This document describes a challenge problem whose scope is two-fold. The first aspect is to develop SAR CCD algorithms that are applicable for X-band SAR imagery collected in an urban environment. The second aspect relates to effective data compression of these complex SAR images, where quality SAR CCD is the metric of performance. A set of X-band SAR imagery is being provided to support this development. To focus research onto specific areas of interest to AFRL, a number of challenge problems are defined. The data provided is complex SAR imagery from an AFRL airborne X-band SAR sensor. Some key features of this data set are: 10 repeat passes, single phase center, and single polarization (HH). In the scene observed, there are multiple buildings, vehicles, and trees. Note that the imagery has been coherently aligned to a single reference.

https://www.sdms.afrl.af.mil/content/challenge_areas/documents/A_challenge_problem_for_SAR_change_detection_and_data.pdf

A Challenge Problem for SAR Change Detection and Data Compression.

In this paper, complex-valued convolutional neural networks are presented, by giving the full deduction of the gradient descent algorithm for training this type of networks. The performances of convolutional neural networks in the real-valued domain for image classification gave rise to the idea of extending them to the complex-valued domain, also. Real-valued image classification experiments done using the MNIST and CIFAR-10 datasets have shown an improvement in performance of complex-valued convolutional neural networks over their real-valued counterparts.

Complex-valued convolutional neural networks for real-valued image classification

We propose two synthetic aperture radar (SAR) complex image compression schemes based on DLWT_IQ and DLWT_FFT. DLWT_IQ encodes the real parts and imaginary parts of the images using directional lifting wavelet transform (DLWT) and bit plane encoder (BPE), while DLWT_FFT encodes the real images converted by fast Fourier transform (FFT). Compared with discrete wavelet transform-IQ (DWT_IQ), DLWT_IQ improves the peak signal-to-noise ratio (PSNR) up to 1.28 dB and reduces the mean phase error (MPE) up to 21.74%; and compared with DWT_FFT, DLWT_FFT improves the PSNR up to 1.22 dB and reduces the MPE up to 20.32%. Moreover, the proposed schemes increase the PSNR up to 3.34 dB and decrease the MPE up to 50.43% as compared with the set partitioning in hierarchical trees (SPIHT) algorithm. In addition to this, we observe a novel phenomenon, that is, DLWT with direction prediction achieves a higher clustering capability for complex SAR images than DWT. Then, coding algorithm based on DLWT requires fewer coding bits than DWT for the same number of coding coefficients, and DLWT outperforms DWT in terms of rate-distortion performance even if the K-term nonlinear approximation of DWT is better than that of DLWT.

Complex SAR Image Compression Based on Directional Lifting Wavelet Transform With High Clustering Capability

Methods and apparatus compress data, comprising an In-phase (I) component and a Quadrature (Q) component. Statistical characteristics of the data are utilized to convert the data into a form that requires fewer bits in accordance with the statistical characteristics. The data may be further compressed by transforming the data and by modifying the transformed data in accordance with a quantization conversion table that is associated with the processed data. Additionally, redundancy may be removed from the processed data with an encoder. Subsequent processing of the compressed data may decompress the compressed data in order to approximate the original data by reversing the process for compressing the data with corresponding inverse operations. Interleaved I and Q components can be processed rather than separating the components before processing the data. The processed data type may be determined by providing metadata to retrieve the appropriate quantization table from a knowledge database.

Measurement and signature intelligence analysis and reduction technique

A technique for compressing synthetic aperture radar raw data using multiresolution representations and subband coding is considered. In particular, we present the performance of a transform coding compression method using wavelet basis, coupled with a threshold quantizer optimized for Gaussian statistics, as well as a proper subband bit allocation strategy. The performances achieved in terms of bit rate reduction and certain quality parameters computed on the images obtained by compressed data have been evaluated. These show an increased performance in the compression method with respect to conventional methods, albeit with a slightly increased complexity in the algorithm implementation.

SAR raw data compression by subband coding

Synthetic aperture radars (SAR) are coherent imaging systems that produce complex-valued images of the ground. Because modern systems can generate large amounts of data, there is substantial interest in applying image compression techniques to these products. We examine the properties of complex-valued SAR images relevant to the task of data compression. We advocate the use of transform-based compression methods but employ radically different quantization strategies than those commonly used for incoherent optical images. The theory, methodology, and examples are presented.

R.W. Ives

Papers

Compression of complex-valued SAR images