Oguz Gungor

Integrating multiple texture methods and NDVI to the Random Forest classification algorithm to detect tea and hazelnut plantation areas in northeast Turkey

Hazelnuts and tea are two major agricultural crops grown in the eastern Black Sea region in Turkey. Since this part of Turkey is not industrialized, most of the local people work in agriculture, making hazelnuts and tea a part of their lives. For the government side, it is crucial to keep records of the amount of harvested croplands to implement agricultural policies. In fact, the harvested area and crop type of each cadastral parcel are collected either during cadastral surveys or with the declaration of individual farmers, yet this information is mostly not up-to-date and does not reflect the current land-use status. This study aims to determine the extent and distribution of hazelnuts and tea grown areas using the Random Forest (RF) classification algorithm. Tea and hazelnuts give similar spectral reflectance values to surrounding vegetation, which makes it difficult to distinguish them using only their spectral properties. To tackle this problem, the normalized difference vegetation index (NDVI) and texture extraction methods such as the Grey Level Co-occurrence Matrix (GLCM) and Gabor filter were integrated with the RF algorithm, and their contributions to the success of the RF classification method were examined. WorldView-2 satellite images, which have eight multispectral bands (MS: 2 m) and one higher spatial resolution panchromatic band (PAN: 0.5 m), were used. Since the study area contains agricultural products grown in different seasons, satellite images belonging to both summer and winter periods were used. Preliminary results acquired using only spectral values indicated that the RF method gives 79.05% and 71.84% overall accuracies for summer and winter periods, respectively. Integrating texture information improves the performance of the RF algorithm such that the overall classification accuracies are increased to 83.54% and 87.89% when texture information extracted with GLCM and the Gabor filter is added. The classification performance of the winter image is also boosted to be 77.41% and 79.73% with the contribution of texture information obtained with GLCM and the Gabor filter, respectively. Finally, produced thematic maps were compared with the latest cadastral maps to validate classification results with ground truth data. The obtained results reveal the success of integrating texture features in classification since the created thematic maps are consistent with actual land use. The results also show that the crops grown on some cadastral parcels are not coherent with the most current cadastral database, which implies that the cadastral maps need to be updated.

Classification of multispectral images using Random Forest algorithm

Rastgele Orman algoritmasi kullanilarak cok bantli goruntulerin siniflandirilmasi Rastgele Orman (RO) algoritmasi en basarili siniflandirma yontemlerinden biri olarak bilinir. Dogasi geregi cok farkli disiplinlere hitap etmesinden dolayi, RO farkli alanlarda calisan arastirmacilarin dikkatini cekmektedir. Bu calisma, farkli konumsal cozunurluge ve karakteristige sahip cok bantli uydu goruntuleri kullanarak RO algoritmasinin performansini incelemeyi amaclamaktadir. Kullanilan uydu goruntuleri dort bantli Ikonos ve QuickBird goruntuleridir. 2005 ve 2008 yillarinda elde edilen QuickBird goruntuleri sirasiyla hem kentsel hem de kirsal alanlari kapsarken, 2003 yilinda alinan Ikonos goruntusu, ozellikle kentsel alani icermektedir. Ayrica, 2005 yilinda alinan QuickBird goruntusu ruzgarli havanin yol actigi dalgalar nedeniyle Karadeniz uzerinde gurultulu oruntuler icermektedir. RO’nun performansini degerlendirmek icin siniflandirma sonuclari, Gentle AdaBoost (GAB), En Cok Benzerlik (ECB) ve Destek Vektor Makineleri (DVM) algoritmalarindan elde edilen sonuclarla karsilastirilmistir. Elde edilen sonuclar RO’nun diger yontemlerden daha yuksek siniflandirma dogrulugu verdigini gostermektedir. Kentsel alan uzerinde cekilen Ikonos goruntusune ait sonuclar, RO algoritmasinin, DVM’ den %10 daha yuksek siniflandirma dogrulugu verdigini, GAB algoritmasinin ise en dusuk siniflandirma dogruluguna sahip oldugunu (RO’dan %14 daha dusuk) gostermektedir. Kirsal alan uzerinde alinan QuickBird goruntusune (2008 yilinda alinan)ait sonuclar diger yontemlerden elde edilen sonuclarla karsilastirildiginda RO’ nun daha iyi sonuc verdigi gorulmustur. Gurultuye benzer oruntuler iceren QuickBird goruntusu icin de RO’nun, DVM’den yaklasik %11 daha yuksek siniflandirma dogrulugu verdigi gozlenmistir.

Evaluation of Different Outlier Detection Methods for GPS Networks

GPS (Global Positioning System) devices can be used in many applications which require accurate point positioning in geosciences. Accuracy of GPS decreases due to outliers resulted from the errors inherent in GPS observations. Several approaches have been developed to detect outliers in geodetic observations. It is important to determine which method is most effective at distinguishing outliers from normal observations. This paper investigates the behavior of conventional statistical test methods (Data Snooping (DS), Tau and t tests), some robust methods (Andrewss M-Estimation, Hubers M-Estimation, Tukeys M-Estimation, Danish Method, Yang-I M-Estimation, Yang-II M-Estimation, and fuzzy logic method in detection of outliers for three GPS networks having different characteristics. Test results are evaluated and the performances of different methods are presented quantitatively.

Comparison of the performances of ground filtering algorithms and DTM generation from a UAV-based point cloud

Abstract Ground filtering algorithms mainly focus on filtering LiDAR (Light Detection and Ranging) point clouds owing to their intrinsic characteristics to classify ground and non-ground points. However, the acquisition and processing of LiDAR data is still costly. Compared to LiDAR technology, UAVs (Unmanned Aerial Vehicle) are cheap and easy to use. In this study, the performances of five widely used ground filtering algorithms (Progressive Morphological 1D/2D, Maximum Local Slope, Elevation Threshold with Expand Window, and Adaptive TIN) were investigated by conducting qualitative and quantitative evaluations on UAV-based point clouds. Evaluation results indicated that the Adaptive TIN algorithm presented the best performance. The result of the Adaptive TIN algorithm was interpolated by using a MATLAB script to generate the DTM (Digital Terrain Model). Field measurements indicated that using UAV-based point clouds may be a reasonable alternative for LiDAR data, depending on the characteristics of the study area.

A novel image fusion method using IKONOS satellite images

In satellite remote sensing, spatial resolutions of multispectral images over a particular region can be enhanced using better spatial resolution panchromatic images for the same region by a process called image fusion, or more generally data fusion. A fusion method is considered successful, if the spatial detail of thepanchromatic image is transferred into the multispectral image and the spectral content of the original multispectral image is preserved in the fused product. This research proposes a novel image fusion algorithm which takes aim at producing both spatially enhanced and spectrally appealing fused multispectral images.In the proposed method, first an intermediary image is created using original panchromatic and multispectral images. This intermediary image contains the high frequency content of the panchromatic source image such that it is the one closest to the given multispectral source image (upsampled) by a natural semi innerproduct defined. The final fused image is obtained by applying a function which performs convex linear combination of the intermediary image and the upsampled multispectral image. The function used depends on the local standard deviations of the source images. To test the performance of the method, the imagesfrom IKONOS sensor are fused using the Brovey, IHS, PCA, wavelet transform based methods, and the proposed method. Both visual and quantitative evaluation results indicate that the proposed method yields to both spectrally and spatially appealing results as the wavelet transform based method, and it gives a betterperformance when both spatial detail enhancement and spectral content preservation in the fused products are considered. It is also obvious that the method has a potential to get better results if a better fitting, more complex function is found.

Investigating the performances of commercial and non-commercial software for ground filtering of UAV-based point clouds

ABSTRACT In recent years, the advent of unmanned aerial vehicles (UAV)-based photogrammetry has enabled the collection of accurate and comprehensive information from the surface of the Earth. Owing to low-altitude flights, it is possible to generate high-density point clouds, which are useful for accurate representation of topography of the land surface. Ground filtering is the removal of the points belonging to above-ground objects in order to retrieve ground points to be used in generating digital terrain models. It is essential in most applications for modelling the environment and is performed by using various types of commercial and non-commercial software. This study investigates the performances of seven widely used ground filtering algorithms on UAV-based point clouds. These algorithms are (1) the adaptive triangulated irregular network implemented into the commercial Agisoft Photoscan Professional software, (2) the multi-scale curvature classification implemented into the commercial global mapper software, (3) the cloth simulation filtering (CSF) applied with a MATLAB script, (4) the interpolation-based Boise Centre Aerospace Laboratory-lidar algorithm embedded in the commercial environment for visualizing images software, (5) the interpolation-based gLiDAR non-commercial software, (6) the 2D progressive morphological algorithms, and (7) elevation threshold with expand window algorithms embedded in the non-commercial airborne lidar data processing and analysis tools software. The results showed that the CSF algorithm presented the best filtering results for both test sites with Total Errors of 6% and 4.5% in the sites 1 and 2, respectively.

Image classification-based ground filtering of point clouds extracted from UAV-based aerial photos

Abstract With the advent of unmanned aerial vehicles (UAVs) for mapping applications, it is possible to generate 3D dense point clouds using stereo images. This technology, however, has some disadvantages when compared to Light Detection and Ranging (LiDAR) system. Unlike LiDAR, digital cameras mounted on UAVs are incapable of viewing beneath the canopy, which leads to sparse points on the bare earth surface. In such cases, it is more challenging to remove points belonging to above-ground objects using ground filtering algorithms generated especially for LiDAR data. To tackle this problem, a methodology employing supervised image classification for filtering 3D point clouds is proposed in this study. A classified image is overlapped with the point cloud to determine the ground points to be used for digital elevation model (DEM) generation. Quantitative evaluation results showed that filtering the point cloud with this methodology has a good potential for high-resolution DEM generation.

Determining the optimum image fusion method for better interpretation of the surface of the Earth

ABSTRACT Image fusion is the production of high-resolution images by combining the spatial details of a high-resolution image with the spectral features of a low-resolution one. Reports of various quality metrics to evaluate the spectral and spatial qualities of fused images have been published. However, metrics may lead to misinterpretation due to inherent limitations in their mathematical algorithms. Hence, the use of additional assessment techniques in quality evaluation is reasonable. The purpose of the study was to compare the performances of several advanced fusion algorithms in order to help users in their choice of an appropriate fusion algorithm. Four different datasets were fused using advanced fusion algorithms, namely UNB PanSharp, Hyperspherical Color Space, High-Pass Filtering, Ehlers, Subtractive, Wavelet Single Band, Gram-Schmidt, Flexible Pixel-Based, and Criteria-Based. The spectral and spatial qualities of the fused images were evaluated using various quantitative procedures to ensure comprehensive and reliable comparison. The results showed that the Flexible Pixel-Based and High-Pass Filtering algorithms were very successful with regard to spatial quality, whereas the Flexible Pixel-Based and Criteria-Based algorithms were very successful with regard to spectral quality. The authors conclude that the Flexible Pixel-Based algorithm can be used for applications that require high spectral and spatial quality.

LAND COVER MAPPING WITH ADVANCED CLASSIFICATION ALGORITHMS

Remote sensing technologies are used in many applications to extract information from the surface of the earth. Image classification, which is one of the most widely-used ways of information extraction, is a controversial topic in remote sensing. This is because all classification algorithms introduced in the literature cause classification errors to some extent. Simple classification algorithms like Minimum Distance, Parallelpiped and Mahalanobis Distance commit a large amount of classification errors. This, of course, has encouraged the remote sensing community to develop more advanced classification algorithms to further increase classification accuracy. This study uses sophisticated classification algorithms Support Vector Machines (SVM), k-Nearest Neighbour (kNN) and Artificial Neural Network (ANN) to classify a WorldView-2 multispectral image in order to produce land cover maps. The accuracies of the produced thematic maps were evaluated with randomly-selected control points. The SVM algorithm classified the imagery with the best classification accuracy of 72.38%.

KİTLE KAYNAĞIN UZAKTAN ALGILAMADA KULLANIMI

Uzaktan algilama yontemleri ile elde edilen veriler ile arazi okyanus ve atmosfer ozellikleri hakkinda bilgi saglanir. Bu veriler islenerek cok cesitli cevresel arastirmalar ve Cografi Bilgi Sistemi uygulamalari yapilmaktadir. Bilgisayar teknolojilerinde meydana gelen buyuk gelismelere ragmen, uydu verilerinin analizi ve yorumlamasinda cesitli zorluklar meydana gelmektedir. Bu zorluklarin asiminda kitle kaynak kullanilabilir. Kitle kaynak; veri elde etme, problem cozme gibi cesitli uygulamalarda insanlarin kullanilmasidir. Kitle kaynak ile kolayca cozumlenemeyen problemlere cozum bulunabildigi gibi, cesitli uygulamalarin yapiminda harcanan zaman, maliyet ve caba da azalmaktadir. Bu calismada sistematik bir literatur taramasi yapilarak, uzaktan algilama biliminde kitle kaynagin kullanim alanlari incelenmistir. Yapilan calisma, uzaktan algilama biliminde kitle kaynagin kullaniminin oldukca yeni oldugunu gostermektedir. Ayrica b u konudaki makalelerin cogunlugunun 2016 ve sonrasinda yayinlanmaya baslandigi gorulmektedir. Basilan yayinlara gore kitle kaynak, uzamsal problemlerin cozumunde geleneksel algoritmalara oranla cok daha dogru sonuc veren cozumler uretebilmektedir.