The rise of data-parallel applications poses a significant challenge to the energy consumption of computing architectures.In-cache computation is a promising solution for achieving high parallelism and energy efficiency because it can eliminate data movement between the cache and the processor.Existing in-cache computing architectures transform a portion of cache arrays into computing arrays, with all rows of these arrays serving as computing lines.The remaining cache arrays are used as cachelines to store the data required by computing arrays or processors.However, in these array-level in-cache computing architectures, only a few computing lines in each computing array are active at runtime while the others are idle, which incurs severe cache capacity loss and space underutilization.In addition, bursty memory accesses of data-parallel applications also cause significant in-cache data movement latency.To address these problems, we propose MagiCache, a virtual in-cache computing engine.First, we design a novel cacheline-level in-cache computing architecture in which each cache array can configure some rows as computing lines and the other rows as cachelines with negligible overhead.Second, a virtual engine is further designed on this novel architecture to dynamically allocate different rows of each array as computing lines or cachelines based on runtime computation and storage requirements, thus realizing efficient cacheline-level space management.Finally, we present an instruction chaining technique to overlap the bursty access latency by enabling asynchronous execution of computing arrays.Evaluation results show that MagiCache achieves a 1.19x-1.61xspeedup over the state-of-the-art in-cache computing architectures with 6.5 KB of additional storage.Our cacheline-level space * These authors contributed equally to this work.

The generality and robustness of a man-made target detection method are of high practical value in SAR applications. However, current methods hardly adapt to different SAR images simultaneously because data property and target characteristics are not identical in images with various resolutions and scenes. The key to achieving high generality and robustness is extracting stable and invariant information in different SAR images. This paper analyzes the scattering of man-made targets and natural backgrounds in SAR images and assumes that the noise characteristics are relatively stable and invariant when resolution and observing background change. Then the Relative Spectral Stationarity (RSS) is proposed based on two-dimensional spectrum analysis to measure the distance between the observed data and a manually generated noise. RSS takes the manually generated noise with known and definite properties as a standard, so it is irrelevant to image parameters and scenes. A low RSS indicates that the characteristics of observing data are dominated by noise, whereas a high value means there may be targets in the observing data weakening the noise characteristics. An efficient segmentation algorithm Sparsity-based Format-free Segmentation within errore(SFFe) is proposed to process the RSS map and complete the detection method. SAR images with resolutions ranging from 0.1m to 8m are employed as testing data. Various testing scenes are constructed to simulate different practical conditions. Experimental results validate that the proposed RSS-based method works well in SAR images with different resolutions, bands, and observing scenes, obtaining reliable and robust detection results and outperforming canonical methods on various criteria.

Built-up area extraction is a primary and fundamental processing in applications such as city planning using remotely sensed high-resolution synthetic aperture radar (SAR) images. One of the critical challenges is that canopy-covered area is always falsely extracted due to similarity between canopy and building in scattering power and texture pattern, resulting in high false alarm and low overall accuracy. In this letter, physical scatterings on built-up areas and canopy-covered areas are analyzed, seeking the distinct differences between these two ground types under various observing scales. A spectral reconfiguration (SR) descriptor is proposed in frequency domain to describe differences that can be strengthened by frequency modulating strategy. Both theoretical and experimental analyses show that the SR descriptor can separate buildings and canopy greatly. Meanwhile, it enjoys both slight computational burden and low operative complexity. Based on this descriptor, an SR-intensity-based built-up extraction algorithm is proposed. Experimental results validate that the SR-intensity-based algorithm acquires low false alarm rate with high accuracy, showing great practical meaning and potential of improvement for the application of urban remote sensing.

A large-scale SAR image covers various land covers and targets with different complexity and statistical characteristics. Single algorithm always fails to deal with the interpretation tasks of large-scale remote sensing images, due to the mismatch between local background conditions and the algorithm requirements. This paper proposes a novel processing method named Image-Key-Element Decomposition, aiming at decomposing a SAR image into several regions under a certain criterion. Every decomposed key element is the subset of the whole image and represents a pixel set with a particular characteristic. Considering that many algorithms have certain requirements about the complexity status of backgrounds, homogeneous is used as the criterion in this paper and the SAR images will be decomposed into homogeneous and heterogeneous elements. The experimental result shows that homogeneous and heterogeneous regions are separated by the proposed model and the following processing may take advantage of the result to improve the performance.

CFAR is a widely used algorithm for target detection in SAR images. The simplicity of computation and stable performance make it a key role in practice. However, for man-made target detection, conventional CFAR has a limited performance because of the non-adaptive processing window and the varieties of categories, sizes and structures of targets. In order to detect man-made targets with different size and complex structures in high resolution SAR images, an improved CFAR algorithm with an adaptive processing window named Adaptive-Window CFAR is proposed. A global guard window obtained by pre-detection adaptively is used to take place of the guard window in conventional CFAR makes AW -CFAR an algorithm with both adaptive threshold and adaptive processing window. In this case, the size of detectable targets is not fixed anymore and targets of different sizes and complex structures are detectable in AW-CF AR. Images with different resolutions and environments, which contain different categories of man-made targets, are used in the experiments. The experimental results show that the AW -CFAR inherits the simplicity of computation and stable performance of the conventional CFAR and has a better performance of man-made target detection in high resolution SAR images.

Hyperspectral images (HSI) contain a wealth of spectral and spatial information, spectral-spatial combination is an effective way in improving the classification accuracy for HSI. To characterize the variability of spatial features at different scales better, a multiscale spectral-spatial classification method with adaptive filtering (MSAF) is proposed. The proposed method consists of the following four steps. Firstly, the spectral features are extracted by a feature selection algorithm. Secondly, the adaptive edge-preserving filtering with different scales are conducted on each feature, and then several stacks of data blocks containing spatial information can be obtained. Thirdly, the combinations of the spectral and spatial data blocks are classified using support vector machine (SVM). Finally, a post-processing is conducted to improve the classification results further. The experiments on the hyperspectral data demonstrate that the proposed method can improve the classification accuracy significantly compared to the SVM classifier, especially need less parameters than the spectral-spatial EPF method.