Workflow of Processing Landsat data for landcover classification

Workflow of Processing Landsat data for landcover classification

I have downloaded Landsat-8 data to make a land-use raster. So far I've made a composite image and clipped it by my study area. As you can see, there is some slight cloud cover. I've gone ahead and downloaded L-LDOPE Toolbelt and the same tile during earlier times in the growing season to possibly interpolate the area covered by clouds.

I'm unsure of what to do next. Should I go ahead and run image classification then go back and account for the clouds or should I deal with the clouds first? What is the standard work flow for processing these images? Does anyone know of good background reading on this? I'm very new to working with Landsat-8 data. I'm using ArcGIS 10.1.

Usually you start with cloud (and cloud shadow) removal then you run the classification. One of the best papers I know about cloud detection on Landsat is Zhu and Woodcock (2012)

Workflow of Processing Landsat data for landcover classification - Geographic Information Systems

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited.

Feature Papers represent the most advanced research with significant potential for high impact in the field. Feature Papers are submitted upon individual invitation or recommendation by the scientific editors and undergo peer review prior to publication.

The Feature Paper can be either an original research article, a substantial novel research study that often involves several techniques or approaches, or a comprehensive review paper with concise and precise updates on the latest progress in the field that systematically reviews the most exciting advances in scientific literature. This type of paper provides an outlook on future directions of research or possible applications.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to authors, or important in this field. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Related Content

March 1, 2019 – Remote Sensing Special Issue Highlights Landsat ARD

Remote Sensing, an open access journal, recently published a special issue titled Science of Landsat Analysis Ready Data (ARD).

17. Image Classification

Back in Chapter 3, we considered the classification of thematic data for choropleth maps. Remember? We approached data classification as a kind of generalization technique, and made the claim that "generalization helps make sense of complex data." The same is true in the context of remotely sensed image data.

A key trend in image classification is the emergence of object-based alternatives to traditional pixel-based techniques. A Penn State lecturer has observed, "For much of the past four decades, approaches to the automated classification of images have focused almost solely on the spectral properties of pixels" (O'Neil-Dunne, 2011). Pixel-based approaches made sense initially, O'Neil-Dunne points out, since "processing capabilities were limited and pixels in the early satellite images were relatively large and contained a considerable amount of spectral information." In recent years, however, pixel-based approaches have begun to be overtaken by object-based image analysis (OBIA) for high-resolution multispectral imagery, especially when fused with lidar data. OBIA is beyond the scope of this chapter, but you can study it in depth in the open-access Penn State courseware GEOG 883: Remote Sensing Image Analysis and Applications.

Pixel-based classification techniques are commonly used in land use and land cover mapping from imagery. These are explained below and in the following case study.

The term land cover refers to the kinds of vegetation that blanket the Earth's surface, or the kinds of materials that form the surface where vegetation is absent. Land use, by contrast, refers to the functional roles that the land plays in human economic activities (Campbell, 1983).

Both land use and land cover are specified in terms of generalized categories. For instance, an early classification system adopted by a World Land Use Commission in 1949 consisted of nine primary categories, including settlements and associated non-agricultural lands, horticulture, tree and other perennial crops, cropland, improved permanent pasture, unimproved grazing land, woodlands, swamps and marshes, and unproductive land. Prior to the era of digital image processing, specially trained personnel drew land use maps by visually interpreting the shape, size, pattern, tone, texture, and shadows cast by features shown in aerial photographs. As you might imagine, this was an expensive, time-consuming process. It's not surprising, then, that the Commission appointed in 1949 failed in its attempt to produce a detailed global land use map.

Part of the appeal of digital image processing is the potential to automate land use and land cover mapping. To realize this potential, image analysts have developed a family of image classification techniques that automatically sort pixels with similar multispectral reflectance values into clusters that, ideally, correspond to functional land use and land cover categories. Two general types of pixel-based image classification techniques have been developed: supervised and unsupervised techniques.

Supervised classification

Human image analysts play crucial roles in both supervised and unsupervised image classification procedures. In supervised classification, the analyst's role is to specify in advance the multispectral reflectance or (in the case of the thermal infrared band) emittance values typical of each land use or land cover class.

For instance, to perform a supervised classification of the Landsat Thematic Mapper (TM) data shown above into two land cover categories, Vegetation and Other, you would first delineate several training fields that are representative of each land cover class. The illustration below shows two training fields for each class however, to achieve the most reliable classification possible, you would define as many as 100 or more training fields per class.

The training fields you defined consist of clusters of pixels with similar reflectance or emittance values. If you did a good job in supervising the training stage of the classification, each cluster would represent the range of spectral characteristics exhibited by its corresponding land cover class. Once the clusters are defined, you would apply a classification algorithm to sort the remaining pixels in the scene into the class with the most similar spectral characteristics. One of the most commonly used algorithms computes the statistical probability that each pixel belongs to each class. Pixels are then assigned to the class associated with the highest probability. Algorithms of this kind are known as maximum likelihood classifiers. The result is an image like the one shown below, in which every pixel has been assigned to one of two land cover classes.

Analysis of Land Use/Land Cover Changes Using Remote Sensing Data and GIS at an Urban Area, Tirupati, India

Land use/land cover (LU/LC) changes were determined in an urban area, Tirupati, from 1976 to 2003 by using Geographical Information Systems (GISs) and remote sensing technology. These studies were employed by using the Survey of India topographic map 57 O/6 and the remote sensing data of LISS III and PAN of IRS ID of 2003. The study area was classified into eight categories on the basis of field study, geographical conditions, and remote sensing data. The comparison of LU/LC in 1976 and 2003 derived from toposheet and satellite imagery interpretation indicates that there is a significant increase in built-up area, open forest, plantation, and other lands. It is also noted that substantial amount of agriculture land, water spread area, and dense forest area vanished during the period of study which may be due to rapid urbanization of the study area. No mining activities were found in the study area in 1976, but a small addition of mining land was found in 2003.

1. Introduction

In an urban environment natural and human-induced environmental changes are of concern today because of deterioration of environment and human health [1]. The study of land use/land cover (LU/LC) changes is very important to have proper planning and utilization of natural resources and their management [2]. Traditional methods for gathering demographic data, censuses, and analysis of environmental samples are not adequate for multicomplex environmental studies [3], since many problems often presented in environmental issues and great complexity of handling the multidisciplinary data set we require new technologies like satellite remote sensing and Geographical Information Systems (GISs). These technologies provide data to study and monitor the dynamics of natural resources for environmental management [4].

Remote sensing has become an important tool applicable to developing and understanding the global, physical processes affecting the earth [5]. Recent development in the use of satellite data is to take advantage of increasing amounts of geographical data available in conjunction with GIS to assist in interpretation [6]. GIS is an integrated system of computer hardware and software capable of capturing, storing, retrieving, manipulating, analyzing, and displaying geographically referenced (spatial) information for the purpose of aiding development-oriented management and decision-making processes [7]. Remote sensing and GIS have covered wide range of applications in the fields of agriculture [8], environments [9], and integrated eco-environment assessment [10]. Several researchers have focused on LU/LC studies because of their adverse effects on ecology of the area and vegetation [11–14].

Present study area witnessed rapid development during past decades in terms of urbanization, industrialization, and also population increase substantially. The main objective of this paper is to detect and quantify the LU/LC in an urban area, Tirupati (Figure 1), from 1976 to 2003 using satellite imagery and topographic map.

2. Study Area Description

The study area, Tirupati region (Figure 1), is located nearby the metropolitan city, Chennai, at a distance of about 145 km in southern peninsular India. Tirupati is a world famous holy pilgrim place for devotees of Lord Sri Venkateswara is situated in Chittoor district of Andhra Pradesh (AP) state at an altitude of 182.9 m (13.05°N latitude and 79.05°E longitude) which represents an urban area surrounded by major industrial and agricultural activities along with dense forest. The town area owes its existence to the sacred world famous temple of Lord Sri Venkateswara situated on the seven hills (Tirumala) adjoining it. The total population of Tirupati region is about 3, 09,000 according to 2001 census of India. Industrial activities have also impact on the overall pollution levels. The major industries are located heavily at Tirupati industrial area situated at the east nearby Renigunta.

The study area covers many water streams, majorly the Swarnamukhi River basin. All the streams including the Swarnamukhi River are ephemeral and rise from the Tirupati hill ranges. The annual rainfall during the study period is 899.8 mm with total number of 43 events, in which the highest rainfall in July (340.6 mm) and the lowest in April (5.6 mm). The streams, while flowing from the upland to lowlands, form steeply dissected valleys often covered with boulders, showing striations. The surface runoff in most of the streams is restricted to a few hours after the rain, while in the Swarnamukhi and Rallakalva Rivers, the flows last for a few days to a few weeks after the rain. Most of the year, they are dry.

3. Data and Methodology

In the present study we have used mainly two types of data. These are topographic map and remote sensing data. The remote sensing data of georeferenced and merged data of LISS III and PAN of IRS ID of 2003 in the digital mode are obtained from the National Remote Sensing Agency (NRSA), Government of India, Hyderabad, and used. The spatial resolutions of LISS III and PAN are 23.5 and 5.8 meters, and spectral resolutions are 4 and 1 meters, respectively.

The topographic map 57 O/6 (1:50,000 scale) is obtained from the Survey of India, Hyderabad, which was surveyed and prepared in 1976 it is converted to digital mode using scanning. The topographic map is georeferenced with longitude and latitudes using the ArcGIS software and spatial analyst tools and demarcated the boundary of study area.

A supervised signature extraction with the maximum likelihood algorithm was employed to classify the digital data of IRS 1D georeferenced and merged LISS III and PAN for land use/land cover mapping for the year 2003. Before the preprocessing and classification of satellite imagery began, an extensive field survey was performed throughout the study area using Global Positioning System (GPS) equipment. This survey was performed in order to obtain accurate locational point data for each land use and land cover class included in the classification scheme as well as for the creation of training sites and for signature generation.

The satellite data was enhanced before classification using histogram equalization in ERDAS Imagine 8.7 to improve the image quality and to achieve better classification accuracy. In supervised classification, spectral signatures are developed from specified locations in the image. These specified locations are given the generic name “training sites” and are defined by the user. Generally a vector layer is digitized over the raster scene. The vector layer consists of various polygons overlaying different land use types. The training sites will help to develop spectral signatures for the outlined areas.

The land use maps pertaining of two different periods were used for postclassification comparison, which facilitated the estimation of changes in the land use category and dynamism with the changes. Postclassification comparison is the most commonly used quantitative method of change detection [15–17] with fairly good results. Postclassification comparison is sometimes referred to as “delta classification” [18]. It involves independently produced spectral classification results from different data sets, followed by a pixel-by-pixel or segment-by-segment comparison to detect changes in the classes. The detailed methodology adopted was given in Figure 2.

4. Results and Discussion

Knowledge about land use/land cover has become important to overcome the problem of biogeochemical cycles, loss of productive ecosystems, biodiversity, deterioration of environmental quality, loss of agricultural lands, destruction of wetlands, and loss of fish and wildlife habitat. The main reason behind the LU/LC changes includes rapid population growth, rural-to-urban migration, reclassification of rural areas as urban areas, lack of valuation of ecological services, poverty, ignorance of biophysical limitations, and use of ecologically incompatible technologies.

Present study area Tirupati is a rapid developing town and is a world famous pilgrim centre for the devotees of Lord Sri Venkateswara. During the past few decades, the study area has witnessed substantial increase in population (Table 1), economic growth, and industrialization, and transportation activities (Table 1) have negative impact on the environmental health of the region.

Due to involvement of multiple data sets, we used latest technologies like remote sensing and GIS to quantify LU/LC. On the basis of interpretation of remote sensing imagery, field surveys, and existing study area conditions, we have classified the study area into eight categories, that is, agriculture, built-up area, dense forest, mining, open forest, other land, plantation, and water spread area (Figures 3 and 4). The study area covers 125 km 2 and LU/LC changes were estimated from 1976 to 2003.

1. Introduction

Deforestation in the Brazilian Amazon since the 1970s has converted a vast area of primary forest into a mosaic of large patches of agricultural lands, pasture, and different stages of successional vegetation (Moran et al. 1994a, Skole et al. 1994, Lucas et al. 2000, Roberts et al. 2002, Lu et al. 2008). The unprecedented tropical deforestation rates have been regarded as an important factor in climate change and environmental degradation at regional and global scales (Skole et al. 1994). In order to better understand the consequences of deforestation and landscape transformations in the region, the timely mapping and monitoring of land use/cover change is required. Remote sensing technologies are useful tools in providing these data sets. Much research has been conducted to classify land cover, especially vegetation classes (Mausel et al. 1993, Moran et al. 1994a, b, Brondízio et al. 1996, Foody et al. 1996, Rignot et al. 1997, Yanasse et al. 1997, Lucas et al. 2002, Vieira et al. 2003, Lu 2005a, Lu et al. 2004a, 2007, 2008). Different classification methods, such as traditional pixel based classifiers (e.g. Euclidean distance and maximum likelihood) (Foody et al. 1996, Yanasse et al. 1997, Vieira et al. 2003), a combination of spectral and spatial information (Mausel et al. 1993, Moran et al. 1994a, b, Lu et al. 2004a), and use of subpixel information (Roberts et al. 1998, Lu et al. 2003) have been examined. Castro et al. (2003) summarizes many approaches using space-borne remotely sensed data to quantify successional forest classification based on biomass or age estimation. Lu (2005a) and Lu et al. (2003, 2008) summarized the major methods for mapping vegetation types, especially successional vegetation stages with remotely sensed data in the moist tropical regions of the Brazilian Amazon.

Previous research has indicated that a major source of confusion often occurs in identifying different successional stages or distinguishing between advanced secondary succession and mature forest (Lu et al. 2003, Lu et al. 2008), since remotely sensed data primarily capture canopy information and the canopy structures between advanced secondary succession and mature forest can be very similar, although they may have different ages, species composition, and biomass density. The smooth transition between different successional stages also causes problems for vegetation classification. Therefore, previous research often provides only coarse vegetation classes, such as primary forest and successional vegetation (Adams et al. 1995, Roberts et al. 2002). However, the biomass densities of different successional stages vary considerably, ranging from less than 2 kg m 𢄢 in initial successional vegetation to greater than 20 kg m 𢄢 in advanced successional vegetation (Lu 2005b). The biomass densities of primary forests also vary considerably, ranging from approximately 12 kg m 𢄢 to greater than 50 kg m 𢄢 in different biophysical environments. Obviously, a single class of primary forest or successional vegetation is not suitable for many applications such as carbon estimations or land degradation assessments.

Remote sensing image classification is a complex process which involves many steps, such as definition of a land cover classification system, collection of data sources (e.g. reference data, different sensor data), extraction of remote sensing variables, selection of classification algorithm, and accuracy assessment (Jensen 2004, Lu and Weng 2007). Great progress in image classification has been achieved, including (1) the development of advanced classification algorithms (e.g. neural network, decision tree, support vector machine, object-based algorithms, and sub-pixel based algorithms) (Tso and Mather 2001, Franklin and Wulder 2002, Lu and Weng 2007, Rogan et al. 2008, Blaschke 2010), (2) the use of multi-source data in a classification process, such as integration of different spatial resolution or sensor images (Solberg et al. 1996, Pohl and Van Genderen 1998, Ali et al. 2009, Ehlers et al. 2010, Zhang 2010) and the integration of remote sensing and ancillary data (Harris and Ventura 1995, Williams 2001, Li 2010), and (3) the development of techniques for modifying classified images by the use of expert knowledge (Stefanov et al. 2001, Hodgson et al, 2003, Zhang et al. 2010).

In practice, Landsat TM images are still the most common data source for land cover classification, even in moist tropical regions due to its suitable spectral and spatial resolutions and long term data availability since the 1970s. Although much research related to land-cover classification has been conducted, a comprehensive analysis of the selection of variables and classification algorithms has not fully investigated. Therefore, this research aims to explore how combinations of different variables can improve land cover classification performance, and which classification algorithm has better classification performance in the moist tropical region of Brazil. This research investigated the roles of vegetation indices and textural images in improving vegetation classification performance based on the comparison of accuracy assessment of the classified images, and compared parametric and nonparametric algorithms in order to understand which classification algorithm is suitable for vegetation classification in the moist tropical region of the Brazilian Amazon, where complex forest stand structure in successional vegetation and primary forest exists. Through this research, we can better understand the classification procedure, including the selection of suitable remote sensing variables and the selection of classification algorithms for the vegetation classification in the Brazilian Amazon.

Workflow of Processing Landsat data for landcover classification - Geographic Information Systems

You have requested a machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Neither SPIE nor the owners and publishers of the content make, and they explicitly disclaim, any express or implied representations or warranties of any kind, including, without limitation, representations and warranties as to the functionality of the translation feature or the accuracy or completeness of the translations.

Translations are not retained in our system. Your use of this feature and the translations is subject to all use restrictions contained in the Terms and Conditions of Use of the SPIE website.

Application of Landsat data to map and monitor agricultural land cover

B. Erdenee, 1 Gegen Tana, 1 Ryutaro Tateishi 1


50 downloads per 1-year subscription

25 downloads per 1 - year subscription

Includes PDF, HTML & Video, when available

Agriculture is one of the major economic sectors of Mongolia and the country's economy is very much dependent on the development of agricultural production. Being the rural and poorest conditions of Mongolia, 60-90% of its labor force employed in agriculture and agricultural sector has a prominent economic role. Mongolian agriculture has been successful in increasing food grains production in the past, guided by the goals of self-sufficiency in the country. The satellite imagery has been effectively utilized for classifying land cover types and detecting land cover conditions. Satellite image classification involves designing and developing efficient image classifiers. With satellite image data and image analysis methods multiplying rapidly, selecting the right mix of data sources and data analysis approaches has become critical to the generation of quality land-use maps. Objective of this study to monitor in the agricultural land cover changes in the Tov aimag, as there is important agricultural producing area in Mongolia. We have developed approaches to map and monitor land cover and land use change across in the Tov aimag using multi-spectral image data. In this study, maximum likelihood supervised classification was applied to Landsat TM and ETM images acquired in 1989 and 2000, respectively, to map cropland area cover changes in the Tov aimag of Mongolia. A supervised classification was carried out on the six reflective bands (bands 1-5 and band 7) for the two images individually with the aid of ground based agricultural monitoring data. Results were then tested using ground check data.

© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

Workflow of Processing Landsat data for landcover classification - Geographic Information Systems


Indonesia has very large mangrove forest area. About 16% of Indonesian mangrove forest area is exist in Kalimantan Island or known as Borneo Island. East Kalimantan is one of four Province in Kalimantan Island that has large area of mangrove forest.This study was conducted to know the existing of mangrove forest in East Kalimantan and also to inventory the degradation of the mangrove forest judging from the mangrove potential area. Mangrove potential area derived from the land system map, and the land system classes that have potential parameters for mangrove living are called as mangrove potential area. The method of this study was to analysis mangrove degradation from Landsat-TM data and Geographic Information Systems (GIS). Land use / land cover derived from Landsat data using Maximum Likelihood Classification. The canopies index vegetation of mangrove derived from Landsat data using Normalized Difference Vegetation Index (NDVI). The result of Land use classification and the canopy index were overlaying with land system digital map, and the field data. The mathematical model for identifying mangrove degradation is TNS = (N x 30)+(Np x 20)+ (L x 15)+(A x 15)+(P x 10)+(C x 10), where TNS is for total score N is for total tree per hectare Np is for total young tree per hectare L is for width of mangrove green belt A is for abrasion P is for pyrite and C is for water pollution.From the data processing result, the mangrove potential area in East Kalimantan is about 759.583,89 ha. Mangrove forest that still exist in East Kalimantan is about 389.426,76 ha, with the area of dense mangrove forest 163.682,82 ha, medium mangrove forest 171.025,83 ha, and sparse mangrove forest 54.718,11 ha. The mangrove potential area that is not degraded is 369.908,19 ha, degraded area 354.814,56 ha, and very degraded area 34.861,14 ha.Judging from the factor that caused the degradation indicated that social economy factor has become the potential factor that caused the degradation. Especially the land use changes between mangrove forest areas become the fishponds.Hlm.791-79

To submit an update or takedown request for this paper, please submit an Update/Correction/Removal Request.

Continental-scale mapping of Adélie penguin colonies from Landsat imagery

Breeding distribution of the Adélie penguin, Pygoscelis adeliae, was surveyed with Landsat-7 Enhanced Thematic Mapper Plus (ETM+) data in an area covering approximately 330° of longitude along the coastline of Antarctica. An algorithm was designed to minimize radiometric noise and to retrieve Adélie penguin colony location and spatial extent from the ETM+ data. In all, 9143 individual pixels were classified as belonging to an Adélie penguin colony class out of the entire dataset of 195 ETM+ scenes, where the dimension of each pixel is 30 m by 30 m, and each scene is approximately 180 km by 180 km. Pixel clustering identified a total of 187 individual Adélie penguin colonies, ranging in size from a single pixel (900 m 2 ) to a maximum of 875 pixels (0.788 km 2 ). Colony retrievals have a very low error of commission, on the order of 1% or less, and the error of omission was estimated to be

3 to 4% by population based on comparisons with direct observations from surveys across east Antarctica. Thus, the Landsat retrievals successfully located Adélie penguin colonies that accounted for

96 to 97% of the regional population used as ground truth. Geographic coordinates and the spatial extent of each colony retrieved from the Landsat data are available publically. Regional analysis found several areas where the Landsat retrievals suggest populations that are significantly larger than published estimates. Six Adélie penguin colonies were found that are believed to be previously unreported in the literature.

Watch the video: How to Download Landsat 8 Atmospherically Corrected Surface Reflectance from EarthExplorer