipitation data is scarce. The results suggest that precipitation measurements only provided accurate information on a small scale, while the satellite remote sensing of precipitation had obvious advantages in basin scale or large scale especially over remote mountainous area.

scholarly journals Violence Detection With Two-Stream Neural Network Based on C3D

2021 ◽  
Vol 15 (4) ◽  
pp. 1-17
Author(s):  
zanzan Lu ◽  
Xuewen Xia ◽  
Hongrun Wu ◽  
Chen Yang
Keyword(s):  
Large Scale ◽  
Research Area ◽  
Video Data ◽  
Small Scale ◽  
Stream Network ◽  
Generalization Ability ◽  
Violence Detection ◽  
Good Ability ◽  
Important Research Area ◽  
Public Datasets

In recent years, violence detection has gradually turned into an important research area in computer vision, and have proposed many models with high accuracy. However, the unsatisfactory generalization ability of these methods over different datasets. In this paper, the authors propose a violence detection method based on C3D two-stream network for spatiotemporal features. Firstly, the authors preprocess the video data of RGB stream and optical stream respectively. Secondly, the authors feed the data into two C3D networks to extract features from the RGB flow and the optical flow respectively. Third, the authors fuse the features extracted by the two networks to obtain a final prediction result. To testify the performance of the proposed model, four different datasets (two public datasets and two self-built datasets) are selected in this paper. The experimental results show that our model has good generalization ability compared to state-of-the-art methods, since it not only has good ability on large-scale datasets, but also performs well on small-scale datasets.

Sign in / Sign up

Export Citation Format

Share Document