Erkan Uslu

Curvelet-Based Synthetic Aperture Radar Image Classification

Curvelet transform (CT) is a multiscale directional transform that enables the use of texture and spatial locality information. In synthetic aperture radar (SAR) imaging, CT is mostly used in speckle noise reduction. This letter utilizes CT for feature extraction in land use classification. Two types of curvelet-based feature extraction methods are implemented for SAR. The first one is defined and used in content-based image retrieval and is based on generalized Gaussian distribution parameter estimation for each curvelet subband. The second implementation is a genuine method that utilizes the use of curvelet subband histograms, namely, histogram of curvelets (HoC). Using the proposed curvelet-based feature extraction method (HoC) on SAR data, better classification accuracies up to 99.56% are achieved compared to original data and H/A/α decomposition features. Compared to speckle-noise-reduced data classification results, it can be said that curvelet-based feature extraction is also robust against speckle noise.

Enhancing land use classification with fusing dual-polarized TerraSAR-X and multispectral RapidEye data

Abstract. The contribution of dual-polarized synthetic aperture radar (SAR) to optical data for the accuracy of land use classification is investigated. For this purpose, different image fusion algorithms are implemented to achieve spatially improved images while preserving the spectral information. To compare the performance of the fusion techniques, both the microwave X-band dual-polarized TerraSAR-X data and the multispectral (MS) optical image RapidEye data are used. Our test site, Gediz Basin, covers both agricultural fields and artificial structures. Before the classification phase, four data fusion approaches: (1) adjustable SAR-MS fusion, (2) Ehlers fusion, (3) high-pass filtering, and (4) Bayesian data fusion are applied. The quality of the fused images was evaluated with statistical analyses. In this respect, several methods are performed for quality assessments. Then the classification performances of the fused images are also investigated using the support vector machines as a kernel-based method, the random forests as an ensemble learning method, the fundamental k-nearest neighbor, and the maximum likelihood classifier methods comparatively. Experiments provide promising results for the fusion of dual polarimetric SAR data and optical data in land use/cover mapping.

Implementation of frontier-based exploration algorithm for an autonomous robot

Exploration is defined as the selection of target points that yield the biggest contribution to a specific gain function at an initially unknown environment. Exploration for autonomous mobile robots is closely related to mapping, navigation, localization and obstacle avoidance. In this study an autonomous frontier-based exploration strategy is implemented. Frontiers are defined as the border points that are calculated throughout the mapping and navigation stage between known and unknown areas. Frontier-based exploration implementation is compatible with the Robot Operating System (ROS). Also in this study, real robot platform is utilized for testing and the effect of different frontier target assignment approaches are comparatively analyzed by means of total path length and thereby total exploration time.

Synthetic Aperture Radar Image Clustering with Curvelet Subband Gauss Distribution Parameters

Curvelet transform is a multidirectional multiscale transform that enables sparse representations for signals. Curvelet-based feature extraction for Synthetic Aperture Radar (SAR) naturally enables utilizing spatial locality; the use of curvelet-based feature extraction is a novel method for SAR clustering. The implemented method is based on curvelet subband Gaussian distribution parameter estimation and cascading these estimated values. The implemented method is compared against original data, polarimetric decomposition features and speckle noise reduced data with use of k-means, fuzzy c-means, spatial fuzzy c-means and self-organizing maps clustering methods. Experimental results show that the curvelet subband Gaussian distribution parameter estimation method with use of self-organizing maps has the best results among other feature extraction-clustering performances, with up to 94.94% overall clustering accuracies. The results also suggest that the implemented method is robust against speckle noise.

Classification of heart arrthymias by using wavelet and merged wavelet packet transforms

In this work efficiency of feature extraction methods based on linear wavelet transform and merged wavelet packets technique are evaluated relatively with different supervised classification methods. Experimental heart arrthymia data has been obtained from MIT-BIH arrthymia database. Total of 1200 training and 1200 test samples have been chosen equally for 6 classes from the database. For the purpose of increasing the accuracy with chosen datasets, mixed noises from different sources in the ECG signals are removed with signal processing methods. Support vector machines (SVM) and statistical neural networks (RBF, PNN and GRNN) are utilized for classification purpose. In the experimental results it has been observed that the best accuracy is accomplished by RBF kernel SVM, trained with any of the two mentioned feature extraction methods.

Obstacle avoidance with Vector Field Histogram algorithm for search and rescue robots

The capability of avoid obstacles is the one of the key issues in autonomous search-and-rescue robots research area. In this study, the avoiding obstacles capability has been provided to the virtula robots in USARSim environment. The aim is finding the minimum movement when robot faces an obstacle in path. For obstacle avoidance we used an real time path planning method which is called Vector Field Histogram (VFH). After experiments we observed that VFH method is successful method for obstacle avoidance. Moreover, the usage of VFH method is highly incresing the amount of the visited places per unit time.

Canny and gauss enhancement on speckle reducing anisotropic diffusion

In this work speckle reduction methods defined in the literature are basically compared according to effect on classification, edge preservation and noise reduction in homogenous regions. The methods are tested on labeled SAR image, simulated SAR image and optical image. It is aimed to present mutual effect of speckle reduction methods on both edge preservation and noise reduction. One of the purposes of this work is to improve SRAD method, which is especially defined for speckle noise reduction, by combining it with Canny edge detection and Gauss filtering. It is proved by the experimental results that proposed method enhances the quality measure results.

Arrhythmia classification by Local Fractional Fourier Transform

Automated analysis of electrocardiography (ECG) signals compose a system for early detection of heart disorders. One of the most important parts of ECG signal classification system is to produce the discriminative features for proper identification of heart disorders. Fractional Fourier Transform (FrFT) as the generalized form of Fourier Transform (FT) gives a hybrid time-frequency representation based on an angle parameter. A genuine method called Local Fractional Fourier Transform (LFrFT) is proposed by means of exploiting local features for non-stationary signals such as heart beats. Experimental results are given for LFrFT features extracted from MIT-BIH arrhythmia ECG dataset with different angle parameters on several classifiers.

Design and implementation of an outdoor line follower robot for RoboCup search and rescue league

RoboCup is the most prestigious robotics contest in the world, with increasing popularity among robot communities and new contests. In this study, design and implementation of a line follower robot for outdoor categories which have been recently added to RoboCup Search and Rescue League is emphasized. In this category, a 100m long colored rope is placed in a field whose floor is made of sand, concrete and laminate and the competing robots are ranked according to the maximum trajectory distance that they cover in 20 minutes. The track is made of right and acute angle parts for the robot which is expected to follow the line including also 15° ramps. The robot which is implemented on ROS (The Robot Operating System) follows the line with the camera sensor above it.

Hazmat charts detection on search and rescue league of RoboCup competitions

RoboCup competitions are among the most prestigious robotics organizations in the world. Contestants compete in different categories with the robots they develop step by step to achieve the objectives of the organization for 2050. One of these categories is designed for search and rescue robots. Robots get points from the following criteria: identifying the victims autonomously in disaster situations, detecting hazmat chars, reading QR code and recognizing the various objects which is identified during the competition and marking their the locations on a map. In this study, some performance tests of various feature extraction methods have been performed for the detection of hazmat signs in competition environments of RoboCup search and rescue league. The tests were conducted in a laboratory which designed as RoboCup search and rescue league competition environment. After some comparison tests, the most appropriate method for real-time operation has been decided considering acceptable accuracy rates.