Deep learning-based image compressive sensing methods have received extensive attention in recent years due to their superior learning ability and fast processing speed. The majority of existing image compressive sensing neural networks use single-scale sampling, whereas multiscale sampling has demonstrated excellent performance compared to single-scale. We propose a multiscale deep network for compressive sensing image reconstruction that consists of a multiscale sampling network and a reconstruction network. First, we use convolution to mimic the linear decomposition of images, and the convolution is learned during the training process. Then a sampling network captures compressive measurements across multiple decomposed scales. The reconstruction network, which includes both the initial and enhanced reconstruction networks, learns an end-to-end mapping between the compressed sensing (CS) measurements and the recovered images of the network. Experimental results indicate that the proposed network framework outperforms the existing CS methods in terms of objective metrics, peak signal to noise ratio (PSNR), structural similarity index, and subjective visual quality. Specifically, at a 0.1 sampling rate, using 10 images for testing, and the average PSNR maximum (minimum) gain is 5.95 dB (0.25 dB).
Compressed sensing (CS) is a signal processing framework for effectively reconstructing signal from a small number of measurements obtained by linear projections of the signal. It is an ongoing challenge for the real-time image reconstruction of the computational imaging, including single pixel imaging based on CS. We built a block-based CS (BCS) image reconstruction framework via a deep learning network with smoothed projected Landweber (SPL). A fully connected network performs both BCS linear sensing and non-linear reconstruction stages, and SPL removes the blocking artifacts due to incorporate Wiener filtering into projected Landweber (PL) method at each iteration. The sensing matrix and nonlinear prediction operator are jointly optimized, and the smoothing filtering is coalesced into the PL framework for eliminating high-frequency oscillatory blocking artifact. Experimental results reveal that the optimized scheme outperforms the approach only based on deep neural network. The reconstruction quality can be improved while being only slightly slower, especially the gain of structural similarity is significantly better than peak signal-to-noise ratio, and the reconstruction image texture details are vivid and natural. At 10% sensing rate, the structural similarity maximum (minimum) gain reaches 0.098 (0.021). The proposed approach is not only far superior to other state-of-the-art CS algorithms in terms of reconstruction time and quality but also comparable with up-to-date deep learning methods.
KEYWORDS: Fiber optics sensors, Machine learning, Artificial neural networks, Data modeling, Temperature metrology, Civil engineering, Aerospace engineering, Neurons, Binary data, Signal to noise ratio
We propose a novel method to extract Brillouin frequency shift (BFS) from Brillouin Gain Spectrum (BGS) in Brillouin distributed fiber sensors. The method is based on machine learning of nearest neighbors. In order to find the BFS from the BGS, we design two datasets, one for storing all possible BGS, and the other for storing the corresponding BFS. By comparing the given BGS with the dataset of BGS, we get the minimal kth BFS. The BFS of the given BGS is determined by voting of the kth BFS. By simulations, we compare the performance of both neighbor-based machine learning and curve-fitting. The results show that the method of neighbor-based machine learning is more robust under a wide range of signal-to-noise ratios, pump pulse widths, and frequency scanning steps. The extracting method of neighbor-based machine learning is highly competitive for future Brillouin distributed fiber sensors.
In this paper, combining with the tenth anniversary load test of Sutong Bridge, we proposed the Brillouin Optical Time Domain Analysis (BOTDA) to conduct health monitoring research on the bridge, and completed the following three aspects: Firstly as an auxiliary measure of strain monitoring, monitor the level of stress and strain along the longitudinal bridge to each position. Secondly through the parking of the vehicles at different positions on the bridge pavement, the influence lines on the overall structural stress of the bridge are obtained. Thirdly through the long-term strain monitoring of the bridge, the impact of the usual traffic load information on the bridge strain is obtained.The results show that the technology not only breaks through the monitoring bottleneck of traditional point sensors, but also realizes distributed measurement of strain on the transmission path; it can also be used for real-time monitoring, damage identification, crack location, settlement monitoring of traffic load information of bridges and other structures.. Compared with the traditional sensor test results, the effectiveness and frontier of the technology are proved, and the significance of the distributed fiber-optic sensor technology for the health monitoring of major structures such as bridges is fully explained.
In this paper we propose a low-cost and stable configuration of Brillouin Optical Time Domain Analysis (BOTDA). Both pump and probe are generated by one single laser source for steady frequency beating. Polarization-maintaining modulators and amplifiers have been applied into the system in order to suppress the ground noise and to control the stability of the pump pulse. The probe is filtered and amplified to obtain the Stokes wave. The bias voltages of modulators are carefully controlled. We implement the prototype of interrogator by using this method and compare it with commercial products. The result shows that the long-term stability of the prototype is three times higher than that of commercial product.
In this paper, Brillouin Optical Time Domain Analysis (BOTDA) is proposed to solve the problem that the traditional point sensor is difficult to realize the comprehensive safety monitoring of bridges and so on. This technology not only breaks through the bottleneck of traditional monitoring point sensor, realize the distributed measurement of temperature and strain on a transmission path; can also be used for bridge and other structures of the damage identification, fracture positioning, settlement monitoring. The effectiveness and frontier of the technology are proved by comparing the test of the indoor model beam and the external field bridge, and the significance of the distributed optical fiber sensing technology to the monitoring of the important structure of the bridge is fully explained.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.