Search

When processing with large data such as satellite images, the computing speed is the problem need to be resolved. This paper introduces a method to improve the computational efficiency of the interval type-2 fuzzy c-means clustering(IT2-FCM) based on GPU platform and applied to land-cover classification from multi-spectral satellite image. GPU-based calculations are high performance solution and free up the CPU. The experimental results show that the performance of the GPU is many times faster than CPU. [ABSTRACT FROM PUBLISHER]

Obstacle detection is a fundamental issue of robot navigation and there have been several proposed methods for this problem. In this paper, we propose a new approach to find out obstacles on Depth Camera streams. The proposed approach consists of three stages. First, preprocessing stage is for noise removal. Second, different depths in a frame are clustered based on the Interval Type-2 Fuzzy Subtractive Clustering algorithm. Third, the objects of interest are detected from the obtained clusters. Beside that, it gives an improvement in the Interval Type-2 Fuzzy Subtractive Clustering algorithm to reduce the time consuming. In theory, it is at least 3700 times better than the original one, and approximate 980100 in practice on our depth frames. The results conducted on frames demonstrate that the distance from the camera to objects retrieved is exact enough for indoor robot navigation problems. [ABSTRACT FROM PUBLISHER]

Nowadays, many studies have focused on representation of type-2 fuzzy sets for reducing complexity of computation of operations, especially type-2 fuzzy logic systems. Advances of nVIDIA computing technology allow to deploy of general purpose applications of GPU by parallelizing algorithms. This paper deals with an approach to representation of type-2 fuzzy sets by dividing domain into grid, called grid-based T2FS. Grid-based T2FS is easily to understand and deploy on platform of Graphics Processor Units (GPU) computing and Compute Unified Device Architecture(CUDA) by taking the advantages of computation on matrices. Experiments are implemented in various applications involving join, meet, negation operation and defuzzification. Results on runtime and accuracy are summarised in comparison with CPU computation to show efficiency of the approach. [ABSTRACT FROM PUBLISHER]

Land cover classification have been developed for specially surveillance of change of land and generating update information. The paper introduces an approach to classification of land cover from multi-spectral satellite imagery using interval type-2 c-means clustering. Two channels (Near Infrared - NIR and Visible Red - NR) are used to generate NDVI image of study area. Then IT2-FCM is used to classify NDVI into six sub-classes presenting for six types of land cover. The method is implemented for two study areas in comparing with ISODATA algorithm and FCM. [ABSTRACT FROM PUBLISHER]