Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-241-2023
S. Tan, O. Mora, C. Tran
Abstract. Landslides are geological events in which masses of rock and soil slide down the slope of a mountain or hillside. They are influenced by topography, geology, weather, and human activity, and can cause extensive damage to the environment and infrastructure, as well as delay transportation networks. Therefore, it is imperative to detect early-warning signs of landslide hazards as a means of prevention. Traditional landslide surveillance consists of field mapping, but the process is costly and time consuming. Modern landslide mapping uses Light Detection and Ranging (LiDAR) derived Digital Elevation Models (DEMs) and sophisticated algorithms to analyze surface roughness and extract spatial features and patterns of landslide and landslide-prone areas. This study follows a previous study performed that demonstrated that it is possible to detect unstable terrain using algorithmic mapping techniques. The focus of this study is to show how spatial resolution can influence the accuracy of the classification results. The DEM data was resampled from 6 to 12, 24, 48 and 96 ft spatial resolution. The surface feature extractors employed (local topographic range, local topographic variability, slope, and roughness) are fused and analyzed simultaneously by applying k-means and Gaussian Mixture Model (GMM) clustering methods. When compared with the detailed, independently compiled landslide reference map, our data shows a decrease in performance as spatial resolution decreases. These results suggest that spatial resolution does impact the performance of landslide classification.
{"title":"EVALUATING THE INFLUENCE OF SPATIAL RESOLUTION ON LANDSLIDE DETECTION: A CASE STUDY IN THE CARLYON BEACH PENINSULA, WASHINGTON","authors":"S. Tan, O. Mora, C. Tran","doi":"10.5194/isprs-archives-xlviii-m-3-2023-241-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-241-2023","url":null,"abstract":"Abstract. Landslides are geological events in which masses of rock and soil slide down the slope of a mountain or hillside. They are influenced by topography, geology, weather, and human activity, and can cause extensive damage to the environment and infrastructure, as well as delay transportation networks. Therefore, it is imperative to detect early-warning signs of landslide hazards as a means of prevention. Traditional landslide surveillance consists of field mapping, but the process is costly and time consuming. Modern landslide mapping uses Light Detection and Ranging (LiDAR) derived Digital Elevation Models (DEMs) and sophisticated algorithms to analyze surface roughness and extract spatial features and patterns of landslide and landslide-prone areas. This study follows a previous study performed that demonstrated that it is possible to detect unstable terrain using algorithmic mapping techniques. The focus of this study is to show how spatial resolution can influence the accuracy of the classification results. The DEM data was resampled from 6 to 12, 24, 48 and 96 ft spatial resolution. The surface feature extractors employed (local topographic range, local topographic variability, slope, and roughness) are fused and analyzed simultaneously by applying k-means and Gaussian Mixture Model (GMM) clustering methods. When compared with the detailed, independently compiled landslide reference map, our data shows a decrease in performance as spatial resolution decreases. These results suggest that spatial resolution does impact the performance of landslide classification.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45533028","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-27-2023
A. Basu, S. Mamgain, A. Roy
Abstract. Climate change has exacerbated the intensity as well as frequency of forest fire events in the Indian state of Uttarakhand. The present study focusses on undertaking forest fire risk mapping across the state by utilizing geospatial technology along with Google Earth Engine. Ten parameters were identified that have a strong influence in determining fire prone areas. The Analytic Hierarchy Process (AHP) was then implemented for the development of the risk map in which criteria weights were assigned to the parameters based on their ability to influence a forest fire event. The analysis revealed that out of the total forest area, 24.22% is under ‘very high’ risk zone, 29.24% is under ‘high’ risk zone, 18.23% is under ‘moderate’ risk zone, 7.69% is under ‘low’ risk zone and 20.62% is under ‘very low’ risk zone of forest fire. Further study was carried out to determine fire risk levels in populated regions and in some of the most critical nature reserves having high ecological importance which reveals that ‘very high’ and ‘high’ risk zones have greater population density indicating the influence of anthropogenic activities on forest fire occurrence. The results additionally indicate that four national parks and wildlife sanctuaries are particularly vulnerable to forest fires at present which is a source of concern and requires intervention from the stakeholders.
{"title":"FOREST FIRE RISK MAPPING FOR THE HIMALAYAN STATE UTTARAKHAND USING GOOGLE EARTH ENGINE","authors":"A. Basu, S. Mamgain, A. Roy","doi":"10.5194/isprs-archives-xlviii-m-3-2023-27-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-27-2023","url":null,"abstract":"Abstract. Climate change has exacerbated the intensity as well as frequency of forest fire events in the Indian state of Uttarakhand. The present study focusses on undertaking forest fire risk mapping across the state by utilizing geospatial technology along with Google Earth Engine. Ten parameters were identified that have a strong influence in determining fire prone areas. The Analytic Hierarchy Process (AHP) was then implemented for the development of the risk map in which criteria weights were assigned to the parameters based on their ability to influence a forest fire event. The analysis revealed that out of the total forest area, 24.22% is under ‘very high’ risk zone, 29.24% is under ‘high’ risk zone, 18.23% is under ‘moderate’ risk zone, 7.69% is under ‘low’ risk zone and 20.62% is under ‘very low’ risk zone of forest fire. Further study was carried out to determine fire risk levels in populated regions and in some of the most critical nature reserves having high ecological importance which reveals that ‘very high’ and ‘high’ risk zones have greater population density indicating the influence of anthropogenic activities on forest fire occurrence. The results additionally indicate that four national parks and wildlife sanctuaries are particularly vulnerable to forest fires at present which is a source of concern and requires intervention from the stakeholders.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47723802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-65-2023
D. J. A. Davis, N. S. Guy
Abstract. Stockpiling aggregate materials is a common practice within the construction industry and with the demand for aggregates rapidly increasing, stockpile owners have taken a greater interest in the effective determination of volumes of inventory to optimize profit and limit waste. Historically, traditional stockpile measurement techniques were inaccurate but with the increase in demand, a higher quality and more reliable assessment of resources is necessary.The evolution of point cloud measurement and mapping technology, such as UAV and Terrestrial Laser Scanning (TLS), now means these techniques can be utilized for stockpile measurements. While some of the advantages over traditional techniques have been well documented, there is still a need to ascertain which of these methods is more applicable for volumetric surveys of different types of aggregate stockpiles.This study involved data collection and analysis from TLS and UAV photogrammetry for volumetric surveys and comparisons with Total Station (TS) measurement of the stockpiles for sharp sand, coarse (gravel) and finer aggregates.The research suggested that TS surveys could only be effectively utilized on sharp sand and coarse aggregates and was impractical for finer aggregates, and their results produced a general under-reporting of stockpile volumes. TLS and UAV provide non-contact collection with increased accuracy. There are differences in accuracy and appropriateness dependent on the aggregate type. It was observed that the TLS outperformed the TS approach whereas UAV demonstrated promise particularly at a lower altitude with greater overlap.Additional recommendations are shared to potentially improve productivity and inventory maintenance for Stockpiling Operations.
{"title":"AN ASSESSMENT OF POINT CLOUD DATA ACQUISITION TECHNIQUES FOR AGGREGATE STOCKPILES AND VOLUMETRIC SURVEYS","authors":"D. J. A. Davis, N. S. Guy","doi":"10.5194/isprs-archives-xlviii-m-3-2023-65-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-65-2023","url":null,"abstract":"Abstract. Stockpiling aggregate materials is a common practice within the construction industry and with the demand for aggregates rapidly increasing, stockpile owners have taken a greater interest in the effective determination of volumes of inventory to optimize profit and limit waste. Historically, traditional stockpile measurement techniques were inaccurate but with the increase in demand, a higher quality and more reliable assessment of resources is necessary.The evolution of point cloud measurement and mapping technology, such as UAV and Terrestrial Laser Scanning (TLS), now means these techniques can be utilized for stockpile measurements. While some of the advantages over traditional techniques have been well documented, there is still a need to ascertain which of these methods is more applicable for volumetric surveys of different types of aggregate stockpiles.This study involved data collection and analysis from TLS and UAV photogrammetry for volumetric surveys and comparisons with Total Station (TS) measurement of the stockpiles for sharp sand, coarse (gravel) and finer aggregates.The research suggested that TS surveys could only be effectively utilized on sharp sand and coarse aggregates and was impractical for finer aggregates, and their results produced a general under-reporting of stockpile volumes. TLS and UAV provide non-contact collection with increased accuracy. There are differences in accuracy and appropriateness dependent on the aggregate type. It was observed that the TLS outperformed the TS approach whereas UAV demonstrated promise particularly at a lower altitude with greater overlap.Additional recommendations are shared to potentially improve productivity and inventory maintenance for Stockpiling Operations.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43028819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-177-2023
E. B. A. Peixoto, E. Chiarani, W. Farias, B. Polli, R. Penteado, C. Freitas, D. Silva, J. A. Centeno
Abstract. The Jirau and Santo Antônio hydroelectric plants in Rondônia implemented a methodology using high-range cameras and artificial intelligence technology to address the challenge of managing logs transported by the river during floods. By applying machine learning techniques and neural networks, the system automatically monitors log transport and accumulation. Python 3, along with libraries like OpenCV, PIL, Numpy, and Pytorch, was utilized for efficient implementation. The methodology includes frame selection, log and debris segmentation, perspective correction, and log counting. Training was conducted using annotated images, and the detection process involved color segmentation, noise removal, and morphological operations. The calculated log and debris occupancy results were stored in a SQL database and presented on Power BI dashboards. The system aims to improve log management, ensuring power generation and ecological order are safeguarded.
{"title":"ARTIFICIAL INTELLIGENCE FOR REAL-TIME MONITORING OF LOGS ON THE MADEIRA RIVER: A CASE STUDY ON JIRAU HYDROELECTRIC PLANT","authors":"E. B. A. Peixoto, E. Chiarani, W. Farias, B. Polli, R. Penteado, C. Freitas, D. Silva, J. A. Centeno","doi":"10.5194/isprs-archives-xlviii-m-3-2023-177-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-177-2023","url":null,"abstract":"Abstract. The Jirau and Santo Antônio hydroelectric plants in Rondônia implemented a methodology using high-range cameras and artificial intelligence technology to address the challenge of managing logs transported by the river during floods. By applying machine learning techniques and neural networks, the system automatically monitors log transport and accumulation. Python 3, along with libraries like OpenCV, PIL, Numpy, and Pytorch, was utilized for efficient implementation. The methodology includes frame selection, log and debris segmentation, perspective correction, and log counting. Training was conducted using annotated images, and the detection process involved color segmentation, noise removal, and morphological operations. The calculated log and debris occupancy results were stored in a SQL database and presented on Power BI dashboards. The system aims to improve log management, ensuring power generation and ecological order are safeguarded.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43462782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-43-2023
D. Bhatt, M. Savarese, N. S. Hewitt, A. Gross, J. Wilder
Abstract. Geospatial data were used to analyze changes to geomorphology of barrier islands and beaches in Southwest Florida resulting from Hurricane Ian in late September 2022. The hurricane generated high intensity winds and storm surge causing more than $112 billion in damages, along with massive sediment mobilization due to erosion and deposition. This study quantified net sediment loss and gain on specific barrier islands by storm surge (Sanibel, Naples, Fort Myers Beach, others, though this paper focuses exclusively on Sanibel) by comparing pre- and post-Ian topography generated by a drone-flown LiDAR sensor; changes in elevation were used to quantify spatial variation in sediment volume. Data were collected immediately after Hurricane Ian and compared against topographic data collected by NOAA in 2018. Digital elevation models (DEMs) were used to compare topography, shoreline positions (relative to Mean High Water), foredune position, and volumetric changes using GIS technology. In general, the shoreline position after Ian changed little, indicating that the incoming surge had little influence on the beach. The foredunes, however, were deflated and set back by surge overwash. The outgoing surge created a much more dramatic geomorphologic change. Erosional surge channels cut through the foredunes and upper beach berm along many regions of the coastline. The ebb erosion also caused extensive damage to physical structures when located immediately behind the foredune. Lastly, this work demonstrates the value of employing GIS and remote sensing technology to problems of beach and dune management, the restoration of coastal ecosystems, the enhancement of resilience capacity of both natural and developed infrastructure, and the development of new policy needed to contend with the effects of climate change.
{"title":"REVEALING THE GEOMORPHOLOGIC IMPACTS OF HURRICANE IAN IN SOUTHWEST FLORIDA USING GEOSPATIAL TECHNOLOGY","authors":"D. Bhatt, M. Savarese, N. S. Hewitt, A. Gross, J. Wilder","doi":"10.5194/isprs-archives-xlviii-m-3-2023-43-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-43-2023","url":null,"abstract":"Abstract. Geospatial data were used to analyze changes to geomorphology of barrier islands and beaches in Southwest Florida resulting from Hurricane Ian in late September 2022. The hurricane generated high intensity winds and storm surge causing more than $112 billion in damages, along with massive sediment mobilization due to erosion and deposition. This study quantified net sediment loss and gain on specific barrier islands by storm surge (Sanibel, Naples, Fort Myers Beach, others, though this paper focuses exclusively on Sanibel) by comparing pre- and post-Ian topography generated by a drone-flown LiDAR sensor; changes in elevation were used to quantify spatial variation in sediment volume. Data were collected immediately after Hurricane Ian and compared against topographic data collected by NOAA in 2018. Digital elevation models (DEMs) were used to compare topography, shoreline positions (relative to Mean High Water), foredune position, and volumetric changes using GIS technology. In general, the shoreline position after Ian changed little, indicating that the incoming surge had little influence on the beach. The foredunes, however, were deflated and set back by surge overwash. The outgoing surge created a much more dramatic geomorphologic change. Erosional surge channels cut through the foredunes and upper beach berm along many regions of the coastline. The ebb erosion also caused extensive damage to physical structures when located immediately behind the foredune. Lastly, this work demonstrates the value of employing GIS and remote sensing technology to problems of beach and dune management, the restoration of coastal ecosystems, the enhancement of resilience capacity of both natural and developed infrastructure, and the development of new policy needed to contend with the effects of climate change.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42592591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-139-2023
P. Loh, Y. Twumasi, Z. H. Ning, M. Anokye, J. Oppong, R. Armah, C. Apraku, J. Namwamba
Abstract. Sea level rise poses risks to coastal areas which is increasingly rendering such areas susceptible to flood and shoreline retreat. Notably, coastal areas like Southern Louisiana located along the Gulf of Mexico has experienced endangering events of land subsidence due to flood inundations resulting from incessant distribution of hurricanes and tropical storms. This research therefore employed remote sensing data to analyze the impacts of sea level rise on coastal flooding and shoreline retreat along the coast of Louisiana. That is, by assessing Sentinel-2 imagery data to evaluate flood prone and flood extent areas particularly during the Louisiana floods and Hurricane Harvey. Based on this, the results show most of the inland parishes in coastal Louisiana such as Assumption, St. James, Livingston, Lafourche and Terrebonne were within high flood risk zones of about 9.3. These parishes also suffered severe damage in terms of affected croplands, potentially flooded areas and affected urban areas. On the other hand, most of the parishes in close proximity to the waterbodies such as the Gulf of Mexico were interestingly within low flood risk zones of about 6.1 suggesting proximity to waterbodies not being the only indicating factor of a flood prone area. This research also highlights that Louisiana's shorelines are rapidly receding at a rate that could result in the loss of one million acres of the state’s land in the next four decades. Hence, the results from this research are anticipated to contribute to sustainable shoreline setback plans and mitigative strategies to protect Louisiana's coast.
{"title":"ANALYZING THE IMPACT OF SEA LEVEL RISE ON COASTAL FLOODING AND SHORELINE CHANGES ALONG THE COAST OF LOUISIANA USING REMOTE SENSORY IMAGERY","authors":"P. Loh, Y. Twumasi, Z. H. Ning, M. Anokye, J. Oppong, R. Armah, C. Apraku, J. Namwamba","doi":"10.5194/isprs-archives-xlviii-m-3-2023-139-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-139-2023","url":null,"abstract":"Abstract. Sea level rise poses risks to coastal areas which is increasingly rendering such areas susceptible to flood and shoreline retreat. Notably, coastal areas like Southern Louisiana located along the Gulf of Mexico has experienced endangering events of land subsidence due to flood inundations resulting from incessant distribution of hurricanes and tropical storms. This research therefore employed remote sensing data to analyze the impacts of sea level rise on coastal flooding and shoreline retreat along the coast of Louisiana. That is, by assessing Sentinel-2 imagery data to evaluate flood prone and flood extent areas particularly during the Louisiana floods and Hurricane Harvey. Based on this, the results show most of the inland parishes in coastal Louisiana such as Assumption, St. James, Livingston, Lafourche and Terrebonne were within high flood risk zones of about 9.3. These parishes also suffered severe damage in terms of affected croplands, potentially flooded areas and affected urban areas. On the other hand, most of the parishes in close proximity to the waterbodies such as the Gulf of Mexico were interestingly within low flood risk zones of about 6.1 suggesting proximity to waterbodies not being the only indicating factor of a flood prone area. This research also highlights that Louisiana's shorelines are rapidly receding at a rate that could result in the loss of one million acres of the state’s land in the next four decades. Hence, the results from this research are anticipated to contribute to sustainable shoreline setback plans and mitigative strategies to protect Louisiana's coast.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46173174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-219-2023
V. S. Sajith Variyar, V. Sowmya, R. Sivanpillai, G. Brown
Abstract. The performance of the deep learning-based image segmentation is highly dependent on two major factors as follows: 1) The organization and structure of the architecture used to train the model and 2) The quality of input data used to train the model. The input image quality and the variety of training samples are highly influencing the features derived by the deep learning filters for segmentation. This study focus on the effect of image quality of a natural dataset of epiphytes captured using Unmanned Aerial Vehicles (UAV), while segmenting the epiphytes from other background vegetation. The dataset used in this work is highly challenging in terms of pixel overlap between target and background to be segmented, the occupancy of target in the image and shadows from nearby vegetation. The proposed study used four different contrast enhancement techniques to improve the image quality of low contrast images from the epiphyte dataset. The enhanced dataset with four different methods were used to train five different segmentation models. The segmentation performances of four different models are reported using structural similarity index (SSIM) and intersection over union (IoU) score. The study shows that the epiphyte segmentation performance is highly influenced by the input image quality and recommendations are given based on four different techniques for experts to work with segmentation with natural datasets like epiphytes. The study also reported that the occupancy of the target epiphyte and vegetation highly influence the performance of the segmentation model.
{"title":"THE EFFECT OF CONTRAST ENHANCEMENT ON EPIPHYTE SEGMENTATION USING GENERATIVE NETWORK","authors":"V. S. Sajith Variyar, V. Sowmya, R. Sivanpillai, G. Brown","doi":"10.5194/isprs-archives-xlviii-m-3-2023-219-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-219-2023","url":null,"abstract":"Abstract. The performance of the deep learning-based image segmentation is highly dependent on two major factors as follows: 1) The organization and structure of the architecture used to train the model and 2) The quality of input data used to train the model. The input image quality and the variety of training samples are highly influencing the features derived by the deep learning filters for segmentation. This study focus on the effect of image quality of a natural dataset of epiphytes captured using Unmanned Aerial Vehicles (UAV), while segmenting the epiphytes from other background vegetation. The dataset used in this work is highly challenging in terms of pixel overlap between target and background to be segmented, the occupancy of target in the image and shadows from nearby vegetation. The proposed study used four different contrast enhancement techniques to improve the image quality of low contrast images from the epiphyte dataset. The enhanced dataset with four different methods were used to train five different segmentation models. The segmentation performances of four different models are reported using structural similarity index (SSIM) and intersection over union (IoU) score. The study shows that the epiphyte segmentation performance is highly influenced by the input image quality and recommendations are given based on four different techniques for experts to work with segmentation with natural datasets like epiphytes. The study also reported that the occupancy of the target epiphyte and vegetation highly influence the performance of the segmentation model.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42606577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-235-2023
Ramesh Sivanpillai, Maria Oreshkina, Paden Bear, Isaac Boettcher, Tyler Bradshaw, Isaac Coleman, Jessica Gifford
Abstract. Identifying newly inundated areas following flood events is essential for planning rescue missions. These maps must be generated quickly as the spatial extent of the inundated areas might change during a single flood event. Several methods exist for generating such maps and several rely on one or more geospatial data to exclude existing waterbodies in an affected area. In this study, we tested a rapid flood mapping method that uses a pair of pre- and post-flood satellite images on seven sites throughout the US. We derived Normalized Difference Water Index (NDWI) and Modified NDWI (MNDWI) images from pre- and post-flood Landsat images and identified the optimal threshold values that highlighted newly inundated areas at these sites. The accuracy of the inundation maps was determined using manually interpreted verification data from the pairs of satellite images. Image analysts have identified the optimal threshold values between 25 and 40 minutes. Maps of newly inundated areas derived from differencing MNDWI and NDWI images had higher overall accuracy > 93%. Results obtained in this study confirms the utility of this rapid flood mapping technique to identify inundated areas using pre- and post-flood satellite images.
{"title":"MAPPING NEWLY INUNDATED AREAS IN POST-FLOOD LANDSAT IMAGES USING THRESHOLDING TECHNIQUES","authors":"Ramesh Sivanpillai, Maria Oreshkina, Paden Bear, Isaac Boettcher, Tyler Bradshaw, Isaac Coleman, Jessica Gifford","doi":"10.5194/isprs-archives-xlviii-m-3-2023-235-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-235-2023","url":null,"abstract":"Abstract. Identifying newly inundated areas following flood events is essential for planning rescue missions. These maps must be generated quickly as the spatial extent of the inundated areas might change during a single flood event. Several methods exist for generating such maps and several rely on one or more geospatial data to exclude existing waterbodies in an affected area. In this study, we tested a rapid flood mapping method that uses a pair of pre- and post-flood satellite images on seven sites throughout the US. We derived Normalized Difference Water Index (NDWI) and Modified NDWI (MNDWI) images from pre- and post-flood Landsat images and identified the optimal threshold values that highlighted newly inundated areas at these sites. The accuracy of the inundation maps was determined using manually interpreted verification data from the pairs of satellite images. Image analysts have identified the optimal threshold values between 25 and 40 minutes. Maps of newly inundated areas derived from differencing MNDWI and NDWI images had higher overall accuracy > 93%. Results obtained in this study confirms the utility of this rapid flood mapping technique to identify inundated areas using pre- and post-flood satellite images.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44733771","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-203-2023
Onteddu Chaitanya Reddy, Illa Dinesh Kumar, Pingali Sathvika, Sajith Variyar, Sowmya, R. Sivanpillai
Abstract. Deep Learning (DL) networks used in image segmentation tasks must be trained with input images and corresponding masks that identify target features in them. DL networks learn by iteratively adjusting the weights of interconnected layers using backpropagation, a process that involves calculating gradients and minimizing a loss function. This allows the network to learn patterns and relationships in the data, enabling it to make predictions or classifications on new, unseen data. Training any DL network requires specifying values of the hyperparameters such as input image size, batch size, and number of epochs among others. Failure to specify optimal values for the parameters will increase the training time or result in incomplete learning. The rationale of this study was to evaluate the effect of input image and batch sizes on the performance of DeepLabV3+ using Sentinel 2 A/B RGB images and labels obtained from Kaggle. We trained DeepLabV3+ network six times with two sets of input images of 128 × 128-pixel, and 256 × 256-pixel dimensions with 4, 8 and 16 batch sizes. The model is trained for 100 epochs to ensure that the loss plot reaches saturation and the model converged to a stable solution. Predicted masks generated by each model were compared to their corresponding test mask images based on accuracy, precision, recall and F1 scores. Results from this study demonstrated that image size of 256 × 256 and batch size 4 achieved highest performance. It can also be inferred that larger input image size improved DeepLabV3+ model performance.
{"title":"EFFECT OF HYPERPARAMETERS ON DEEPLABV3+ PERFORMANCE TO SEGMENT WATER BODIES IN RGB IMAGES","authors":"Onteddu Chaitanya Reddy, Illa Dinesh Kumar, Pingali Sathvika, Sajith Variyar, Sowmya, R. Sivanpillai","doi":"10.5194/isprs-archives-xlviii-m-3-2023-203-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-203-2023","url":null,"abstract":"Abstract. Deep Learning (DL) networks used in image segmentation tasks must be trained with input images and corresponding masks that identify target features in them. DL networks learn by iteratively adjusting the weights of interconnected layers using backpropagation, a process that involves calculating gradients and minimizing a loss function. This allows the network to learn patterns and relationships in the data, enabling it to make predictions or classifications on new, unseen data. Training any DL network requires specifying values of the hyperparameters such as input image size, batch size, and number of epochs among others. Failure to specify optimal values for the parameters will increase the training time or result in incomplete learning. The rationale of this study was to evaluate the effect of input image and batch sizes on the performance of DeepLabV3+ using Sentinel 2 A/B RGB images and labels obtained from Kaggle. We trained DeepLabV3+ network six times with two sets of input images of 128 × 128-pixel, and 256 × 256-pixel dimensions with 4, 8 and 16 batch sizes. The model is trained for 100 epochs to ensure that the loss plot reaches saturation and the model converged to a stable solution. Predicted masks generated by each model were compared to their corresponding test mask images based on accuracy, precision, recall and F1 scores. Results from this study demonstrated that image size of 256 × 256 and batch size 4 achieved highest performance. It can also be inferred that larger input image size improved DeepLabV3+ model performance.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41357619","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-05DOI: 10.5194/isprs-archives-xlviii-m-3-2023-109-2023
Kalingga Titon, Nur Ihsan, A. B. Harto, Dimara Sakti, K. Wikantika
Abstract. The coastal area is an area that has a dense population with a lot of human activities that occur there. Due to environmental changes and human activities, changes often occur in coastal areas ranging from erosion and sedimentation. Changes must continuously be monitored to plan countermeasures due to the occurring phenomena. This study aims to create a website-based application to monitor coastal areas. This study will use Landsat data 5,7,8, and 9 to see changes in coastal areas. The analysis can be provided from 1985 until recent data by integrating four Landsat satellites. The NDWI index (Normalized Difference Wetness Index) analyzes changes occurring in coastal areas and differentiates between water and land area. The analysis is not only in the form of changes that occur in coastal areas but also in time series analysis, and trends that occur at a point can be analyzed using land trend analysis. The resulting website based on Cloud Computation in Google Earth Engine can be seen at the link https://bit.ly/MonitoringPesisir. This website can automatically update, and users can choose the location to monitor. This research is expected to be used by policymakers to monitor and plan the development and regulation of coastal areas.
{"title":"MONITORING COASTAL AREAS USING NDWI FROM LANDSAT IMAGE DATA FROM 1985 BASED ON CLOUD COMPUTATION GOOGLE EARTH ENGINE AND APPS","authors":"Kalingga Titon, Nur Ihsan, A. B. Harto, Dimara Sakti, K. Wikantika","doi":"10.5194/isprs-archives-xlviii-m-3-2023-109-2023","DOIUrl":"https://doi.org/10.5194/isprs-archives-xlviii-m-3-2023-109-2023","url":null,"abstract":"Abstract. The coastal area is an area that has a dense population with a lot of human activities that occur there. Due to environmental changes and human activities, changes often occur in coastal areas ranging from erosion and sedimentation. Changes must continuously be monitored to plan countermeasures due to the occurring phenomena. This study aims to create a website-based application to monitor coastal areas. This study will use Landsat data 5,7,8, and 9 to see changes in coastal areas. The analysis can be provided from 1985 until recent data by integrating four Landsat satellites. The NDWI index (Normalized Difference Wetness Index) analyzes changes occurring in coastal areas and differentiates between water and land area. The analysis is not only in the form of changes that occur in coastal areas but also in time series analysis, and trends that occur at a point can be analyzed using land trend analysis. The resulting website based on Cloud Computation in Google Earth Engine can be seen at the link https://bit.ly/MonitoringPesisir. This website can automatically update, and users can choose the location to monitor. This research is expected to be used by policymakers to monitor and plan the development and regulation of coastal areas.\u0000","PeriodicalId":30634,"journal":{"name":"The International Archives of the Photogrammetry Remote Sensing and Spatial Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47152860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：