Fig 3 - uploaded by Jacob Shermeyer
Content may be subject to copyright.
The 2000 to 2010 conventional supervised classification change map generated for the study site overlaid on July 27, 2000, Landsat TM image visualized in a false color composite using bands 7 (red), 4 (green), and 3 (blue).
Source publication
Two forestry-change detection methods are described, compared, and contrasted for estimating deforestation and growth in threatened forests in southern Peru from 2000 to 2010. The methods used in this study rely on freely available data, including atmospherically corrected Landsat 5 Thematic Mapper and Moderate Resolution Imaging Spectroradiometer...
Context in source publication
Context 1
... signature extraction, a k-NN classifier was applied to the 2000 and 2010 Landsat imagery. FNF maps were generated and a change map was subsequently created using postclas- sification differencing (Fig. 3). An initial visual assessment reveals a substantial amount of forest loss. Dense forests in the southeast are appropriately classified with a minor amount of speckled misclassification. An area of false-positive growth can be seen in the southwestern corner of the study area. Areas of loss are predominantly located near existing ...
Citations
... Contemporary literature on the topic seems to be geared towards machine learning approaches. Just to name a few examples we refer to [10] that employed deep learning for deforestation detection in the Brazilian Amazon and [15] that presented k-nearest neighbour classification to track deforestation in Peru. Bayesian approach has also seen use, e.g. in [12] on time series data from Fiji. ...
This paper is devoted to the problem of detection of forest and non-forest areas on Earth images. We propose two statistical methods to tackle this problem: one based on multiple hypothesis testing with parametric distribution families, another one -- on non-parametric tests. The parametric approach is novel in the literature and relevant to a larger class of problems -- detection of natural objects, as well as anomaly detection. We develop mathematical background for each of the two methods, build self-sufficient detection algorithms using them and discuss numerical aspects of their implementation. We also compare our algorithms with those from standard machine learning using satellite data.
... RS-IEA is also the main function of RS application system. RS-IEA includes image segmentation [67], spectral-based classification [68], scene classification, and other pixel-based image processing methods; as well as target detection [69], target change and tracking [70], target classification, target recognition [33], and other target-based information extraction algorithms. ...
With the rapid development of high-resolution earth observation system, the data processing, algorithm design and system development of Remote Sensing Spatial-Temporal Big Data (RS-STBD) have gradually become the bottleneck problems in the application and development of earth observation system. The research on the model, algorithm and system of RS-STBD processing involves complex scientific problems, technical bottlenecks and inconstant requirements of engineering applications. This paper summarizes the data type and processing theory model of RS-STBD, the high performance algorithm design based on cloud service and intelligent computing, and the architecture design and engineering development methods of the complex remote sensing application system. Furthermore, the existing problems in the current research are analyzed, and the related solutions are given. Finally, the future development trend of scientific exploration, technical research and application development of RS-STBD is prospected.
... RS-IEA is also the main function of the RS application system. RS-IEA includes image segmentation [67], spectral-based classification [68], scene classification, and other pixel-based image processing methods, as well as target detection [31], target change and tracking [69], target classification, target recognition [33], and other target-based information extraction algorithms. ...
Since the amount of remote sensing data that can be obtained every day has reached ten terabytes, the storage and sharing of massive remote sensing data and the construction of product production application architecture have become imminent research topics. The purpose of this study is to build an efficient system that can meet the needs of massive remote sensing data storage, calculation, and distribution. The article elaborates on the latest research progress of cloud service remote sensing data distribution and intelligent production from the four aspects of remote sensing cloud service development, data distribution technology, cloud service technology, and operation mechanism of remote sensing cloud computing processing model, and proposes cloud service remote sensing Key issues facing data distribution, and summarize and prospect related research, with a view to promoting the further development of research on cloud service remote sensing data distribution and intelligent production.
... Traditionally, when differencing maps, the result of the comparison is compounding misclassification error that leads to a lower overall agreement. Expected overall accuracy for change products or agreement during a comparison of maps typically can be calculated by multiplying the two products' overall accuracies together [45,46]. For example, under ideal conditions we would expect our two maps with potential accuracies of 92% and 93% to have a maximal agreement of around 85.6%. ...
A unique, multi-tiered approach was applied to map crop-residue cover on the Eastern Shore of the Chesapeake Bay, United States. Field measurements of crop-residue cover were used to calibrate residue mapping using shortwave infrared (SWIR) indices derived from WorldView-3 imagery for a 12-km × 12-km footprint. The resulting map was then used to calibrate and subsequently classify crop residue mapping using Landsat imagery at a larger spatial resolution and extent. This manuscript describes how the method was applied and presents results in the form of crop-residue cover maps, validation statistics, and quantification of conservation tillage implementation in the agricultural landscape. Overall accuracy for maps derived from Landsat 7 and Landsat 8 were comparable at roughly 92% (+/− 10%). Tillage class-specific accuracy was also strong and ranged from 75% to 99%. The approach, which employed a 12-band image stack of six tillage spectral indices and six individual Landsat bands, was shown to be adaptable to variable soil-moisture conditions—under dry conditions (Landsat 7, 14 May 2015) the majority of predictive power was attributed to SWIR indices, and under wet conditions (Landsat 8, 22 May 2015) single band reflectance values were more effective at explaining variability in residue cover. Summary statistics of resulting tillage class occurrence matched closely with conservation tillage implementation totals reported by Maryland and Delaware to the Chesapeake Bay Program. This hybrid method combining WorldView-3 and Landsat imagery sources shows promise for monitoring progress in the adoption of conservation tillage practices and for describing crop-residue outcomes associated with a variety of agricultural management practices.
... A k-nn classifier was then trained on these signatures and applied to each image in the pre-storm time series and the post-storm time series. The k-nn classifier was chosen primarily for its ease of implementation, for its exhibition of high accuracies in previous land cover mapping studies [28][29][30] , and for its fast training and inference time (under one minute for each (64GB RAM CPU)). The value of k was set equal to one and the Euclidian distance metric with an equal neighborweighting scheme was chosen for this study. ...
... The value of k was set equal to one and the Euclidian distance metric with an equal neighborweighting scheme was chosen for this study. No fine-tuning of hyper-parameters was performed; instead relying on the strength of previous research that showed these parameters produced high accuracies for land-cover mapping [28][29][30] . It should be noted that tuning these parameters might have slightly increased the accuracy, precision, and recall of the final output maps. ...
... This map depicts the percentage of impervious surface for the entire island for the year 2001 10 .The percentage of impervious surface is correlated to the amount of urban development and total population10,16 . A transfer learning approach similar to the one described inShermeyer and Haack (2015) 28 was devised to leverage the existing impervious surface map, and to create two updated impervious surface maps: pre-storm (January 2016 -August 2017), and post-storm from (September 2017 -May 2018).i B U DI = i x B ...
... A k-nn classifier was then trained on these signatures and applied to each image in the pre-storm time series and the post-storm time series. The k-nn classifier was chosen primarily for its ease of implementation, for its exhibition of high accuracies in previous land cover mapping studies [28][29][30] , and for its fast training and inference time (under one minute for each (64GB RAM CPU)). The value of k was set equal to one and the Euclidian distance metric with an equal neighborweighting scheme was chosen for this study. ...
... The value of k was set equal to one and the Euclidian distance metric with an equal neighborweighting scheme was chosen for this study. No fine-tuning of hyper-parameters was performed; instead relying on the strength of previous research that showed these parameters produced high accuracies for land-cover mapping [28][29][30] . It should be noted that tuning these parameters might have slightly increased the accuracy, precision, and recall of the final output maps. ...
... This map depicts the percentage of impervious surface for the entire island for the year 2001 10 .The percentage of impervious surface is correlated to the amount of urban development and total population10,16 . A transfer learning approach similar to the one described inShermeyer and Haack (2015) 28 was devised to leverage the existing impervious surface map, and to create two updated impervious surface maps: pre-storm (January 2016 -August 2017), and post-storm from (September 2017 -May 2018).i B U DI = i x B ...
Puerto Rico suffered severe damage from the category 5 hurricane (Maria) in September 2017. Total monetary damages are estimated to be ~92 billion USD, the third most costly tropical cyclone in US history. The response to this damage has been tempered and slow moving, with recent estimates placing 45% of the population without power three months after the storm. Consequently, we developed a unique data-fusion mapping approach called the Urban Development Index (UDI) and new open source tool, Comet Time Series (CometTS), to analyze the recovery of electricity and infrastructure in Puerto Rico. Our approach incorporates a combination of time series visualizations and change detection mapping to create depictions of power or infrastructure loss. It also provides a unique independent assessment of areas that are still struggling to recover. For this workflow, our time series approach combines nighttime imagery from the Suomi National Polar-orbiting Partnership Visible Infrared Imaging Radiometer Suite (NPP VIIRS), multispectral imagery from two Landsat satellites, US Census data, and crowd-sourced building footprint labels. Based upon our approach we can identify and evaluate: 1) the recovery of electrical power compared to pre-storm levels, 2) the location of potentially damaged infrastructure that has yet to recover from the storm, and 3) the number of persons without power over time. As of May 31, 2018, declined levels of observed brightness across the island indicate that 13.9% +/- ~5.6% of persons still lack power and/or that 13.2% +/- ~5.3% of infrastructure has been lost. In comparison, the Puerto Rico Electric Power Authority states that less than 1% of their customers still are without power.
... A year 2000 forest/non-forest mask derived from methods similar to those described in [38] was used to mask out low forest cover pixels. Six of the most cloud-free Landsat TM and ETM+ images from the year 2000 were acquired and cloud masked. ...
Landsat time series data were used to characterize forest degradation in Lam Dong Province, Vietnam. We conducted three types of image change analyses using Landsat time series data to characterize the land cover changes. Our analyses concentrated on the timeframe of 1973–2014, with much emphasis on the latter part of that range. We conducted a field trip through Lam Dong Province to develop a better understanding of the ground conditions of the region, during which we obtained many photographs of representative forest sites with Global Positioning System locations to assist us in our image interpretations. High-resolution Google Earth imagery and Landsat data of the region were used to validate results. In general, our analyses indicated that many land-use changes have occurred throughout Lam Dong Province, including gradual forest to non-forest transitions. Recent changes are most marked along the relatively narrow interfaces between agricultural and
forest areas that occur towards the boundaries of the province. One important observation is that the most highly protected national reserves in the region have not changed much over the entire Landsat timeframe (1972–present). Spectral changes within these regions have not occurred at the same levels as those areas adjacent to the reserves.
... CD is of great significance in the field of land use and land cover investigation, resource survey, urban expansion monitoring, environment assessment, and rapid response to disaster events. [1][2][3][4] In the past decades, numerous CD methods have been proposed, and the investigations can mainly be divided into pixel-based and object-based methods. 5 In the first case, the change features from two images are compared for each pixel independently. ...
In order to fully consider the local spatial constraints between neighboring objects in object-based change detection (OBCD), an OBCD approach is presented by introducing a refined Markov random field (MRF). First, two periods of images are stacked and segmented to produce image objects. Second, object spectral and textual histogram features are extracted and G-statistic is implemented to measure the distance among different histogram distributions. Meanwhile, object heterogeneity is calculated by combining spectral and textual histogram distance using adaptive weight. Third, an expectation-maximization algorithm is applied for determining the change category of each object and the initial change map is then generated. Finally, a refined change map is produced by employing the proposed refined object-based MRF method. Three experiments were conducted and compared with some state-of-the-art unsupervised OBCD methods to evaluate the effectiveness of the proposed method. Experimental results demonstrate that the proposed method obtains the highest accuracy among the methods used in this paper, which confirms its validness and effectiveness in OBCD. © 2017 Society of Photo-Optical Instrumentation Engineers (SPIE).
... CD is of great significance in the field of land use and land cover investigation, resource survey, urban expansion monitoring, environment assessment, and rapid response to disaster events. [1][2][3][4] In the past decades, numerous CD methods have been proposed, and the investigations can mainly be divided into pixel-based and object-based methods. 5 In the first case, the change features from two images are compared for each pixel independently. ...
Object-based change detection method using refined Markov random field, " Abstract. In order to fully consider the local spatial constraints between neighboring objects in object-based change detection (OBCD), an OBCD approach is presented by introducing a refined Markov random field (MRF). First, two periods of images are stacked and segmented to produce image objects. Second, object spectral and textual histogram features are extracted and G-statistic is implemented to measure the distance among different histogram distributions. Meanwhile, object heterogeneity is calculated by combining spectral and textual histogram distance using adaptive weight. Third, an expectation-maximization algorithm is applied for determining the change category of each object and the initial change map is then generated. Finally, a refined change map is produced by employing the proposed refined object-based MRF method. Three experiments were conducted and compared with some state-of-the-art unsuper-vised OBCD methods to evaluate the effectiveness of the proposed method. Experimental results demonstrate that the proposed method obtains the highest accuracy among the methods used in this paper, which confirms its validness and effectiveness in OBCD.
... Where: K = kappa statistic The total number of pixels occupied by each class was converted to hectares for easy interpretation. Given that the Landsat images obtained were at a resolution of 30 m, each pixel was taken to represent an area of 900 m 2 (Shermeyer and Haack 2015). This was multiplied by the total number of pixels covered by the entire municipality for each year and the product converted to hectares. ...
The study was aimed at assessing the dynamics of land conversions for urban development and their impact on environmental planning in addition to assessing the characteristics of urban sprawl in Mbarara Municipality. To determine the dynamics of land conversion for urban development in Mbarara Municipality since 1984, Landsat images for the years 1984, 1999, and 2014 were classified using multi-spectral classification techniques to enable the creation of land cover maps. Population and built-up area density were used as a measure of sprawl for Mbarara Municipality. The built-up area had increased by 107 % between 1984 and 1999 and by 37 % between 1999 and 2014 while the overall growth of built-up area between 1984 and 2014 was found out to be 182 %. This variation in growth is attributed to the introduction of environmental controls and policies that largely checked the rate of growth between 1999 and 2014. The overall growth has affected the size of the area covered by other land uses which were seen to greatly fluctuate over the years. The characteristics of urban sprawl in Mbarara Municipality typically depict strip, cluster, and leapfrog sprawl. Based on the Organization for Economic Cooperation and Development (OECD) sprawl index, Mbarara Municipality was found to have sprawled at a rate of 7.7 % between 1984 and 1999 and −7.6 % between 1999 and 2014. The overall sprawl rate between 1984 and 2014 was −4.3 %. The study suggests that smart growth strategies, upholding zoning practices and the enactment of laws to check illegal land conversions are important to check sprawl.
