On the Application of Remote Sensing Time Series Analysis for Land Cover Mapping: Spectral Indices for Crops Classification (original) (raw)
Related papers
Land cover data from Landsat single-date archive imagery: an integrated classification approach
Proceedings of SPIE, 2013
The analysis of land cover dynamics provides insight into many environmental problems. However, there are few data sources which can be used to derive consistent time series, remote sensing being one of the most valuable ones. Due to their multi-temporal and spatial coverage needs, such analysis is usually based on large land cover datasets, which requires automated, objective and repeatable procedures. The USGS Landsat archives provide free access to multispectral, high-resolution remotely sensed data starting from the mid-eighties; in many cases, however, only single date images are available. This paper suggests an objective approach for generating land cover information from 30m resolution and single date Landsat archive satellite imagery. A procedure was developed integrating pixel-based and object-oriented classifiers, which consists of the following basic steps: i) pre-processing of the satellite image, including radiance and reflectance calibration, texture analysis and derivation of vegetation indices, ii) segmentation of the pre-processed image, iii) its classification integrating both radiometric and textural properties. The integrated procedure was tested for an area in Sardinia Region, Italy, and compared with a purely pixel-based one. Results demonstrated that a better overall accuracy, evaluated against the available land cover cartography, was obtained with the integrated (86%) compared to the pixel-based classification (68%) at the first CORINE Land Cover level. The proposed methodology needs to be further tested for evaluating its trasferability in time (constructing comparable land cover time series) and space (for covering larger areas).
5ο Πανελλήνιο Γεωγραφικό Συνέδριο, 1999
Remote Sensing makes possible to monitor and observe extensive geographical areas economically Unlike Photogrammetric data, Remote sensing data are only in digital form. These data are uscd l()r the measurement of both geometric and thematic properties of earth's environment from a distance Thematic maps are currently used in many studies such as environmental monitoring. tceton\( fault/fold observations, etc. Thematic maps are generated in two ways: either manually hy photointel1'retation, or automatically using a classification procedure. The objective of this paper i, !n use both ways to produce a land cover map, evaluate and comment the resulting maps. For the plIrpo,e of this paper we used a part of a multispectral LandSat TM imagery over the Greek Island of Le~\ \), To perform the photointel1'retation task the MAPINFO GIS was used. The automatic land lI'C classification was carried out using the "Artificial Neural Networks for GEOSIS Inlegr~llL'd Environment (ANNGIE)", developed by National Research Center "DEMOKRITOS" in tilL· framework of Geonickel BRITE EURAM project. Results show that both approaches of thcmatic 111-1p generation have their own purpose and are often complementary.
Remote Sensing and Spatial Analysis for Land-Take Assessment in Basilicata Region (Southern Italy)
Remote Sensing
Land use is one of the drivers of land-cover change (LCC) and represents the conversion of natural to artificial land cover. This work aims to describe the land-take-monitoring activities and analyze the development trend in test areas of the Basilicata region. Remote sensing is the primary technique for extracting land-use/land-cover (LULC) data. In this study, a new methodology of classification of Landsat data (TM–OLI) is proposed to detect land-cover information automatically and identify land take to perform a multi-temporal analysis. Moreover, within the defined model, it is crucial to use the territorial information layers of geotopographic database (GTDB) for the detailed definition of the land take. All stages of the classification process were developed using the supervised classification algorithm support vector machine (SVM) change-detection analysis, thus integrating the geographic information system (GIS) remote sensing data and adopting free and open-source software a...
Accuracy assessment procedures are usually recommended in land-use/land-cover change studies to ascertain the quality of thematic maps produced by means of automated image classification methods. In this regard, the error matrix and accompanying accuracy measures (AMs) are widely used. Concerning thematic maps of crop-shelter coverage (CSC), highresolution remote sensing images have been found to be suitable for the automated classification of greenhouses and several methodologies have been developed. However, methods to report the thematic map accuracy derived from these classifications still need to be standardized. Therefore, this study was undertaken with the aim to provide guidelines for determining a set of AMs suitable to report the accuracy of CSC thematic maps produced at local scale. In the case study, QuickBird RGB-band layers were combined with textural information computed on degraded QuickBird panchromatic layer to perform CSC supervised classifications in a study area located in South-Eastern Sicily (Italy) where protected cultivation is widespread. Statistical tests were performed on a number of AMs computed on the error matrices obtained from the automated image classifications of three different areas selected within the study area.
2007
Remote sensing data provide required inputs for environmental modelling at regional to global scales and play an important role in monitoring agricultural land use with regard to EU agricultural subsidy and sustainable watershed management. To characterise the dynamics of agricultural parcels, it is important to have information about the crop type and the vegetation cover (VC) during the year. In this study, the potential of multitemporal high resolution satellite-images (SPOT) for monitoring agricultural land use was evaluated. The study area is located in the Eifel region (49°8’N, 6°5’E), in the South-West of Germany. Data sets used in this study cover the growing season between April and September 2003 and comprise five satellite scenes from SPOT 4 and SPOT 5 systems. The data was geometrically and radiometrically corrected using standard procedures. To ensure which crop types can be separated with the available image data, a Spectral Mixture Analysis (SMA) was executed for ever...
Land Use Cover Mapping of Water Melon and Cereals in Southern Italy
Italian Journal of Agronomy, 2010
The new high-resolution images from the satellites as IKONOS, SPOT5, Quickbird2 give us the opportunity to map ground features, which were not detectable in the past, by using medium resolution remote sensed data (LAND-SAT). More accurate and reliable maps of land cover can then be produced. However, classification procedure with these images is more complex than with the medium resolution remote sensing data for two main reasons: firstly, because of their exiguous number of spectral bands, secondly, owing to high spatial resolution, the assumption of pixel independence does not generally hold. It is then necessary to have a multi-temporal series of images or to use classifiers taking into account also proximal information. The data in this study were (i) a remote sensing image taken by SPOT5 satellite in July 2007 and used to discriminate the water melon cover class and, (ii) three multi-temporal remote sensing images taken by SPOT5 satellite in May, June and July 2008 used to discriminate water melon and cereal crop cover classes. For water melon recognition, providing a single image in 2007, an object-oriented technique was applied instead of a traditional, per pixel technique obtaining an increase of overall accuracy of 15%. In 2008, since it was available a multi-temporal data set, a traditional 'Maximum Likelihood' technique was applied for both water melon and cereal crop cover class. The overall accuracy is greater than 95%.
Use of intra-annual satellite imagery time-series for land cover characterization purposes
2007
Automatic image classification often fails at separating a large number of land cover classes that punctually may present similar spectral reflectances. To improve the classification accuracy in such situations, multi-temporal satellite data has proven to be valuable auxiliary information. In this paper, we present a study exploring the usefulness of intra-annual satellite images timeseries for automatic land cover classification. The reported work aims at producing a land cover classification of continental Portugal from multi-spectral and multi-temporal MODIS satellite images acquired at a spatial resolution of 500 metres for the year 2000. We started our study by performing a single date classification to define the month with the best score as a benchmark to compare with classification accuracies obtained with sets of images from various dates. Then, we considered various combinations of twelve intra-annual image observations (one per month) to quantify the gain when integrating temporal information in the classification process. Curiously, the results we obtained show that multi-temporal information does not significantly improve overall classification accuracy, but in particular it permits to better separate similar land cover classes even if those remain wrongly identified. Surprisingly also, we show that only few (typically 2) dates are sufficient to reach optimal performance of our multi-temporal classifier. In our study we used a Support Vector Machine learning approach.
Land Cover Classification by using Sentinel-2 Images: A case study in the city of Rome
Proceedings of the 2nd World Congress on Civil, Structural, and Environmental Engineering, 2019
Land cover mapping of land area in Mediterranean climate regions from satellite images is not simple, due to the similarity of the spectral characteristics of the urban area and city surroundings. In this study, satellite images from Sentinel 2A by ESA (European Space Agency) were used to classify the land cover of Rome city, Italy. This paper presents two methods aiming at improving the land cover classification accuracy by using multispectral satellite images. The classification process was performed by using two different algorithms, namely: Maximum Likelihood (ML) and Support Vector Machine (SVM). The supervised "Maximum Likelihood" and "Support Vector Machine" classification algorithms available in ENVI (Environment for Visualizing Images) software, were used to detect five land cover classes: urban, forest, water, agriculture and empty land classes. The results show the ML method applied to Sentinel-2A images provides a higher overall accuracy and kappa coefficient than the SVM method. The main reasons are to increase two amounts in the ML method and over the next few years in this study, first: carry out three steps for increase accuracy and kappa with Sieve Classes, Clump Classes and Majority/Minority Analysis. And second reason the most accurate classification for both approaches was allocated to the year 2018, possibly due to the higher image quality and on-time training sample sites compared to previous 2015, 2016, 2017 years. The results of both methods in this study have been compared.
International Journal of Applied Earth Observation and Geoinformation, 2011
Several previous studies have shown that the inclusion of the LST (Land Surface Temperature) parameter to a NDVI (Normalized Difference Vegetation Index) based classification procedure is beneficial to classification accuracy. In this work, the Yearly Land Cover Dynamics (YLCD) approach, which is based on annual behavior of LST and NDVI, has been used to classify an agricultural area into crop types. To this end, a time series of Landsat-5 images for year 2009 of the Barrax (Spain) area has been processed: georeferenciation, destriping and atmospheric correction have been carried out to estimate NDVI and LST time series for year 2009, from which YLCD parameters were estimated. Then, a maximum likelihood classification was carried out on these parameters based on a training dataset obtained from a crop census. This classification has an accuracy of 87% (kappa = 0.85) when crops are subdivided in irrigated and non-irrigated fields, and when cereal crops are aggregated in a single crop, and performs better than a similar classification from Landsat bands only. These results show that a good crop differentiation can be obtained although detailed crop separation may be difficult between similar crops (barley, wheat and oat) due to similar annual NDVI and LST behavior. Therefore, the YLCD approach is suited for vegetation classification at local scale. As regards the assessment of the YLCD approach for classification at regional and global scale, it will be carried out in a further study.
International Journal of Applied Earth Observation and Geoinformation
A key factor in the implementation of productive and sustainable cultivation procedures is the frequent and area-wide monitoring of cropland and grassland. In particular, attention is focused on assessing the actual status, identifying basic trends and mitigating major threats with respect to land-use intensity and its changes in agricultural and semi-natural areas. Here, multi-seasonal analyses based on satellite Earth Observation (EO) data can provide area-wide, spatially detailed and up-to-date geo-information on the distribution and intensity of land use in agricultural and grassland areas. This study introduces an operational, application-oriented approach towards the categorization of agricultural cropland and grassland based on a novel scheme combining multi-resolution EO data with ancillary geo-information available from currently existing databases. In this context, multi-seasonal high (HR) and medium resolution (MR) satellite imagery is used for both a land parcel-based determination of crop types as well as a cropland and grassland differentiation, respectively. In our experimental analysis, two HR IRS-P6 LISS-3 images are first employed to delineate the field parcels in potential agricultural and grassland areas (determined according to the German Official Topographic Cartographic Information System-ATKIS). Next, a stack of seasonality indices is generated based on 5 image acquisitions (i.e., the two LISS scenes and three additional IRS-P6 AWiFS scenes). Finally, a C5.0 tree classifier is applied to identify main crop types and grassland based on the input imagery and the derived seasonality indices. The classifier is trained using sample points provided by the European Land Use/Cover Area Frame Survey (LUCAS). Experimental results for a test area in Germany assess the effectiveness of the proposed approach and demonstrate that a multi-scale and multi-temporal analysis of satellite data can provide spatially detailed and thematically accurate geo-information on crop types and the cropland-grassland distribution, respectively.