scholarly journals Joint Unsupervised Learning of Depth, Pose, Ground Normal Vector and Ground Segmentation by a Monocular Camera Sensor

Sensors ◽  
2020 ◽  
Vol 20 (13) ◽  
pp. 3737
Author(s):  
Lu Xiong ◽  
Yongkun Wen ◽  
Yuyao Huang ◽  
Junqiao Zhao ◽  
Wei Tian
Keyword(s):  
Geometric Constraints ◽  
Estimation Accuracy ◽  
Average Error ◽  
Normal Vector ◽  
Learning Procedure ◽  
Ground Segmentation ◽  
Unsupervised Approach ◽  
Normal Vectors ◽  
Self Learning ◽  
Scene Depth

We propose a completely unsupervised approach to simultaneously estimate scene depth, ego-pose, ground segmentation and ground normal vector from only monocular RGB video sequences. In our approach, estimation for different scene structures can mutually benefit each other by the joint optimization. Specifically, we use the mutual information loss to pre-train the ground segmentation network and before adding the corresponding self-learning label obtained by a geometric method. By using the static nature of the ground and its normal vector, the scene depth and ego-motion can be efficiently learned by the self-supervised learning procedure. Extensive experimental results on both Cityscapes and KITTI benchmark demonstrate the significant improvement on the estimation accuracy for both scene depth and ego-pose by our approach. We also achieve an average error of about 3 ∘ for estimated ground normal vectors. By deploying our proposed geometric constraints, the IOU accuracy of unsupervised ground segmentation is increased by 35% on the Cityscapes dataset.

scholarly journals Camera-based measurement of cyclist motion

2018 ◽  
Vol 233 (7) ◽  
pp. 1793-1805 ◽  
Author(s):  
Chris Eddy ◽  
Christopher de Saxe ◽  
David Cebon
Keyword(s):  
Contact Point ◽  
Ground Plane ◽  
Blind Spot ◽  
Geometric Constraints ◽  
Average Error ◽  
Camera System ◽  
Position Data ◽  
Close Range ◽  
Heavy Goods Vehicles ◽  
Full Scale Tests

Heavy goods vehicles are overrepresented in cyclist fatality statistics in the United Kingdom relative to their proportion of total traffic volume. In particular, the statistics highlight a problem for vehicles turning left across the path of a cyclist on their inside. In this article, we present a camera-based system to detect and track cyclists in the blind spot. The system uses boosted classifiers and geometric constraints to detect cyclist wheels, and Canny edge detection to locate the ground contact point. The locations of these points are mapped into physical coordinates using a calibration system based on the ground plane. A Kalman Filter is used to track and predict the future motion of the cyclist. Full-scale tests were conducted using a construction vehicle fitted with two cameras, and the results compared with measurements from an ultrasonic-sensor system. Errors were comparable to the ultrasonic system, with average error standard deviation of 4.3 cm when the cyclist was 1.5 m from the heavy goods vehicles, and 7.1 cm at a distance of 1 m. When results were compared to manually extracted cyclist position data, errors were less than 4 cm at separations of 1.5 and 1 m. Compared to the ultrasonic system, the camera system requires simple hardware and can easily differentiate cyclists from stationary or moving background objects such as parked cars or roadside furniture. However, the cameras suffer from reduced robustness and accuracy at close range and cannot operate in low-light conditions.

Extracting Geometric Edges from 3D Point Clouds Based on Normal Vector Change

2014 ◽  
Vol 571-572 ◽  
pp. 729-734
Author(s):  
Jia Li ◽  
Huan Lin ◽  
Duo Qiang Zhang ◽  
Xiao Lu Xue
Keyword(s):  
Point Clouds ◽  
Detection Algorithm ◽  
Growth Strategy ◽  
Normal Vector ◽  
Geometric Morphometric ◽  
3D Point Clouds ◽  
Real Scene ◽  
Edge Points ◽  
Normal Vectors ◽  
Change Intensity

Normal vector of 3D surface is important differential geometric property over localized neighborhood, and its abrupt change along the surface directly reflects the variation of geometric morphometric. Based on this observation, this paper presents a novel edge detection algorithm in 3D point clouds, which utilizes the change intensity and change direction of adjacent normal vectors and is composed of three steps. First, a two-dimensional grid is constructed according to the inherent data acquisition sequence so as to build up the topology of points. Second, by this topological structure preliminary edge points are retrieved, and the potential directions of edges passing through them are estimated according to the change of normal vectors between adjacent points. Finally, an edge growth strategy is designed to regain the missing edge points and connect them into complete edge lines. The results of experiment in a real scene demonstrate that the proposed algorithm can extract geometric edges from 3D point clouds robustly, and is able to reduce edge quality’s dependence on user defined parameters.

scholarly journals A special interpretation of the concept constant breadth for a space curve

2019 ◽  
Vol 23 (Suppl. 1) ◽  
pp. 371-382
Author(s):  
Tuba Agirman-Aydin
Keyword(s):  
Variable Coefficients ◽  
Space Curve ◽  
Normal Vector ◽  
Exit Point ◽  
Constant Breadth ◽  
Normal Vectors ◽  
Linear Differential ◽  
Curve Of Constant Breadth ◽  
Definition Of ◽  
Curvature And Torsion

The definition of curve of constant breadth in the literature is made by using tangent vectors, which are parallel and opposite directions, at opposite points of the curve. In this study, normal vectors of the curve, which are parallel and opposite directions are placed at the exit point of the concept of curve of constant breadth. In this study, on the concept of curve of constant breadth according to normal vector is worked. At the conclusion of the study, is obtained a system of linear differential equations with variable coefficients characterizing space curves of constant breadth according to normal vector. The coefficients of this system of equations are functions depend on the curvature and torsion of the curve. Then is obtained an approximate solution of this system by using the Taylor matrix collocation method. In summary, in this study, a different interpretation is made for the concept of space curve of constant breadth, the first time. Then this interpretation is used to obtain a characterization. As a result, this characterization we?ve obtained is solved.

scholarly journals A posture measurement approach for an articulated manipulator by RGB-D cameras

2019 ◽  
Vol 16 (2) ◽  
pp. 172988141983813
Author(s):  
Haobin Shi ◽  
Meng Xu ◽  
Kao-Shing Hwang ◽  
Chia-Hung Hung
Keyword(s):  
Point Cloud ◽  
Original Method ◽  
Consensus Method ◽  
Average Error ◽  
Normal Vector ◽  
Point Cloud Data ◽  
Sample Standard Deviation ◽  
Cylindrical Model ◽  
Cloud Data ◽  
Posture Measurement

The objective of this article aims at the safety problems where robots and operators are highly coupled in a working space. A method to model an articulated robot manipulator by cylindrical geometries based on partial cloud points is proposed in this article. Firstly, images with point cloud data containing the posture of a robot with five resolution links are captured by a pair of RGB-D cameras. Secondly, the process of point cloud clustering and Gaussian noise filtering is applied to the images to separate the point cloud data of three links from the combined images. Thirdly, an ideal cylindrical model fits the processed point cloud data are segmented by the random sample consensus method such that three joint angles corresponding to three major links are computed. The original method for calculating the normal vector of point cloud data is the cylindrical model segmentation method, but the accuracy of posture measurement is low when the point cloud data is incomplete. To solve this problem, a principal axis compensation method is proposed, which is not affected by number of point cloud cluster data. The original method and the proposed method are used to estimate the three joint angular of the manipulator system in experiments. Experimental results show that the average error is reduced by 27.97%, and the sample standard deviation of the error is reduced by 54.21% compared with the original method for posture measurement. The proposed method is 0.971 piece/s slower than the original method in terms of the image processing velocity. But the proposed method is still feasible, and the purpose of posture measurement is achieved.

scholarly journals Extraction and Filter Algorithm of Roll Angular Rate for High Spinning Projectiles

2019 ◽  
Vol 2019 ◽  
pp. 1-15
Author(s):  
Hui Zhao ◽  
Zhong Su ◽  
Fuchao Liu ◽  
Chao Li ◽  
Qing Li ◽  
...  
Keyword(s):  
Measurement Problem ◽  
Frequency Estimation ◽  
Angular Rate ◽  
Estimation Accuracy ◽  
Signal Frequency ◽  
Average Error ◽  
Accurate Estimation ◽  
Test Results ◽  
Narrowband Signal ◽  
Spinning Projectile

The accurate measurement of roll angular rate for high spinning projectile has long been a challenging problem. Aiming to obtain the accurate roll angular rate of high spinning projectile, a novel extraction and filter algorithm, BSCZT-KF, is proposed in this paper. Firstly, a compound angular motion model of high spinning projectile is established. According to the model, we translate the roll angular rate measurement problem into a frequency estimation problem. Then the improved CZT algorithm, BSCZT, was employed to realize an accurate estimation of the narrowband signal frequency. Combined with the peak detection method, the BSCZT-KF algorithm is presented to further enhance the frequency estimation accuracy and the real-time performance. Finally, two sets of actual flight tests were conducted to verify the effectiveness and accuracy of the algorithm. The test results show that the average error of estimated roll angular rate is about 0.095% of the maximum of roll angular rate. Compared with the existing methods, the BSCZT-KF has the highest frequency estimation accuracy for narrowband signal.

scholarly journals Infrastructure-Free Indoor Pedestrian Tracking with Smartphone Acoustic-Based Enhancement

Sensors ◽  
2019 ◽  
Vol 19 (11) ◽  
pp. 2458 ◽  
Author(s):  
Chao Liu ◽  
Sining Jiang ◽  
Shuo Zhao ◽  
Zhongwen Guo
Keyword(s):  
Doppler Effect ◽  
Location Estimation ◽  
Location Based Services ◽  
Estimation Accuracy ◽  
Measurement Unit ◽  
Average Error ◽  
Pedestrian Tracking ◽  
Tracking Accuracy ◽  
Acoustic Frequency ◽  
Indoor Pedestrian Tracking

Indoor pedestrian tracking has been identified as a key technology for indoor location-based services such as emergency locating, advertising, and gaming. However, existing smartphone-based approaches to pedestrian tracking in indoor environments have various limitations including a high cost of infrastructure constructing, labor-intensive fingerprint collection, and a vulnerability to moving obstacles. Moreover, our empirical study reveals that the accuracy of indoor locations estimated by a smartphone Inertial Measurement Unit (IMU) decreases severely when the pedestrian is arbitrarily wandering with an unstable speed. To improve the indoor tracking performance by enhancing the location estimation accuracy, we exploit smartphone-based acoustic techniques and propose an infrastructure-free indoor pedestrian tracking approach, called iIPT. The novelty of iIPT lies in the pedestrian speed reliability metric, which characterizes the reliability of the pedestrian speed provided by the smartphone IMU, and in a speed enhancing method, where we adjust a relatively less reliable pedestrian speed to the more reliable speed of a passing by “enhancer” based on the acoustic Doppler effect. iIPT thus changes the encountered pedestrians from an“obstacle” into an “enhancer.” Extensive real-world experiments in indoor scenarios have been conducted to verify the feasibility of realizing the acoustic Doppler effect between smartphones and to identify the applicable acoustic frequency range and transmission distance while reducing battery consumption. The experiment results demonstrate that iIPT can largely improve the tracking accuracy and decrease the average error compared with a conventional IMU-based method.

Computing Offsets of NURBS Curve and Surface

2012 ◽  
Vol 542-543 ◽  
pp. 537-540
Author(s):  
Ying Yue ◽  
Jun Jia
Keyword(s):  
Curve Surface ◽  
Computational Time ◽  
Normal Vector ◽  
Unit Normal Vector ◽  
Nurbs Curve ◽  
Vector Direction ◽  
Normal Vectors ◽  
Offset Curve ◽  
Offset Error ◽  
Unit Normal

This paper presents an algorithm for the offsetting of NURBS curve/surface. First the unit normal vectors of the progenitor NURBS curve/surface is computed precisely, then the offset curve/surface can be obtained by offsetting the progenitor curve/surface in the normal vector direction with the required distance. Considerable extra computational time can be saved, especially when they are to be offset by several times. As the method successfully computes the unit normal vector of the progenitors, the offset error of this method is zero. The method can also be generalized to other degree NURBS curve/surface.

scholarly journals Second-order normal vectors to a convex epigraph

1994 ◽  
Vol 50 (1) ◽  
pp. 123-134 ◽  
Author(s):  
Alberto Seeger
Keyword(s):  
Convex Function ◽  
Mathematical Object ◽  
Second Order ◽  
Normal Vector ◽  
Subdifferential Mapping ◽  
Normal Vectors

The second–order behaviour of a nonsmooth convex function f is reflected by the so–called second–order subdifferential mapping ∂2f. This mathematical object has been intensively studied in recent years. Here we study ∂2f in connection with the geometric concept of “second-order normal vector” to the epigraph of f.

scholarly journals A Self-Improving Photosensitizer Discovery System via Bayesian Optimization and Quantum Chemical Calculation

2021 ◽  
Author(s):  
Shidang Xu ◽  
Jiali Li ◽  
Pengfei Cai ◽  
Xiaoli Liu ◽  
Bin Liu ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Quantum Chemical ◽  
Prediction Accuracy ◽  
High Performance ◽  
Materials Science ◽  
Bayesian Optimization ◽  
Average Error ◽  
Chemical Calculation ◽  
Discovery System ◽  
Self Learning

Artificial intelligence (AI) based self-learning or self-improving material discovery system is the holy grail of next-generation material discovery and materials science. Herein, we demonstrate how to combine accurate prediction of material performance via quantum chemical calculations and Bayesian optimization-based active learning to realize a self-improving discovery system for high-performance photosensitizers (PS). Through self-improving cycles, such a system can improve the model prediction accuracy (best mean average error of 0.09 eV for singlet-triplet spitting) and high-performance PS search ability, realizing the efficient discovery of PS. From a molecular space with more than 7 million molecules, 5950 potential high-performance PSs were discovered.

scholarly journals The Geometry of Power Systems Steady-State Equations– Part II: a Power Surface Study

2020 ◽  
pp. 47-53
Author(s):  
B. Ayuev ◽  
V. Davydov ◽  
P. Erokhin ◽  
V. Neuymin ◽  
A. Pazderin
Keyword(s):  
Steady State ◽  
Power Systems ◽  
Transmission Loss ◽  
Normal Vector ◽  
State Equations ◽  
Marginal State ◽  
Normal Vectors ◽  
System States ◽  
Steady State Solutions ◽  
System Power

Steady-state equations play an essential part in the theory of power systems and the practice of computations. These equations are directly or mediately used almost in all areas of the theory of power system states, constituting its basis. This two-part study deals with a geometrical interpretation of steady-state solutions in a power space. Part I has proposed considering the power system's steady states in terms of power surface. Part II is devoted to an analytical study of the power surface through its normal vectors. An interrelationship between the entries of the normal vector is obtained through incremental transmission loss coefficients. Analysis of the normal vector has revealed that in marginal states, its entry of the slack bus active power equals zero, and the incremental transmission loss coefficient of the slack bus equals one. Therefore, any attempts of the slack bus to maintain the system power balance in the marginal state are fully compensated by associated losses. In real-world power systems, a change in the slack bus location in the marginal state makes this steady state non-marginal. Only in the lossless power systems, the marginal states do not depend on a slack bus location.

Sign in / Sign up

Export Citation Format

Share Document