Neural networks

We present a survey on maritime object detection and tracking approaches, which are essential for the development of a navigational system for autonomous ships. The electro-optical (EO) sensor considered here is a video camera that operates in the visible or the infrared spectra, which conventionally complements radar and sonar for situational awareness at sea and has demonstrated its effectiveness over the last few years. This paper provides a comprehensive overview of various approaches of video processing for object detection and tracking in the maritime environment. We follow an approach-based taxonomy wherein the advantages and limitations of each approach are compared. The object detection system consists of the following modules: horizon detection, static background subtraction, and foreground segmentation. Each of these has been studied extensively in maritime situations and has been shown to be challenging due to the presence of background motion especially due to waves and wakes. The key processes involved in object tracking include video frame registration, dynamic background subtraction, and the object tracking algorithm itself. The challenges for robust tracking arise due to camera motion, dynamic background, and low contrast of tracked object, possibly due to environmental degradation. The survey also discusses multisensor approaches and commercial maritime systems that use EO sensors. The survey also highlights methods from computer vision research, which hold promise to perform well in maritime EO data processing. Performance of several maritime and computer vision techniques is evaluated on Singapore Maritime Dataset.

Video Processing From Electro-Optical Sensors for Object Detection and Tracking in a Maritime Environment: A Survey

Nowadays, through wall imaging (TWI) is an emerging topic of research in which one of the most important tasks is to minimize the clutter through which detection accuracy can be improved. Clutter in TWI is due to many reasons like wall coupling, antenna coupling, multiple re∞ections etc. To analyze the clutter reduction techniques, flrstly we indigenously assembled a TWI system (i.e., step frequency continuous wave radar (SFCW)) in UWB range (freq. 3.95GHz to 5.85GHz), and difierent observations have been taken. We have considered metallic plate and one more material with low dielectric constant (Te∞on) as a target and kept them behind the plywood wall. A-scan and B-scan observations have been carried out. The observed data are preprocessed for imaging and then difierent types of clutter reduction techniques like Principal Component Analysis (PCA), Independent Component Analysis (ICA), Factor Analysis (FA) and Singular Value Decomposition (SVD) have been applied, and results were analyzed. Signal to noise ratio (SNR) of the flnal images (i.e., after clutter removal with difierent techniques) has been computed to compare the results and know the efiectiveness of individual clutter removal techniques. It is observed that ICA has better capability to remove the clutter in comparison to other applied techniques; especially it is found that ICA has a capability to distinguish the difierence between clutter and low dielectric target whereas other clutter removal techniques are not showing signiflcant result.

https://www.jpier.org/pierb/pierb17/03.09060903.pdf

Analysis of Clutter Reduction Techniques for through Wall Imaging in UWB Range

An optimization on pictogram identification for the road-sign recognition task using SVMs

Attention has been focused on the moving target detection in heavy sea clutter. On the basis of detection model of moving target with fluctuant amplitudes, a novel adaptive algorithm in fractional Fourier transform (FRFT) domain is proposed, which combines statistic-based and FRFT-based detection method. FRFT has good energy concentration property on linear frequency modulation (LFM) signal with the optimal transform angle, which is determined by calculating spectral kurtosis (SK) in FRFT domain. Grading iterative search method is used for good accuracy of parameter estimation and fast calculation speed. A novel adaptive line enhancer (ALE) in FRFT domain is proposed to suppress sea clutter and improve signal-to-clutter ratio (SCR), which provides less error and faster convergence. Leakage factor is introduced into the update equation of weight vector to reduce `memory effect` and step size is normalised by the power of input signal with better convergence characteristic. In the end, both X-band and S-band real sea clutter is used for verification and the results present that the proposed algorithm has good convergence property and small mean square error (MSE). Weak moving target in low SCR environment (SCR = -6 dB) can be well detected and estimated, which indicates the effectiveness of the algorithm.

Adaptive fractional fourier transform-based detection algorithm for moving target in heavy sea clutter

This paper proposes a novel algorithm for retrieving the ocean wind vector from marine radar image sequences in real time. It is presented as an alternative to mitigate anemometer problems, such as blockage, shadowing, and turbulence. Since wind modifies the sea surface, the proposed algorithm is based on the dependence of the sea surface backscatter on wind direction and speed. This algorithm retrieves the wind vector using radar measurements in the range of 200–1500 m. Wind directions are retrieved from radar images integrated over time and smoothed (averaged) in space by searching for the maximum radar cross section in azimuth as the radar cross section is largest for upwind directions. Wind speeds are retrieved by an empirical third-order polynomial geophysical model function (GMF), which depends on the range distance in the upwind direction to a preselected intensity level and the intensity level. This GMF is approximated from a dataset of collocated in situ wind speed and radar measurements ...

/pdf/real-time-ocean-wind-vector-retrieval-from-marine-radar-8odqcgoy6l.pdf

Real-Time Ocean Wind Vector Retrieval from Marine Radar Image Sequences Acquired at Grazing Angle

The presence of sea clutter in marine radar signals is sometimes not desired. So, efficient radar signal processing techniques are needed to reduce it. In this way, nonlinear signal processing techniques based on neural networks (NNs) are used in the proposed clutter reduction system. The developed experiments show promising results characterized by different subjective (visual analysis of the processed radar images) and objective (clutter reduction, target enhancement and signal-to-clutter ratio improvement) criteria. Moreover, a deep study of the NN structure is done, where the low computational cost and the high processing speed of the proposed NN structure are emphasized.

https://www.mdpi.com/1424-8220/9/3/1913/pdf

Sea Clutter Reduction and Target Enhancement by Neural Networks in a Marine Radar System

In this paper, curve-fitting and intensity-level-selection (ILS)-based algorithms for wind parameter extraction from shipborne X-band nautical radar images are investigated First, to exclude the rain cases and low-backscatter images, a data quality control process is designed for both algorithms An additional process is then introduced for the ILS-based method to improve the accuracy of wind measurements, including the recognition of blockages and islands in the temporally integrated radar images For the low sea states, a dual-curve-fitting is proposed These wind algorithms are tested using radar images and shipborne anemometer data collected on the east coast of Canada It is shown that the dual-curve-fitting algorithm produces improvements in the mean differences between the radar and the anemometer results for wind direction and speed of about 57° and 03 m/s, respectively, under sea states with significant wave height lower than 230 m Also, a harmonic function that is least-squares fitted to the selected range distances vector as a function of antenna look direction is applied Compared with the original ILS-based algorithm, the modified procedure reduces the standard deviation for wind direction and speed by about 4° and 02 m/s, respectively Finally, the performance of these two modified methods are compared

/pdf/comparison-of-algorithms-for-wind-parameters-extraction-from-2rqgzxztfo.pdf

Comparison of Algorithms for Wind Parameters Extraction From Shipborne X-Band Marine Radar Images

The mean-shift (MS) algorithm is applied for reducing speckle noise and segmenting synthetic aperture radar (SAR) images. Two coastal images acquired by Envisat's advanced SAR (ASAR) [European Space Agency (ESA)] are used. Studies of the MS parameters are carried out according to the desired product: a speckle filtered image where textures and edges are preserved, or a segmented image, where land and sea are distinguished, as a previous stage for obtaining a land mask and detecting the coastal line. In all cases, Gaussian kernels are used. Speckle filtering results are compared with those obtained using uniform kernels, proving that the former provides better results than the latter. A segmentation approach based on the positions and frequencies at which the MS converges is applied. The use of a combined spatial-range processing and the corresponding bandwidths makes the MS suitable for the two proposed problems. The solid theoretical basis of this procedure allows designing a guided search of the best parameters according to the desired solution, avoiding a tedious trial-and-error process. Although the used images have different characteristics, results prove that similar sets of parameters can be used, showing some degree of robustness with respect to the image, for a given sensor and image acquisition mode.

Spatial-Range Mean-Shift Filtering and Segmentation Applied to SAR Images

One of the most relevant parameters to characterize the severity of ocean waves is the significant wave height (H
 s
 ). The estimate of H
 s
 from remotely sensed data acquired by non-coherent X-band marine radars is a problem not completely solved nowadays. A method commonly used in the literature (standard method) uses the square root of the signal-to-noise ratio (SNR) to linearly estimate H
 s
 . This method has been widely used during the last decade, but it presents some limitations, especially when swell-dominated sea states are present. To overcome these limitations, a new non-linear method incorporating additional sea state information is proposed in this article. This method is based on artificial neural networks (ANNs), specifically on multilayer perceptrons (MLPs). The information incorporated in the proposed MLP-based method is given by the wave monitoring system (WaMoS II) and concerns not only to the square root of the SNR, as in the standard method, but also to the peak wave length and mean wave period. Results for two different platforms (Ekofisk and FINO 1) placed in different locations of the North Sea are presented to analyze whether the proposed method works regardless of the sea states observed in each location or not. The obtained results empirically demonstrate how the proposed non-linear solution outperforms the standard method regardless of the environmental conditions (platform), maintaining real-time properties.

/pdf/estimate-of-significant-wave-height-from-non-coherent-marine-3du6kdarxn.pdf

Raul Vicen-Bueno

Papers

Real-Time Ocean Wind Vector Retrieval from Marine Radar Image Sequences Acquired at Grazing Angle

Sea Clutter Reduction and Target Enhancement by Neural Networks in a Marine Radar System

Comparison of Algorithms for Wind Parameters Extraction From Shipborne X-Band Marine Radar Images

Spatial-Range Mean-Shift Filtering and Segmentation Applied to SAR Images

Estimate of significant wave height from non-coherent marine radar images by multilayer perceptrons