Real-Time Detection of Concealed Threats with Passive Millimeter Wave and Visible Images Via Deep Neural Networks

Passive millimeter wave has been employed in security inspection owing to a good penetrability to clothing and harmlessness. However, the passive millimeter wave images (PMMWIs) suffer from low resolution and inherent noise. The published methods have rarely improved the quality of images for PMMWI and performed the detection only based on PMMWI with bounding box, which cause a high rate of false alarm. Moreover, it is difficult to identify the low-reflective non-metallic threats by the differences in grayscale. In this paper, a method of detecting concealed threats in human body is proposed. We introduce the GAN architecture to reconstruct high-quality images from multi-source PMMWIs. Meanwhile, we develop a novel detection pipeline involving semantic segmentation, image registration, and comprehensive analyzer. The segmentation network exploits multi-scale features to merge local and global information together in both PMMWIs and visible images to obtain precise shape and location information in the images, and the registration network is proposed for privacy concerns and the elimination of false alarms. With the grayscale and contour features, the detection for metallic and non-metallic threats can be conducted, respectively. After that, a synthetic strategy is applied to integrate the detection results of each single frame. In the numerical experiments, we evaluate the effectiveness of each module and the performance of the proposed method. Experimental results demonstrate that the proposed method outperforms the existing methods with 92.35% precision and 90.3% recall in our dataset, and also has a fast detection rate.

SLIDING MODE BASED PREDICTIVE CONTROLLER OF A SPHEROIDAL UNDERWATER VEHICLE

This paper focuses on hydrodynamic modeling and control of spheroidal underwater vehicle. The vehicle considered in this paper is appendage free and unstable. Water jet propulsion system is used in this vehicle. The dynamics of the vehicle is highly unstable due to munk moment. The spheroidal shape underwater robot is used in nuclear reactor inspection, port security inspection, defence and ocean surveillance where external appendages are not required. A new and innovative control technique, Sliding mode based model predictive control is introduced in this paper. Sliding mode control technique is used to stabilize the vehicle and once the vehicle model is stabilized it is easy to apply Model Predictive Control (MPC). Model Predictive control technique is used to control the heading of spheroidal underwater vehicle. Simulation results show that the Sliding mode based predictive control performance is better than simple PD control and state feedback controller.

A Suspicious Multi-Object Detection and Recognition Method for Millimeter Wave SAR Security Inspection Images Based on Multi-Path Extraction Network

There are several major challenges in detecting and recognizing multiple hidden objects from millimeter wave SAR security inspection images: inconsistent clarity of objects, similar objects, and complex background interference. To address these problems, a suspicious multi-object detection and recognition method based on the Multi-Path Extraction Network (MPEN) is proposed. In MPEN, You Only Look Once (YOLO) v3 is used as the base network, and then the Multi-Path Feature Pyramid (MPFP) module and modified residual block distribution are proposed. MPFP is designed to output the deep network feature layers separately. Then, to distinguish similar objects more easily, the residual block distribution is modified to improve the ability of the shallow network to capture details. To verify the effectiveness of the proposed method, the millimeter wave SAR images from the laboratory’s self-developed security inspection system are utilized in conducting research on multi-object detection and recognition. The detection rate (probability of detecting a target) and average false alarm (probability of error detection) rate of our method on the target are 94.6% and 14.6%, respectively. The mean Average Precision (mAP) of recognizing multi-object is 82.39%. Compared with YOLOv3, our method shows a better performance in detecting and recognizing similar targets.

A Depth Map Generation Method for Path Planning of Mobile Robot

Recently, the unmanned mobile robots have received broad applications, such as industrial and security inspection, disinfection and epidemic prevention, warehousing logistics, agricultural picking, etc. In order to drive autonomously from departure to destination, an unmanned mobile robot mounts different sensors to collect information around it and further understand its surrounding environment based on the perceptions. Here we proposed a method to generate high-resolution depth map for given sparse LiDAR point cloud. Our method fits the point cloud into a 3D curve and projects LiDAR data onto the curve surface, and then we make appropriate interpolations of the curve and finally implement the Delaunay triangulation algorithm to all the data points on the 3D curve. The experimental results show that our approach can effectively improve the resolution of depth maps from sparse LiDAR measurements.

Dangerous goods identification based on multi-channel neural network

In the current society, there is an increasing demand for dangerous goods identification technology in X-ray images, but at the current stage, most of the identification of dangerous goods in X-ray images still relies on artificial eye recognition. In order to solve this problem, this paper proposes A method for automatically and intelligently identifying dangerous goods in X-ray images based on the transformation of the convolutional neural network. By adding multi-channel convolution and normalization to the convolutional neural network, the target features are extracted to achieve automatic detection of dangerous goods. The purpose of better identification. In the simulation experiment, using the public data set and self-built data set in the X-ray security inspection field, the accuracy of the identification of dangerous goods in the X-ray image was obtained more satisfactory results than the traditional dangerous goods identification. The improved Alex Net network’ s testing accuracy on contraband knives and guns is 8.53% and 11.6% higher than the training accuracy of the original Alex Net network.

Broadband Tunable Terahertz Beam Deflector Based on Liquid Crystals and Graphene

Terahertz (THz) technology has unique applications in, for example, wireless communication, biochemical characterization, and security inspection. However, high-efficiency, low-cost, and actively tunable THz modulators are still scarce. We propose a broadband tunable THz beam deflector based on liquid crystals (LCs). By a periodic gradual distribution of the orientation of the LC in one direction, a frequency-independent geometric phase modulation is obtained. The LC device with this specific orientation distribution was obtained through ultraviolet polarization exposure. We have verified the broadband beam deflection in both the simulation and experiment. The device can achieve a good spin-coupled beam deflection effect in the 0.8–1.2 Thz band, and the average polarization conversion efficiency exceeds 70%. Moreover, because the electro-optical responsivity of LCs is excellent, graphene transparent electrode layers introduced on the upper and lower substrates enable the deflection modulation to be switched and dynamic tuning to be achieved.

X-Ray Image Recognition Based on Improved Mask R-CNN Algorithm

It is an important part of security inspection to carry out security and safety screening with X-ray scanners. Computer vision plays an important role in detection, recognition, and location analysis in intelligent manufacturing. The object detection algorithm is an important part of the intelligent X-ray machine. Existing threat object detection algorithms in X-ray images have low detection precision and are prone to missed and false detection. In order to increase the precision, a new improved Mask R-CNN algorithm is proposed in this paper. In the feature extraction network, an enhancement path is added to fuse the features of the lower layer into the higher layer, which reduces the loss of feature information. By adding an edge detection module, the training effect of the sample model can be improved without accurate labeling. The distance, overlap rate, and scale difference between objects and region proposals are solved using DIoU to improve the stability of the region proposal’s regression, thus improving the accuracy of object detection; SoftNMS algorithm is used to overcome the problem of missed detection when the objects to be detected overlap each other. The experimental results indicate that the mean Average Precision (mAP) of the improved algorithm is 9.32% higher than that of the Mask R-CNN algorithm, especially for knife and portable batteries, which are small in size, simple in shape, and easy to be mistakenly detected, and the Average Precision (AP) is increased by 13.41% and 15.92%, respectively. The results of the study have important implications for the practical application of threat object detection in X-ray images.

Deep Feature Fusion Based Dual Branch Network for X-ray Security Inspection Image Classification

Automatic computer security inspection of X-ray scanned images has an irresistible trend in modern life. Aiming to address the inconvenience of recognizing small-sized prohibited item objects, and the potential class imbalance within multi-label object classification of X-ray scanned images, this paper proposes a deep feature fusion model-based dual branch network architecture. Firstly, deep feature fusion is a method to fuse features extracted from several model layers. Specifically, it operates these features by upsampling and dimension reduction to match identical sizes, then fuses them by element-wise sum. In addition, this paper introduces focal loss to handle class imbalance. For balancing importance on samples of minority and majority class, it assigns weights to class predictions. Additionally, for distinguishing difficult samples from easy samples, it introduces modulating factor. Dual branch network adopts the two components above and integrates them in final loss calculation through the weighted sum. Experimental results illustrate that the proposed method outperforms baseline and state-of-art by a large margin on various positive/negative ratios of datasets. These demonstrate the competitivity of the proposed method in classification performance and its potential application under actual circumstances.