Sea-Based Infrared Scene Interpretation by Background Type Classification and Coastal Region Detection for Small Target Detection

V2T-GAN: Three-Level Refined Light-Weight GAN with Cascaded Guidance for Visible-to-Thermal Translation

Infrared image simulation is challenging because it is complex to model. To estimate the corresponding infrared image directly from the visible light image, we propose a three-level refined light-weight generative adversarial network with cascaded guidance (V2T-GAN), which can improve the accuracy of the infrared simulation image. V2T-GAN is guided by cascading auxiliary tasks and auxiliary information: the first-level adversarial network uses semantic segmentation as an auxiliary task, focusing on the structural information of the infrared image; the second-level adversarial network uses the grayscale inverted visible image as the auxiliary task to supplement the texture details of the infrared image; the third-level network obtains a sharp and accurate edge by adding auxiliary information of the edge image and a displacement network. Experiments on the public dataset Multispectral Pedestrian Dataset demonstrate that the structure and texture features of the infrared simulation image obtained by V2T-GAN are correct, and outperform the state-of-the-art methods in objective metrics and subjective visualization effects.

A Novel Method for Sea-Land Clutter Separation Using Regularized Randomized and Kernel Ridge Neural Networks

Classification of clutter, especially in the context of shore based radars, plays a crucial role in several applications. However, the task of distinguishing and classifying the sea clutter from land clutter has been historically performed using clutter models and/or coastal maps. In this paper, we propose two machine learning, particularly neural network, based approaches for sea-land clutter separation, namely the regularized randomized neural network (RRNN) and the kernel ridge regression neural network (KRR). We use a number of features, such as energy variation, discrete signal amplitude change frequency, autocorrelation performance, and other statistical characteristics of the respective clutter distributions, to improve the performance of the classification. Our evaluation based on a unique mixed dataset, which is comprised of partially synthetic clutter data for land and real clutter data from sea, offers improved classification accuracy. More specifically, the RRNN and KRR methods offer 98.50% and 98.75% accuracy, outperforming the conventional support vector machine and extreme learning based solutions.

Fast infrared horizon detection algorithm based on gradient directional filtration

Infrared imaging has been widely used in the field of sea surface monitoring. Horizon detection is a key step before a target's detection, locating, and tracking in the sea-sky infrared scene. Reducing processing time while ensuring accuracy is the research focus of infrared horizon detection. This paper proposes a novel method of a line segment detector (LSD) algorithm with gradient direction filtering. First, the rough extraction of the sea-sky region is used to limit the size of the detected image, and then the potential horizon line segment is extracted, applying the improved LSD algorithm in the sea-sky region, which probably contains many false extraction results. Then, gradient direction filtering is designed to pick the horizon line segments in this step. Finally, the horizon line segments are stitched to obtain the whole horizon line based on random sample consensus. The results of the comparative experiments show that this novel method has high detection accuracy, and the processing time is significantly shortened; what is more, we can also conclude that this method has a good performance on the detection stability.

Coarse-Fine-Stitched: A Robust Maritime Horizon Line Detection Method for Unmanned Surface Vehicle Applications

The horizon line has numerous applications for an unmanned surface vehicles (USV), such as autonomous navigation, attitude estimation, obstacle detection and target tracking. However, maritime horizon line detection is quite a challenging problem. The pixel points of the horizon line features are far fewer than the pixel points of the entire image, on the one hand. Conversely, the detection results might be impacted negatively by the complex maritime environment, waves, light changing, and partial occlusions due to maritime vessels or islands, for example. To solve these problems, a robust horizon line detection method named coarse-fine-stitched (CFS) is proposed in this paper. First, in the coarse step of CFS, a line segment detection approach using gradient features is applied to build a line candidate pool, which probably contains many false detection results. Then, hybrid feature filtering is designed to pick the horizon line segments from the pool in the fine step. Finally, the fine line segments are stitched to obtain the whole horizon line based on random sample consensus (RANSAC). Using real data in the maritime environment, the experimental results demonstrate the effectiveness of CFS, compared to the existing methods in terms of accuracy and robustness.

Sparsity Adaptive Matching Pursuit Detection Algorithm Based on Compressed Sensing for Radar Signals

In this paper, the application of the emerging compressed sensing (CS) theory and the geometric characteristics of the targets in radar images are investigated. Currently, the signal detection algorithms based on the CS theory require knowing the prior knowledge of the sparsity of target signals. However, in practice, it is often impossible to know the sparsity in advance. To solve this problem, a novel sparsity adaptive matching pursuit (SAMP) detection algorithm is proposed. This algorithm executes the detection task by updating the support set and gradually increasing the sparsity to approximate the original signal. To verify the effectiveness of the proposed algorithm, the data collected in 2010 at Pingtan, which located on the coast of the East China Sea, were applied. Experiment results illustrate that the proposed method adaptively completes the detection task without knowing the signal sparsity, and the similar detection performance is close to the matching pursuit (MP) and orthogonal matching pursuit (OMP) detection algorithms.

A Sea-Sky Line Detection Method for Unmanned Surface Vehicles Based on Gradient Saliency

Special features in real marine environments such as cloud clutter, sea glint and weather conditions always result in various kinds of interference in optical images, which make it very difficult for unmanned surface vehicles (USVs) to detect the sea-sky line (SSL) accurately. To solve this problem a saliency-based SSL detection method is proposed. Through the computation of gradient saliency the line features of SSL are enhanced effectively, while other interference factors are relatively suppressed, and line support regions are obtained by a region growing method on gradient orientation. The SSL identification is achieved according to region contrast, line segment length and orientation features, and optimal state estimation of SSL detection is implemented by introducing a cubature Kalman filter (CKF). In the end, the proposed method is tested on a benchmark dataset from the “XL” USV in a real marine environment, and the experimental results demonstrate that the proposed method is significantly superior to other state-of-the-art methods in terms of accuracy rate and real-time performance, and its accuracy and stability are effectively improved by the CKF.

An Ensemble Successive Project Algorithm for Liquor Detection Using Near Infrared Sensor

Spectral analysis technique based on near infrared (NIR) sensor is a powerful tool for complex information processing and high precision recognition, and it has been widely applied to quality analysis and online inspection of agricultural products. This paper proposes a new method to address the instability of small sample sizes in the successive projections algorithm (SPA) as well as the lack of association between selected variables and the analyte. The proposed method is an evaluated bootstrap ensemble SPA method (EBSPA) based on a variable evaluation index (EI) for variable selection, and is applied to the quantitative prediction of alcohol concentrations in liquor using NIR sensor. In the experiment, the proposed EBSPA with three kinds of modeling methods are established to test their performance. In addition, the proposed EBSPA combined with partial least square is compared with other state-of-the-art variable selection methods. The results show that the proposed method can solve the defects of SPA and it has the best generalization performance and stability. Furthermore, the physical meaning of the selected variables from the near infrared sensor data is clear, which can effectively reduce the variables and improve their prediction accuracy.

Labeled RFS-Based Track-Before-Detect for Multiple Maneuvering Targets in the Infrared Focal Plane Array

The problem of jointly detecting and tracking multiple targets from the raw observations of an infrared focal plane array is a challenging task, especially for the case with uncertain target dynamics. In this paper a multi-model labeled multi-Bernoulli (MM-LMB) track-before-detect method is proposed within the labeled random finite sets (RFS) framework. The proposed track-before-detect method consists of two parts-MM-LMB filter and MM-LMB smoother. For the MM-LMB filter, original LMB filter is applied to track-before-detect based on target and measurement models, and is integrated with the interacting multiple models (IMM) approach to accommodate the uncertainty of target dynamics. For the MM-LMB smoother, taking advantage of the track labels and posterior model transition probability, the single-model single-target smoother is extended to a multi-model multi-target smoother. A Sequential Monte Carlo approach is also presented to implement the proposed method. Simulation results show the proposed method can effectively achieve tracking continuity for multiple maneuvering targets. In addition, compared with the forward filtering alone, our method is more robust due to its combination of forward filtering and backward smoothing.