MODIS Collection 5 global land cover: Algorithm refinements and characterization of new datasets

Remote Sensing of Environment, Journal Year: 2009, Volume and Issue: 114(1), P. 168 - 182

Published: Oct. 15, 2009

Language: Английский

---

A survey of image classification methods and techniques for improving classification performance

International Journal of Remote Sensing, Journal Year: 2007, Volume and Issue: 28(5), P. 823 - 870

Published: March 1, 2007

Image classification is a complex process that may be affected by many factors. This paper examines current practices, problems, and prospects of image classification. The emphasis placed on the summarization major advanced approaches techniques used for improving accuracy. In addition, some important issues affecting performance are discussed. literature review suggests designing suitable image‐processing procedure prerequisite successful remotely sensed data into thematic map. Effective use multiple features selection method especially significant Non‐parametric classifiers such as neural network, decision tree classifier, knowledge‐based have increasingly become multisource Integration remote sensing, geographical information systems (GIS), expert system emerges new research frontier. More research, however, needed to identify reduce uncertainties in chain improve

Language: Английский

---

Change detection techniques

International Journal of Remote Sensing, Journal Year: 2004, Volume and Issue: 25(12), P. 2365 - 2401

Published: May 27, 2004

Timely and accurate change detection of Earth's surface features is extremely important for understanding relationships interactions between human natural phenomena in order to promote better decision making. Remote sensing data are primary sources extensively used recent decades. Many techniques have been developed. This paper summarizes reviews these techniques. Previous literature has shown that image differencing, principal component analysis post-classification comparison the most common methods detection. In years, spectral mixture analysis, artificial neural networks integration geographical information system remote become applications. Different algorithms their own merits no single approach optimal applicable all cases. practice, different often compared find best results a specific application. Research still an active topic new needed effectively use increasingly diverse complex remotely sensed available or projected be soon from satellite airborne sensors. comprehensive exploration major approaches implemented as found literature. Abbreviations this 6S second simulation signal solar spectrum ANN ASTER Advanced Spaceborne Thermal Emission Reflection Radiometer AVHRR Very High Resolution AVIRIS Airborne Visible/Infrared Imaging Spectrometer CVA vector EM expectation–maximization algorithm ERS-1 Earth Resource Satellite-1 ETM+ Enhanced Thematic Mapper Plus, Landsat 7 GIS Geographical Information System GS Gramm–Schmidt transformation J-M distance Jeffries–Matusita KT Kauth–Thomas tasselled cap LSMA linear LULC land cover MODIS Moderate Spectroradiometer MSAVI Modified Soil Adjusted Vegetation Index MSS Multi-Spectral Scanner NDMI Normalized Difference Moisture NDVI NOAA National Oceanic Atmospheric Administration PCA RGB red, green blue colour composite RTB ratio tree biomass total aboveground SAR synthetic aperture radar SAVI SPOT HRV Satellite Probatoire d'Observation de la Terre (SPOT) high resolution visible TM VI

Language: Английский

---

Good practices for estimating area and assessing accuracy of land change

Remote Sensing of Environment, Journal Year: 2014, Volume and Issue: 148, P. 42 - 57

Published: April 12, 2014

Language: Английский

---

Death to Kappa: birth of quantity disagreement and allocation disagreement for accuracy assessment

International Journal of Remote Sensing, Journal Year: 2011, Volume and Issue: 32(15), P. 4407 - 4429

Published: Aug. 10, 2011

Abstract The family of Kappa indices agreement claim to compare a map's observed classification accuracy relative the expected baseline maps that can have two types randomness: (1) random distribution quantity each category and (2) spatial allocation categories. Use has become part culture in remote sensing other fields. This article examines five different indices, some which were derived by first author 2000. We expose indices' properties mathematically illustrate their limitations graphically, with emphasis on Kappa's use randomness as baseline, often-ignored conversion from an sample matrix estimated population matrix. concludes these are useless, misleading and/or flawed for practical applications we seen. After more than decade working recommend profession abandon purposes assessment map comparison, instead summarize cross-tabulation much simpler summary parameters: disagreement disagreement. shows how compute parameters using examples taken peer-reviewed literature. Acknowledgements United States' National Science Foundation (NSF) supported this work through its Coupled Natural Human Systems program via grant BCS-0709685. NSF supplied additional funding Long Term Ecological Research network OCE-0423565 supplemental DEB-0620579. Any opinions, findings, conclusions or recommendation expressed those authors do not necessarily reflect funders. Clark Labs produced GIS software Idrisi, computes components endorses. Anonymous reviewers constructive feedback helped improve article.

Language: Английский

---

Finer resolution observation and monitoring of global land cover: first mapping results with Landsat TM and ETM+ data

International Journal of Remote Sensing, Journal Year: 2012, Volume and Issue: 34(7), P. 2607 - 2654

Published: Dec. 21, 2012

We have produced the first 30 m resolution global land-cover maps using Landsat Thematic Mapper (TM) and Enhanced Plus (ETM+) data. classified over 6600 scenes of TM data after 2006, 2300 ETM+ before all selected from green season. These images cover most world's land surface except Antarctica Greenland. Most these came United States Geological Survey in level L1T (orthorectified). Four classifiers that were freely available employed, including conventional maximum likelihood classifier (MLC), J4.8 decision tree classifier, Random Forest (RF) support vector machine (SVM) classifier. A total 91,433 training samples collected by traversing each scene finding representative homogeneous samples. 38,664 test at preset, fixed locations based on a globally systematic unaligned sampling strategy. Two software tools, Global Analyst developed extending functionality Google Earth, used developing sample databases referencing Moderate Resolution Imaging Spectroradiometer enhanced vegetation index (MODIS EVI) time series for 2010 high Earth. unique classification system was can be crosswalked to existing Nations Food Agriculture Organization (FAO) as well International Geosphere-Biosphere Programme (IGBP) system. Using four algorithms, we obtained initial set maps. The SVM highest overall accuracy (OCA) 64.9% assessed with our samples, RF (59.8%), (57.9%), MLC (53.9%) ranked second fourth. also estimated OCAs subset (8629) which represented area greater than 500 × m. this subset, found OCA 71.5%. As consistent source estimating coverage types world, estimation shows only 6.90% world is planted agricultural production. cropland 11.51% if unplanted croplands are included. forests, grasslands, shrublands 28.35%, 13.37%, 11.49% respectively. impervious covers 0.66% world. Inland waterbodies, barren lands, snow ice 3.56%, 16.51%, 12.81%

Language: Английский

---

Recent advances in techniques for hyperspectral image processing

Remote Sensing of Environment, Journal Year: 2009, Volume and Issue: 113, P. S110 - S122

Published: April 15, 2009

Language: Английский

---

Continuous change detection and classification of land cover using all available Landsat data

Remote Sensing of Environment, Journal Year: 2014, Volume and Issue: 144, P. 152 - 171

Published: Feb. 11, 2014

Language: Английский

---

Remote sensing imagery in vegetation mapping: a review

Journal of Plant Ecology, Journal Year: 2008, Volume and Issue: 1(1), P. 9 - 23

Published: March 1, 2008

Mapping vegetation through remotely sensed images involves various considerations, processes and techniques. Increasing availability of due to the rapid advancement remote sensing technology expands horizon our choices imagery sources. Various sources are known for their differences in spectral, spatial, radioactive temporal characteristics thus suitable different purposes mapping. Generally, it needs develop a classification at first classifying mapping cover from either community level or species level. Then, correlations types (communities species) within this system with discernible spectral have be identified. These classes finally translated into image interpretation process, which is also called processing. This paper presents an overview how use classify map cover. Specifically, focuses on comparisons popular sensors, commonly adopted processing methods prevailing accuracy assessments. The basic concepts, available techniques related were introduced, analyzed compared. advantages limitations using provided iterate importance thorough understanding concepts careful design technical procedures, can utilized study images.

Language: Английский

---

Optical remotely sensed time series data for land cover classification: A review

ISPRS Journal of Photogrammetry and Remote Sensing, Journal Year: 2016, Volume and Issue: 116, P. 55 - 72

Published: March 23, 2016

Accurate land cover information is required for science, monitoring, and reporting. Land changes naturally over time, as well a result of anthropogenic activities. Monitoring mapping change in consistent robust manner large areas made possible with Earth Observation (EO) data. products satisfying range science policy needs are currently produced periodically at different spatial temporal scales. The increased availability EO data—particularly from the Landsat archive (and soon to be augmented Sentinel-2 data)—coupled improved computing storage capacity novel image compositing approaches, have resulted annual, large-area, gap-free, surface reflectance data products. In turn, these support development annual that can both informed constrained by detection outputs. inclusion time series process provides on class stability informs logical transitions (both temporally categorically). this review, we present issues opportunities associated generating validating time-series products, identify methods suited incorporating other inputs characterization.

Language: Английский

---