The despeckling of synthetic aperture radar images using two different convolutional neural network architectures is presented in this paper. The first method presents a novel Siamese convolutional neural network with a dilated convolutional network in each branch. Recently, attention mechanisms have been introduced to convolutional networks to better model and recognize features. Therefore, we propose a novel design for a convolutional neural network using an attention mechanism for an encoder–decoder-type network. The framework consists of a multiscale spatial attention network to improve the modeling of semantic information at different spatial levels and an additional attention mechanism to optimize feature propagation. Both proposed methods are different in design but they provide comparable despeckling results in subjective and objective measurements in terms of correlated speckle noise. The experimental results are evaluated on both synthetically generated speckled images and real SAR images. The methods proposed in this paper are able to despeckle SAR images and preserve SAR features.

This paper presents the design of a high-voltage pulse-based radar and a supervised data processing method for soil moisture estimation. The goal of this research was to design a pulse-based radar to detect changes in soil moisture using a cross-hole approach. The pulse-based radar with three transmitting antennas was placed into a 12 m deep hole, and a receiver with three receive antennas was placed into a different hole separated by 100 m from the transmitter. The pulse generator was based on a Marx generator with an LC filter, and for the receiver, the high-frequency data acquisition card was used, which can acquire signals using 3 Gigabytes per second. Used borehole antennas were designed to operate in the wide frequency band to ensure signal propagation through the soil. A deep regression convolutional network is proposed in this paper to estimate volumetric soil moisture using time-sampled signals. A regression convolutional network is extended to three dimensions to model changes in wave propagation between the transmitted and received signals. The training dataset was acquired during the period of 73 days of acquisition between two boreholes separated by 100 m. The soil moisture measurements were acquired at three points 25 m apart to provide ground truth data. Additionally, water was poured into several specially prepared boreholes between transmitter and receiver antennas to acquire additional dataset for training, validation, and testing of convolutional neural networks. Experimental results showed that the proposed system is able to detect changes in the volumetric soil moisture using Tx and Rx antennas.

This paper presents an automatic classification of plastic material’s inorganic pigment using terahertz spectroscopy and convolutional neural networks (CNN). The plastic materials were placed between the THz transmitter and receiver, and the acquired THz signals were classified using a supervised learning approach. A THz frequency band between 0.1–1.2 THz produced a one-dimensional (1D) vector that is almost impossible to classify directly using supervised learning. This paper proposes a novel pre-processing of 1D THz data that transforms 1D data into 2D data, which are processed efficiently using a convolutional neural network. The proposed pre-processing algorithm consists of four steps: peak detection, envelope extraction, and a down-sampling procedure. The last main step introduces the windowing with spectrum dilatation that reorders 1D data into 2D data that can be considered as an image. The spectrum dilation techniques ensure the classifier’s robustness by suppressing measurement bias, reducing the complexity of the THz dataset with negligible loss of accuracy, and speeding up the network classification. The experimental results showed that the proposed approach achieved high accuracy using a CNN classifier, and outperforms 1D classification of THz data using support vector machine, naive Bayes, and other popular classification algorithms.

This paper investigates the use of ground-penetrating radar (GPR) technology for detecting shallow buried objects, utilizing an air-coupled stepped frequency continuous wave (SFCW) radar system that operates within a 2 GHz bandwidth starting at 500 MHz. Different GPR data acquisition methods for air-coupled systems are compared, specifically down-looking, side-looking, and circular acquisition strategies, employing the back projection algorithm to provide focusing of the acquired GPR data. Experimental results showed that the GPR can penetrate up to 0.6 m below the surface in a down-looking mode. The developed radar and the back projection focusing algorithm were used to acquire data in the side-looking and circular mode, providing focused images with a resolution of 0.1 m and detecting subsurface objects up to 0.3 m below the surface. The proposed approach transforms B-scans of the GPR-based data into 2D images. The provided approach has significant potential for advancing shallow object detection capabilities by transforming hyperbola-based features into point-like features.

This paper presents a frequency-modulated optical signal generator in the THz band. The proposed method is based on a fast optical frequency sweep of a single narrowband laser diode used together with an optical fiber interferometer. The optical frequency sweep using a single laser diode is achieved by generating short current pulses with a high amplitude, which are driving the laser diode. Theoretical analysis showed that the modulation frequency could be changed by the optical path difference of the interferometer or optical frequency sweep rate of a laser diode. The efficiency of the optical signal generator with Michelson and Fabry–Perot interferometers is theoretically analyzed and experimentally evaluated for three different scenarios. Interferometers with different optical path differences and a fixed optical frequency sweep rate were used in the first scenario. Different optical frequency sweep rates and fixed optical path differences of the interferometers were used in the second scenario. This paper presents a method for optical chirp generation using a programmable current pulse waveform, which drives a laser diode to achieve nonlinear optical sweep with a fixed optical path difference of the interferometer. The experimental results showed that the proposed signals could be generated within a microwave (1–30 GHz) and THz band (0.1–0.3 THz).

With shorter wavelengths than microwaves and greater penetration depth than infrared light, waves in the terahertz spectrum offer unique material testing opportunities. Terahertz technology offers non-invasive and non-destructive testing in the form of spectroscopy and imaging. The most used systems for terahertz imaging are time-domain spectroscopy systems. However, frequency domain spectroscopy systems could offer excellent frequency resolution and be more suitable for biomedical applications. Terahertz imaging based on frequency domain spectroscopy systems is slow, and suffers from frequency tuning errors. A novel one-dimensional imaging principle is presented in this paper. In addition, frequency range optimization based on convolutional neural networks and occlusion sensitivity is utilized for frequency range optimization. Frequency range optimization is used to determine the optimal frequency range for data acquisition. The optimal frequency range or bandwidth should be wide enough for effective phase detection, and should be at the intersection of several spectral footprints in the observed medium. The intersection of spectral footprints is estimated using the proposed frequency range optimization algorithm based on a convolutional neural network and occlusion sensitivity algorithm. The proposed algorithm selects the most sensitive frequency band of THz spectrum automatically, and enables very fast acquisitions for object inspection and classification.