Combining spectral with texture features into object-oriented classification in mountainous terrain using advanced land observing satellite image (original) (raw)
Related papers
Remote Sensing, 2019
The paper presents a comparison of the efficacy of several texture analysis methods as tools for improving land use/cover classification in satellite imagery. The tested methods were: gray level co-occurrence matrix (GLCM) features, Laplace filters and granulometric analysis, based on mathematical morphology. The performed tests included an assessment of the classification accuracy performed based on spectro-textural datasets: spectral images with the addition of images generated using different texture analysis methods. The class nomenclature was based on spectral and textural differences and included the following classes: water, low vegetation, bare soil, urban, and two (coniferous and deciduous) forest classes. The classification accuracy was assessed using the overall accuracy and kappa index of agreement, based on the reference data generated using visual interpretation of the images. The analysis was performed using very high-resolution imagery (Pleiades, WorldView-2) and high-resolution imagery (Sentinel-2). The results show the efficacy of selected GLCM features and granulometric analysis as tools for providing textural data, which could be used in the process of land use/cover classification. It is also clear that texture analysis is generally a more important and effective component of classification for images of higher resolution. In addition, for classification using GLCM results, the Random Forest variable importance analysis was performed.
Comparing pixel based and object based approaches in land use classification in mountainous areas
Remote sensing imagery needs to be converted into tangible information which can be utilized in conjunction with other data sets, often within widely used Geospatial Information Systems (GIS). Remote sensing data help in mapping land resources, especially in mountainous areas where accessibility is limited. Classification of remote sensing data in mountainous terrain is problematic because of variations in the sun illumination angle. Traditional approaches have many problems in these conditions. In object based approach can utilized GIS tools for improvement of classification results. In the present work we used pixel based and object based approaches that in both we imported GIS concepts and ancillary data for refining of classification results. The results showed that object based approach have higher accuracy than pixel based approach.
International Journal of Environment and Geoinformatics
There are several remotely sensed images of varied resolutions available. As a result, several classification techniques exist, which are roughly classified as pixel-based and object-based classification methods. Based on the foregoing, this study provided an integrated method of deriving land use from a coarse satellite image. Location coordinates of the land uses were acquired with a handheld Global Positioning System (GPS) instrument as primary data. The study classified the image quantitatively (pixel-based) into built-up, water, riparian, cultivated, and uncultivated land cover classes with no mixed pixels, and then qualitatively into educational, commercial, health, residential, and security land use classes that were conflicting due to spectral similarity. The total accuracy and kappa coefficient of the pixel-based land cover classification were 92.5% and 94% respectively. The results showed that residential land use occupied an area of 5500.01ha, followed by educational (280...
A textural approach for land cover classification of remotely sensed image
Texture features play a vital role in land cover classification of remotely sensed images. Local binary pattern (LBP) is a texture model that has been widely used in many applications. Many variants of LBP have also been proposed. Most of these texture models use only two or three discrete output levels for pattern characterization. In the case of remotely sensed images, texture models should be capable of capturing and discriminating even minute pattern differences. So a multivariate texture model is proposed with four discrete output levels for effective classification of land covers. Remotely sensed images have fuzzy land covers and boundaries. Support vector machine is highly suitable for classification of remotely sensed images due to its inherent fuzziness. It can be used for accurate classification of pixels falling on the fuzzy boundary of separation of classes. In this work, texture features are extracted using the proposed multivariate descriptor, MDLTP/MVAR that uses multivariate discrete local texture pattern (MDLTP) supplemented with multivariate variance (MVAR). The classification accuracy of the classified image obtained is found to be 93.46 %.
Texture analysis of AIRSAR images for land cover classification
Proceeding of the 2011 IEEE International Conference on Space Science and Communication (IconSpace), 2011
Remote sensing technology and the advance of science are very useful information in land cover classification study. In this study, data from an airborne radar system, AIRSAR (Airborne Synthetic Aperture Radar) containing C, P and L bands each with HH, VV and HV polarizations were used to identify land cover features in two study areas in Kedah. The main objective of this study was to investigate the performance of each band and polarization for land cover classification by applying supervised classifiers. Texture measure such as VI, VA, VL, U, homogeineity, contrast, mean, standard deviation, entropy, angular second moment, GLDV angular second moment, GLDV entropy, mean Euclidean distance, skewness, kurtosis, energy, lacunarity and semivariogram were applied. For each measure, signature separability between classes selected from training areas was determined using Bhattacharyya distance. The texture measures were then used for supervised classification using Maximum Likelihood Classifier (MLC) and K Nearest Neighbor (kNN) classifiers. Accuracy assessment for each measure was carried out using random ground samples.
2017
The main focus of this study is to measure accuracy of classification for combined spectral features and textural features of fused images obtained after applying different fusion techniques. Study area is selected with a variety of land cover features so as to understand effect of fusion on different land cover features. IHS, GS, PC, CN and Brovey fusion methods are used. A different layer stacked method has been used to create a composite image of bands 5, 4 and 3 of Landsat8, since they show maximum spectral reflectance variations in the land features present in the selected area of study. Resampling is carried out using Nearest Neighbor and Cubic Convolution, where Nearest Neighbor gave better accuracy of fused image. Classification is performed using spectral features of fused images, textural features of fused images and composite images created using spectral and textural features. Fused images are assessed for distortion using visual analysis and statistical parameters. Clas...
Abstract: Automated features extraction techniques have gone a long way to ameliorate pains on imageprocessing and information extraction from remotely sensed image; this has opened many doors for manyapplications. However, pixel based image classification and feature extraction of Nigerian urban areas are stillvery difficult due to the spectral similarity of many of the urban features such as roads, bare ground, rockoutcrop, built-up areas and many more. This is as a result of the nature of the roads (untarred) and otherfeatures like built-up areas with large area coverage of bare ground rather than well landscaped with grassgardens. This makes high classification accuracy of urban feature extraction very difficult to achieve. Theperformance of urban feature extraction from LANDSAT imagery of Abuja, Nigeria with eCognition isoverwhelming. This research compared the result of object oriented image classification (i.e. eCognition) andother traditional (i.e. Erdas Imagine) methods of image classification vis-à-vis urban feature extraction frommedium resolution remotely sensed image of Abuja Nigeria which has a very highly similar spectral reflectancevalue. The result shows overall accuracy of 94.9% as against the overall accuracy of 89% in using thetraditional method in Erdas Imagine. While the ever difficult and very confused class differentiation in imageclassification in Nigerian urban areas between built-up and bare ground was clearly separated with high levelof accuracy Keyword: Pixel Based Classification, Object-Oriented Classification, Automatic Feature Extraction, eCognition.
International Journal of Remote Sensing, 2003
Object-oriented classification techniques based on image segmentation are gaining interest as methods for producing output maps directly storable into Geophysical Information System (GIS) databases. A limitation in efficiently applying image segmentation is often represented by the spatial resolution of the image. This contribution proposes a method for overcoming this problem, based on the integrated use of images of different resolution. A high-resolution black and white (b/w) orthophoto and a subscene of a Landsat Thematic Mapper (TM) image have been used to obtain an object-oriented classification of the land cover of a study area in northern Italy. The method consists of a sequential application of segmentation and classification techniques. First, the TM image was classified using the maximum likelihood classifier and additional empirical rules. Subsequently, the orthophoto was segmented by applying a region-based segmentation algorithm. Finally, the classification of the segmented images was performed using as a reference the TM image previously classified. The resulting land cover map was tested for accuracy and the results are dicusssed.
An integrated classification method for thematic mapper imagery of plain and highland terrains
Journal of Zhejiang University SCIENCE A, 2008
The classification of thematic mapper imagery in areas with strong topographic variations has proven problematic in the past using a single classifier, due to the changing sun illumination geometry. This often results in the phenomena of identical object with dissimilar spectrum and different objects with similar spectrum. In this paper, an integrated classification method that combines a decision tree with slope data, tasseled cap transformation indices and maximum likelihood classifier is introduced, to find an optimal classification method for thematic mapper imagery of plain and highland terrains. A Landsat 7 ETM+ image acquired over Hangzhou Bay, in eastern China was used to test the method. The results indicate that the performance of the integrated classifier is acceptably good in comparison with that of the existing most widely used maximum likelihood classifier. The integrated classifier depends on hypsography (variation in topography) and the characteristics of ground truth objects (plant and soil). It can greatly reduce the influence of the homogeneous spectrum caused by topographic variation. This integrated classifier might potentially be one of the most accurate classifiers and valuable tool for land cover and land use mapping of plain and highland terrains.