Search

Aircraft detection in synthetic aperture radar (SAR) images is a tough problem in remote sensing image interpretation, due to different scattering characteristics and scales of aircrafts as well as...

This paper proposes a new method for feature extraction of synthetic aperture radar (SAR) image based on parametric supervised t-stochastic neighbor embedding (PS-tSNE). To mitigate the rejection of dissimilar targets from the same class, aspect angles of targets are modeled to supervise and construct the distributional properties of the training data in the original space. Then, an explicit nonlinear mapping using kernel trick is proposed by an extension of non-parametric t-SNE supervised by the information of aspect angles. This method preserves the local structure of the targets of SAR images as well as possible and enables explicit out-of-sample extensions. Experimental results based on moving and stationary target automatic recognition (MSTAR) dataset illustrate the effective performance of the proposed method on visualization and recognition.

Ocean surveillance is an important application of synthetic aperture radar (SAR) image interpretation. In the ocean, the ship target usually attracts the attention of researchers, hence many interpretation works focus on it. The precise ship segmentation can facilitate further processing, such as feature extraction and target recognition. This paper proposes a refined segmentation method for ship target, in which the ratio of exponentially weighted averages (ROEWA) operator is used to produce the edge map and the elliptical constraint (EC) is incorporated into the energy function of the gradient vector flow (GVF) snake model to extract the ship target precisely. With this prior shape constraint, the snake contour can overcome unexpected distortions stemming from speckle noise and side lobe effect, etc. experimental results based on TerraSAR X-band and RADARSAT-2 C-band data illustrate the effective performance of the proposed method on refined segmentation for ship targets.

Flood detection from synthetic aperture radar (SAR) images should be performed with accurate, stable, automated and time-efficient algorithms; however, few methods meet all these requirements. Recently, Giustarini et al. proposed an automated promising methodology, capable of providing satisfactory results in flood detection. The algorithm is based on the assumption that a flood image contains a relatively high number of pixels with low backscatter values, exhibiting a bimodal histogram. For the case of a histogram that is not bimodal, the optimization of the theoretical curve describing the water pixels has to be manually constrained in a user-defined range. To overcome this shortcoming, this letter proposes an alternative procedure for core water body identification. First, by thresholding the difference image, derived by change detection between the flood and reference images, a mask of core water bodies is identified. Then, the mask is applied on the flood image, to extract the water pixels located in...

With its weather- and illumination-independent characteristics, synthetic aperture radar (SAR) has become an important tool for change detection. There are two critical steps in SAR image change detection: designing a change detector and choosing a decision rule. Given a measure from a change detector, the change detection results could be sensitive to the decision rule, such as the selection of a threshold. This letter presents a change detection measure based on a likelihood ratio and the statistical distribution of SAR intensity images. The likelihood ratio is defined as the ratio between the joint probability density functions (PDFs) of a pair of SAR images. Under the condition that both PDFs follow the gamma distribution, the histogram of this change detection measure deduced from the likelihood ratio has a single and steep peak that can be used to reliably and easily determine the change detection threshold. Analyses of SAR image pairs from different platforms show that the proposed change detection...