USGS Science⌕ Search
Home/Archive/USGS

USGS · 70023862

AVHRR composite period selection for land cover classification

S.K. Maxwell·R.M. Hoffer·P.L. Chapman
View original source ↗Export metadata ↓

Abstract

Multitemporal satellite image datasets provide valuable information on the phenological characteristics of vegetation, thereby significantly increasing the accuracy of cover type classifications compared to single date classifications. However, the processing of these datasets can become very complex when dealing with multitemporal data combined with multispectral data. Advanced Very High Resolution Radiometer (AVHRR) biweekly composite data are commonly used to classify land cover over large regions. Selecting a subset of these biweekly composite periods may be required to reduce the complexity and cost of land cover mapping. The objective of our research was to evaluate the effect of reducing the number of composite periods and altering the spacing of those composite periods on classification accuracy. Because inter-annual variability can have a major impact on classification results, 5 years of AVHRR data were evaluated. AVHRR biweekly composite images for spectral channels 1-4 (visible, near-infrared and two thermal bands) covering the entire growing season were used to classify 14 cover types over the entire state of Colorado for each of five different years. A supervised classification method was applied to maintain consistent procedures for each case tested. Results indicate that the number of composite periods can be halved-reduced from 14 composite dates to seven composite dates-without significantly reducing overall classification accuracy (80.4% Kappa accuracy for the 14-composite data-set as compared to 80.0% for a seven-composite dataset). At least seven composite periods were required to ensure the classification accuracy was not affected by inter-annual variability due to climate fluctuations. Concentrating more composites near the beginning and end of the growing season, as compared to using evenly spaced time periods, consistently produced slightly higher classification values over the 5 years tested (average Kappa) of 80.3% for the heavy early/late case as compared to 79.0% for the alternate dataset case).

Explore related subjects

International Journal of Remote SensingArticle

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

S.K. Maxwell, R.M. Hoffer, P.L. Chapman. 2010-11-25. AVHRR composite period selection for land cover classification. https://doi.org/10.1080/01431160210145579

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Bridging remote sensing advances and management needs for small Prairie Pothole waterbodies using a multiscale accuracy assessment

Remote sensing of surface water provides a powerful tool to inform the management of waterfowl habitat, but there is little information available to directly assess the relative accuracy of different remote sensing datasets. Our objective was to understand how the characteristics of remotely sensed inundation datasets inform dataset accuracy, the reliable detection of small waterbodies, the distribution of surface water, observation frequency and completeness of data: all attributes relevant to the management of waterfowl habitat. We compared surface water composites from 10 remotely sensed surface water datasets from 2016 to 2021 to in-situ surface water data for three complexes of small waterbodies in the U.S. Prairie Pothole Region and evaluated their accuracy at the pixel, waterbody and local landscape scales. While all products had high per-pixel balanced accuracies (>0.75), we found distinct differences in waterbody area and landscape distribution estimates among datasets. Sentinel-1-based datasets provided a more complete set of observations over time and were more sensitive in detecting water presence in smaller waterbodies but were less accurate at identifying waterbody area than other datasets evaluated. Landsat datasets, alternatively, produced simpler landscape distributions that largely omitted the smallest waterbodies. While the datasets that either fused Sentinel-1 and −2 data collections or utilized local training data had the highest performances (e.g. balanced accuracy = 0.92), all datasets had use-case scenarios for which they may be informative. Our comparisons revealed differences that were not evident in traditional pixel-scale accuracy assessment, such as an 18-fold difference in the number of inundated waterbodies identified across remote sensing datasets. These findings provide novel insights for waterfowl conservation management on howremote sensing datasets may differ in their ability to monitor annual spring surface water presence within landscapes dominated by small waterbodies, such as the Prairie Pothole Region.

Minnesota, North Dakota↗
USGS2026

Sentinel-2 for chlorophyll-a water quality monitoring: A review of validation evidence and application potential

Water quality monitoring is integral to preserving the health of freshwater ecosystems, and satellite remote sensing has emerged as one monitoring method. Sentinel-2, in particular, has been valuable for water quality monitoring due to its 5-day global temporal revisit time and spatial resolution that ranges from 10 to 60 metres. Sentinel-2 can be used to measure and monitor chlorophyll-a, which historically has been used as an indicator of water quality, eutrophication and harmful algal blooms. Our goal was to review aquatic chlorophyll-a Sentinel-2 research to assess the types of validation evidence reported. Validation evidence is defined here as the set of information key to assessing algorithm performance, and include the spatial and temporal scales of satellite validation, reported in situ sampling method context information, demonstration of validation results through plots, and appropriate algorithm performance metrics. We highlight how the body of literature collectively contributes to advancing a national scale chlorophyll-a product that could support future resource management applications. Our review of 122 published studies indicated that much of the validation evidence corresponded to early stages, as defined by the NASA data maturity framework, due to a limited focus on individual lakes and limited detail on methodology for reproducibility. Prioritizing data accessibility for both in situ data and satellite workflows used in published studies; reporting methods with transparency and consistency; and using standard algorithm performance metrics could provide a consistent framework to support and enhance the utility of satellite inland water quality research. These three quality assurance mechanisms can promote effective evaluation of approaches for remote sensing of chlorophyll-a. Adopting these quality criteria could enable the integration of validation evidence from multiple studies, supporting more spatially and temporally representative products that would advance these approaches towards maturation for broader application.

International Journal of Remote Sensing↗
USGS2025

Deep learning approaches on Landsat observations: A review and meta-analysis

Although multiple summary reviews of deep learning in remote sensing have been conducted, the field currently lacks a comprehensive analysis of deep learning utilizing Landsat observations in single time pixel-based classification, both qualitatively and quantitatively. This study addresses this gap by conducting a systematic review based on 352 case studies extracted from 279 peer-reviewed articles that supported direct comparisons across deep learning and traditional classifiers. Qualitatively, our findings are as follows: (i) Two dimensional Convolutional Neural Networks (CNNs) remain the dominant methods; however, exploring other architectures is becoming a growing trend. (ii) Most studies focused on single countries (e.g., China, USA, and India) and the last decade, (iii) Sentinel-2 are the most commonly fused sensors, and vegetation indices are the most commonly used spectral indices. The quantitative meta-analysis revealed that Deep Neural Networks (DNNs) offer an approximate 5% improvement for non-DNN accuracy of 80%, with potentially higher improvements for more challenging tasks. By contrasting between popular DNN methods, UNets achieve higher overall accuracy (OA) than CNNs (approximately 10% UNet improvement when CNN OA is at 80%). However, UNets are more challenging to implement because they require labels for large contiguous reference patches, while CNNs, at least at 30m Landsat pixel, pose lower spatial requirements. In another comparison, Transformers achieve higher OA than CNNs (approximately 5% Transformer improvement when CNN OA is at 80%), although the accuracy improvements are smaller compared to the increases associated with UNet benefits. With greater flexibility and more standardized architecture, Transformers can potentially simplify the design process of models. However, Transformers also face challenges having higher computational resource and memory requirements. Increasing availability of reference datasets, further algorithmic developments in the machine learning field, and improvements in cloud-based computing power and data storage, could increase the utility of deep learning methods on Landsat observations.

International Journal of Remote Sensing↗