K. Vani

Classification of satellite images using new Fuzzy cluster centroid for unsupervised classification algorithm

Vagueness in the boundaries of land cover classes is one of the important problems in the image classification. Fuzzy c means (FCM) is a traditional clustering algorithm that has been widely used in the satellite image classification. However, this algorithm has the drawback of falling into a local minimum and it needs much time to accomplish the classification for a large data set. In order to overcome these drawbacks, a New Fuzzy Cluster Centroid (NFCC) for unsupervised classification algorithm is proposed to improve the traditional FCM and fuzzy weighted c means (FWCM) algorithm. In this work a, new objective function is formulated by adding the new term along with the distance between the pixels and cluster centers in the spectral domain. This new term is formulated by multiplying the Lagranges multiplier with the membership values of the pixel for a particular class is subtracted with one. It gives weightage to the instance of a particular pixel. The inclusion of the fuzzy centroid for each cluster increases the stability of the algorithm and the inclusion of the new term reduces the number of iterations for image classification. The technique was applied to both IKONOS and QuickBird images. Overall accuracy statistics indicates that NFCC for unsupervised classification algorithm increases the accuracy of the satellite image classification at the pixel level.

Automatic mountain detection in lunar images using texture of DTM data

This paper presents the development and implementation of a novel approach for automatic detection of the lunar mountains using Digital Terrain Model (DTM). The approach consists preprocessing the data, denoising, extracting texture information of the DTM, choosing an appropriate threshold using the Renyi Entropy threshold selection method, then the post-processing to extract the boundary of the mountain. The approach is applied to eight test sites, which were chosen in a manner so that the mountains are isolated. The detected mountains are assessed by their morphometric properties. The accuracy of the approach was assessed by determining its accuracy in detecting manually defined mountains and by comparing diameter estimates with separately determined values. The results are in agreement with the manually detected mountains, with the average detection performance of 0.91 and average precision of 0.97. The proposed approach extracts the mountain boundaries slightly more precisely than K-Means clustering which had an average detection performance of 0.89. We have proposed a novel approach for automatic mountain detection in lunar images using texture based approach.An approach has been attempted on DTM of lunar images.We examine the terrains roughness in lunar images using entropy measure.Mountain region with high slope and height are detected accurately.The mountain regions with low slope and low height are not detected accurately.

Automatic detection of ridges in lunar images using phase symmetry and phase congruency

Lunar surface exploration is increasing rapidly. These exploring satellites provide a large number of high resolution images containing topographical information. The topographical information in lunar surface are craters, ridges, mountains and grabens. Extracting this topographical information manually is time-consuming. Hence, an automatic feature extraction is favored. This paper presents a novel approach using image processing techniques to automatically detect ridges in lunar images. The approaches adopted for this development includes phase symmetry, phase congruency and morphological operations to automatically detect significant ridges. The phase symmetry extracts symmetry features with discontinuities, phase congruency extracts features lying in low contrast regions and morphological operations such as thinning and pruning are used to obtain significant ridges. The proposed novel approach experiments on a test set of different regions. These different region images are obtained from different sensors (LROC, Selene and Clementine) having different spatial resolution and illumination variation. The results obtained are compared with the plan curvature method; and they are evaluated based on true and false detection of ridge pixels. Irrespective of illumination variation and spatial resolution, the proposed approach provides better results than the plan curvature method and its detection rate is approximately 92%. We have proposed a novel approach for automatic ridge detection in lunar images using phase symmetry and phase congruency method.All ridge detection techniques are available only for terrestrial data. This is the first attempt of automatically identifying ridges in lunar data.The automatic detection approach has been attempted on different resolution of lunar images.We examine the symmetric nature of the ridges in lunar images using the phase symmetry component.The ridges are extracted automatically using image processing techniques.

Fusion of Satellite Images Using Multiwavelet Transform

The image sensors of the satellite can focus only on a particular operating range and environmental conditions, so it may not receive all the information necessary and each of the input images conveys important information that cannot be discarded. Scalar wavelet fusion scheme cannot possess all the properties necessary for producing images of sophisticated multisource data. Multiwavelet fusion scheme can offer more defined features in the fused image than any of the individual source images because of its ability to merge images in multiwavelet space, different frequency are processed differently. The paper is an attempt to combine multisensor images to obtain a single composite image with extended information using multiwavelet fusion scheme. The paper proposes a new fusion algorithm for combining the two source images get better quality image.

Satellite image processing

Satellite image processing plays a vital role for research and developments in Astronomy, Remote Sensing, GIS, Agriculture Monitoring, Disaster Management and many other fields of study. Satellite images are recorded in digital forms and then processed by the computers to extract information. Variations in the scene characteristics are represented as variations in brightness on images. A particular part of scene reflecting more energy will appear bright while a different part of the same scene that reflecting less energy will appear black. Digital image consists of discrete picture elements called pixels. Each pixel is a number represented as DN (Digital Number), that depicts the average radiance of relatively small area within a scene.

Clustering-based dome detection in lunar images using DTM data

ABSTRACT As the quest for lunar exploration is increasing, large numbers of lunar images are being acquired by many satellites. These images provide more information about topographic features on the lunar surface. Detecting and mapping these features using satellite images are of great scientific interest to understand them in detail. This article presents the development and implementation of an approach for automatic lunar domes detection from digital terrain model using clustering techniques. This approach consists of pre-processing, denoising, clustering, segmentation, post-processing, and boundary detection. This approach also examines domes morphometric property such as diameter, circularity index, and aspect ratio. The proposed method is experimented on nine test sites and evaluated by manual analysis for accurate detection with the detailed qualitative assessment. The manual analysis depicts that the results are in agreement with the automatic detection, while the overall statistical results reveal the detection performance as the true detection rate and false detection rate, which is achieved as 83.75% and 16.25%. In addition, the evaluated results also depict the morphometric parameters diameter, circularity index, and aspect ratio from the detected dome.

Optimization of UAV Video Frames for Tracking

In this digital era, UAV is becoming a trend setter in surveillance and gathering traffic information. Shortage of time to view the entire video necessitates video optimization. In this paper, a novel method has been proposed for optimizing the video frames without variation in the tracking path of the object. The keyframes are extracted using absolute difference of histogram of consecutive frames with mean as threshold. Then principal keyframes are selected at regular interval and finally compiled into an optimized video. In the tracking session, the region of interest is obtained from the first frame of the video and the SURF features are extracted and tracked with KLT tracker. A SURF feature is tracked along the video and position is tabulated. The tracking path is represented graphically to evaluate the tracking deviation from original and optimized video. The proposed method had successfully achieved the average time saved as 90.68% with negligible tracking deviation.

Multi-sensor image fusion of the lunar image data using DT-CWT and curvelet transform

Multisensor image fusion is the process of combining relevant information from high spatial resolution image and high spectral resolution image. This paper proposes a new image fusion method based on Dual Tree-Complex Wavelet Transform (DTCWT), and Curvelet transform for remotely sensed lunar image data in order to extract features accurately. Different fusion techniques have been used in the past separately for spatial and spectral quality image enhancement. In this study, we use a new image fusion technique based on Dual Tree — Complex Wavelet Transform (DT-CWT) and Curvelet transforms. Results indicate that the fused lunar image shows good spatial fidelity and the spectral resolution of the fused product was preserved after image data fusion. It is seen that 95.98% of the spectral content is preserved by curvelet fusion. From the results of statistical evaluation parameters demonstrated for the two study sites, it is found that curvelet transform gives better results than the other techniques commonly used.

Feature Extraction from Lunar Images

Automatic feature identification from orbital imagery would be of wide use in planetary science. For geo scientific applications, automatic shape-based feature detection offers a fast and non-subjective means of identifying geological structures within data. Most previously published examples of circular feature detection for geo scientific applications aimed to identify impact craters from optical or topographic data. Various techniques used include the texture analysis, template matching, and machine learning. In this paper, we propose a new method for the extraction of features from the planetary surface, based on the combination of several image processing techniques, including a shadow removal, watershed segmentation and the Circular Hough Transform (CHT). The original edge map of craters is detected by canny operator. In most literatures Hough transform is generally used for crater detection but we have added a shadow removal which includes a novel color image fusion method, based on the multi-scale Retinex (MSR) and discrete wavelet transform (DWT), is proposed. This proposed method is capable of detecting partially visible craters, and overlapping craters.