A visual fingerprint update algorithm based on crowdsourced localization and deep learning for smart IoV

Recently, deep learning and vision-based technologies have shown their great significance for the prospective development of smart Internet of Vehicle (IoV). When the smart vehicle enters the indoor parking of a shopping mall, the vision-based localization technology can provide reliable parking service. As known, the vision-based technique relies on a visual map without a change in the position of the reference object. Although, some researchers have proposed a few automatic visual fingerprinting (AVF) methods, which are aiming at reducing the cost of building the visual map database. However, the AVF method still costs too much under such a situation, since it is impossible to determine the specific location of the displaced object. Given the smart IoV and the development of deep learning approach, we propose an algorithm for solving the problem based on crowdsourcing and deep learning in this paper. Firstly, we propose a Region-based Fully Convolutional Network (R-FCN) based method with the feedback of crowdsourced images to locate the specific displaced object in the visual map database. Secondly, we propose a method based on quadratic programming (QP) for solving the translation vector of the displaced objects, which finally solves the problem of updating the visual map database. The simulation results show that our method can provide a higher detection sensitivity and correction accuracy as well as the relocation results. It means that our proposed algorithm outperforms the compared one, which is verified by both synthetic and real data simulation.

Thermal Imaging Camera Supporting the Navigation of UAVs

The topic of this paper is an evaluation of developed sensor intended for navigation aid of unmanned aerial vehicles (UAVs). Its operation is based on processing images acquired with a thermal camera operating in the long-wave infrared band (LWIR) placed underneath a vehicle’s chassis. The vehicle’s spatial displacement is determined by analyzing movement of characteristic thermal radiation points (ground, forest, buildings, etc.) in pictures acquired by the thermal camera. Magnitude and direction of displacement is obtained by processing the stream of consecutive pictures with optical-flow based algorithm in real time. Radiation distribution analysis allows to calculate camera’s self-translation vector. Advantages of measuring translation based on thermal image analysis is lack of drift effect, resistance to magnetic field variations, low susceptibility to electromagnetic interference and change in weather conditions as compared to traditional inertial navigation sensors. As opposed to visible light situational awareness sensors, it offers operation in complete darkness (harsh weather, nights and indoors).The topic of this paper is an evaluation of developed sensor intended for navigation aid of unmanned aerial vehicles (UAVs). Its operation is based on processing images acquired from a thermal camera operating in the long wave infrared band (LWIR) placed underneath a vehicle’s chassis. The vehicle’s spatial displacement is determined by analyzing movement of characteristic thermal radiation points (ground, forest, buildings, etc.) in pictures acquired by the thermal camera. Magnitude and direction of displacement is obtained by processing the stream of consecutive pictures with optical-flow based algorithm in real time. Radiation distribution analysis allows to calculate camera’s self-translation vector. Advantages of measuring translation based on thermal image analysis is lack of drift effect, resistance to magnetic field variations, low susceptibility to electromagnetic interference and change in weather conditions as compared to traditional inertial navigation sensors. As opposed to visible light situational awareness sensors, it offers operation in complete darkness (harsh weather, nights and indoors).

A Visual Fingerprint Update Algorithm Based On Crowdsourced Localization And Deep Learning For Smart IoV

Recently, the deep learning and vision-based technologies has shown their great significance for the prospective development of smart Internet of Vehicle (IoV). When the smart vehicle enters the indoor parking of a shopping mall, the vision-based localization technology can provide reliable parking service. As known, the vision-based technique relies on a visual map without a change in the position of the reference object. Although, some researchers have proposed a few automatic visual fingerprinting (AVF) methods, which are aiming at reducing the cost of building the visual map database. However, the AVF method still costs too much under such situation, since it is impossible to determine the specific location of the displaced object. In view of the smart IoV and the development of deep learning approach, we propose a crowdsourcing and deep learning based algorithm for solving the problem in this paper. Firstly, we propose a Region-based Fully Convolutional Network (R-FCN) based method with the feedback of crowdsourced images to locate the specific displaced object in the visual map database. Secondly, we propose a method based on quadratic programming (QP) for solving the translation vector of the displaced objects, which finally solves the problem of updating the visual map database. The simulation results show that our method can provide a higher detection sensitivity and correction accuracy as well as the relocation results. It means that our proposed algorithm outperforms the compared one, which is verified by both synthetic and real data simulation.

Improved extended Gaussian image and its application on 3D point clouds registration

To solve the problem of lacking geometric and topological information for conventional 3D point clouds registration algorithm, this paper proposed a novel 3D point clouds registration algorithm based on improved extended Gaussian image. The proposed registration algorithm first estimates a normal vector and curvature of every point by least square method. Then, according to the normal vector to calculate extended Gaussian image (EGI) and complex extended Gaussian image (CEGI). By using the calculated EGI/CEGI and spherical harmonic function, a correlated function is constructed to calculate 3D rotation space to obtain initial positions coarse registration result. At last, by using the Fourier transform to estimate translation vector and coarse registration, the iterative closest point algorithm is used to obtain the fine registration results. Experiments on three groups of different 3D point clouds are performed to validate the proposed registration algorithm. Experimental results have shown that the proposed algorithm has good performances on registration of different forms of 3D point clouds. The robustness and efficiency of our proposed algorithm can effectively solve the problem that it is difficult to find the target or the homonymic feature points in the registration process of 3D point clouds.

Connection of Two Conventional Enhanced to Encrypt Color Images

This document introduces a new cryptosystem mixing two improvement standards generally used for text encryption, in order to give birth a new color image encryption algorithm capable of dealing with known attacks. Firstly, two substitution matrixes attached to a strong replacement function will be generated for advanced Vigenere technique application. At the end of this first round, the output vector is subdivided into size blocks according to the used chaotic map, for acting a single enhanced Hill circuit insured by a large inversible matrix. A detailed description of such a large involutive matrix constructed using Kronecker products will be given. accompanied by a dynamic translation vector to eliminate any linearity. A solid chaining is established between the encrypted block and the next clear block to avoid any differential attack. Simulations carried out on a large volume of images of different sizes and formats ensure that our approach is not exposed to any known attacks.

Efficient Rock Mass Point Cloud Registration Based on Local Invariants

Point cloud registration is one of the basic research hotspots in the field of 3D reconstruction. Although many previous studies have made great progress, the registration of rock point clouds remains an ongoing challenge, due to the complex surface, arbitrary shape, and high resolution of rock masses. To overcome these challenges, a novel registration method for rock point clouds, based on local invariants, is proposed in this paper. First, to handle the massive point clouds, a point of interest filtering method based on a sum vector is adopted to reduce the number of points. Second, the remaining points of interest are divided into several cluster point sets and the centroid of each cluster is calculated. Then, we determine the correspondence between the original point cloud and the target point cloud by proving the inherent similarity (using the trace of the covariance matrix) of the remaining point sets. Finally, the rotation matrix and translation vector are calculated, according to the corresponding centroids, and a correction method is used to adjust the positions of the centroids. To illustrate the superiority of our method, in terms of accuracy and efficiency, we conducted experiments on multiple datasets. The experimental results show that the method has higher accuracy (about ten times) and efficiency than similar existing methods.

ON DIFFERENT DEFINITIONS OF STRAIN TENSORS IN GENERAL SHELL THEORIES OF VEKUA-AMOSOV TYPE

Several possible definiions of strains in a general shell theory of I.N. Vekua – A.A. Amosov type are considered. The higher-order shell model is definedon a two-dimensional manifold within a set of fieldvariables of the firstkind determined by the expansion factors of the spatial vector fieldof the translation. Two base vector systems are introduced, the firs one so-called concomitant corresponds to the cotangent fibrtion of the modelling surface while the other is defind on a surface equidistant to the modelling one. The distortion appears as a two-point tensor referred to both base systems after covariant differentiationof the translation vector feld. Thus, two main definition of the strain tensor become possible, the firstone referred to the main basis whereas the second to the concomitant one. Some possible simplificationsof these tensors are considered, and the interrelation between the general theory of A.A. Amosov type and the classical ones is shown.

Testing the Accuracy of the Modified ICP Algorithm with Multimodal Weighting Factors

SLAM technology is increasingly used to self-locate mobile robots in an unknown environment. One of the methods used in this technology is called scan matching. Increasing evidence is placed on the accuracy and speed of the methods used in terms of navigating mobile robots. The aim of this article is to present a modification to the standard method of Iterative Closest Point (ICP) environment scan matching using the authors’ three original weighting factors based on the error modeling. The presented modification was supported by a simulation study whose aim was not exclusively to check the effect of the factors but also to examine the effect of the number of points in scans on the correct and accurate development of the rotation matrix and the translation vector. The study demonstrated both an increase in the accuracy of ICP results following the implementation of the proposed modification and a noticeable increase in accuracy with an increase in the mapping device’s angular resolution. The proposed method has a positive impact on reducing number of iteration and computing time. The research results have shown to be promising and will be extended to 3D space in the future.

LiDAR-Camera Calibration Using Line Correspondences

In this paper, we introduce a novel approach to estimate the extrinsic parameters between a LiDAR and a camera. Our method is based on line correspondences between the LiDAR point clouds and camera images. We solve the rotation matrix with 3D–2D infinity point pairs extracted from parallel lines. Then, the translation vector can be solved based on the point-on-line constraint. Different from other target-based methods, this method can be performed simply without preparing specific calibration objects because parallel lines are commonly presented in the environment. We validate our algorithm on both simulated and real data. Error analysis shows that our method can perform well in terms of robustness and accuracy.

Efficient Coarse Registration of Pairwise TLS Point Clouds Using Ortho Projected Feature Images

The degree of automation and efficiency are among the most important factors that influence the availability of Terrestrial light detection and ranging (LiDAR) Scanning (TLS) registration algorithms. This paper proposes an Ortho Projected Feature Images (OPFI) based 4 Degrees of Freedom (DOF) coarse registration method, which is fully automated and with high efficiency, for TLS point clouds acquired using leveled or inclination compensated LiDAR scanners. The proposed 4DOF registration algorithm decomposes the parameter estimation into two parts: (1) the parameter estimation of horizontal translation vector and azimuth angle; and (2) the parameter estimation of the vertical translation vector. The parameter estimation of the horizontal translation vector and the azimuth angle is achieved by ortho projecting the TLS point clouds into feature images and registering the ortho projected feature images by Scale Invariant Feature Transform (SIFT) key points and descriptors. The vertical translation vector is estimated using the height difference of source points and target points in the overlapping regions after horizontally aligned. Three real TLS datasets captured by the Riegl VZ-400 and the Trimble SX10 and one simulated dataset were used to validate the proposed method. The proposed method was compared with four state-of-the-art 4DOF registration methods. The experimental results showed that: (1) the accuracy of the proposed coarse registration method ranges from 0.02 m to 0.07 m in horizontal and 0.01 m to 0.02 m in elevation, which is at centimeter-level and sufficient for fine registration; and (2) as many as 120 million points can be registered in less than 50 s, which is much faster than the compared methods.