Depth Prediction without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos

Learning to predict scene depth from RGB inputs is a challenging task both for indoor and outdoor robot navigation. In this work we address unsupervised learning of scene depth and robot ego-motion where supervision is provided by monocular videos, as cameras are the cheapest, least restrictive and most ubiquitous sensor for robotics. Previous work in unsupervised image-to-depth learning has established strong baselines in the domain. We propose a novel approach which produces higher quality results, is able to model moving objects and is shown to transfer across data domains, e.g. from outdoors to indoor scenes. The main idea is to introduce geometric structure in the learning process, by modeling the scene and the individual objects; camera ego-motion and object motions are learned from monocular videos as input. Furthermore an online refinement method is introduced to adapt learning on the fly to unknown domains. The proposed approach outperforms all state-of-the-art approaches, including those that handle motion e.g. through learned flow. Our results are comparable in quality to the ones which used stereo as supervision and significantly improve depth prediction on scenes and datasets which contain a lot of object motion. The approach is of practical relevance, as it allows transfer across environments, by transferring models trained on data collected for robot navigation in urban scenes to indoor navigation settings. The code associated with this paper can be found at https://sites.google.com/view/struct2depth.

Download Full-text

Principles of adaptive element spacing in linear array antennas

Scientific Reports ◽

10.1038/s41598-021-84874-7 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Tanzeela Mitha ◽

Maria Pour

Keyword(s):

Antenna Arrays ◽

Communication Systems ◽

Linear Array ◽

Main Idea ◽

Phase Center ◽

Array Configuration ◽

Array Antennas ◽

Center Location ◽

Novel Approach ◽

Array Performance

AbstractA novel approach to linear array antennas with adaptive inter-element spacing is presented for the first time. The main idea is based upon electronically displacing the phase center location of the antenna elements, which determine their relative coordinates in the array configuration. This is realized by employing dual-mode microstrip patch antennas as a constitutive element, whose phase center location can be displaced from its physical center by simultaneously exciting two modes. The direction and the amount of displacement is controlled by the amplitude and phase of the modes at the element level. This in turn facilitates reconfiguring the inter-element spacing at the array level. For instance, a uniformly-spaced array could be electronically transformed into a non-uniform one without any mechanical means. The proposed idea is demonstrated in two- and three-element linear antenna arrays. The technique has the potential to control the radiation characteristics such as sidelobe levels, position of the nulls, and the beamwidths in small arrays, which are useful for adaptively controlling the array performance in emerging wireless communication systems and radars.

Download Full-text

Extracting Stops from Spatio-Temporal Trajectories within Dynamic Contextual Features

Sustainability ◽

10.3390/su13020690 ◽

2021 ◽

Vol 13 (2) ◽

pp. 690

Author(s):

Tao Wu ◽

Huiqing Shen ◽

Jianxin Qin ◽

Longgang Xiang

Keyword(s):

Moving Objects ◽

Major Effect ◽

Location Based Services ◽

Trajectory Data ◽

The Road ◽

Contextual Features ◽

Novel Approach ◽

Gps Trajectories ◽

Time Distance ◽

Spatio Temporal

Identifying stops from GPS trajectories is one of the main concerns in the study of moving objects and has a major effect on a wide variety of location-based services and applications. Although the spatial and non-spatial characteristics of trajectories have been widely investigated for the identification of stops, few studies have concentrated on the impacts of the contextual features, which are also connected to the road network and nearby Points of Interest (POIs). In order to obtain more precise stop information from moving objects, this paper proposes and implements a novel approach that represents a spatio-temproal dynamics relationship between stopping behaviors and geospatial elements to detect stops. The relationship between the candidate stops based on the standard time–distance threshold approach and the surrounding environmental elements are integrated in a complex way (the mobility context cube) to extract stop features and precisely derive stops using the classifier classification. The methodology presented is designed to reduce the error rate of detection of stops in the work of trajectory data mining. It turns out that 26 features can contribute to recognizing stop behaviors from trajectory data. Additionally, experiments on a real-world trajectory dataset further demonstrate the effectiveness of the proposed approach in improving the accuracy of identifying stops from trajectories.

Download Full-text

Sensor Fusion-Based Approach to Eliminating Moving Objects for SLAM in Dynamic Environments

Sensors ◽

10.3390/s21010230 ◽

2021 ◽

Vol 21 (1) ◽

pp. 230

Author(s):

Xiangwei Dang ◽

Zheng Rong ◽

Xingdong Liang

Keyword(s):

State Estimation ◽

Autonomous Navigation ◽

Moving Objects ◽

Physical Simulation ◽

Point Clouds ◽

Dynamic Environments ◽

Accurate Localization ◽

Novel Approach ◽

Localization And Mapping ◽

Real World Datasets

Accurate localization and reliable mapping is essential for autonomous navigation of robots. As one of the core technologies for autonomous navigation, Simultaneous Localization and Mapping (SLAM) has attracted widespread attention in recent decades. Based on vision or LiDAR sensors, great efforts have been devoted to achieving real-time SLAM that can support a robot’s state estimation. However, most of the mature SLAM methods generally work under the assumption that the environment is static, while in dynamic environments they will yield degenerate performance or even fail. In this paper, first we quantitatively evaluate the performance of the state-of-the-art LiDAR-based SLAMs taking into account different pattens of moving objects in the environment. Through semi-physical simulation, we observed that the shape, size, and distribution of moving objects all can impact the performance of SLAM significantly, and obtained instructive investigation results by quantitative comparison between LOAM and LeGO-LOAM. Secondly, based on the above investigation, a novel approach named EMO to eliminating the moving objects for SLAM fusing LiDAR and mmW-radar is proposed, towards improving the accuracy and robustness of state estimation. The method fully uses the advantages of different characteristics of two sensors to realize the fusion of sensor information with two different resolutions. The moving objects can be efficiently detected based on Doppler effect by radar, accurately segmented and localized by LiDAR, then filtered out from the point clouds through data association and accurate synchronized in time and space. Finally, the point clouds representing the static environment are used as the input of SLAM. The proposed approach is evaluated through experiments using both semi-physical simulation and real-world datasets. The results demonstrate the effectiveness of the method at improving SLAM performance in accuracy (decrease by 30% at least in absolute position error) and robustness in dynamic environments.

Download Full-text

A Novel Approach for Robust Perceptual Image Hashing

Computer and Information Science ◽

10.5539/cis.v14n3p38 ◽

2021 ◽

Vol 14 (3) ◽

pp. 38

Author(s):

Azhar Hadmi ◽

Awatif Rouijel

Keyword(s):

Side Information ◽

Main Idea ◽

Jpeg Compression ◽

Image Hashing ◽

Novel Approach ◽

Perceptual Hash ◽

Short Signature ◽

Robust Image ◽

Tampered Image ◽

The Stability

Perceptual image hashing system generates a short signature called perceptual hash attached to an image before transmission and acts as side information for analyzing the trustworthiness of the received image. In this paper, we propose a novel approach to improve robustness for perceptual image hashing scheme for generating a perceptual hash that should be resistant to content-preserving manipulations, such as JPEG compression and Additive white Gaussian noise (AWGN) also should differentiate the maliciously tampered image and its original version. Our algorithm first constructs a robust image, derived from the original input by analyzing the stability of the extracted features and improving their robustness. From the robust image, which does perceptually resemble the original input, we further extract the final robust features. Next, robust features are suitably quantized allowing the generation of the final perceptual hash using the cryptographic hash function SHA1. The main idea of this paper is to transform the original image into a more robust one that allows the extraction of robust features. Generation of the robust image turns out be quite important since it introduces further robustness to the perceptual image hashing system. The paper can be seen as an attempt to propose a general methodology for more robust perceptual image hashing. The experimental results presented in this paper reveal that the proposed scheme offers good robustness against JPEG compression and Additive white Gaussian noise.

Download Full-text

Explaining Algorithms

Intelligent Information Technologies ◽

10.4018/978-1-59904-941-0.ch131 ◽

2011 ◽

pp. 2261-2279

Author(s):

Tomasz Muldner ◽

Elhadi Shakshuki

Keyword(s):

Distance Education ◽

Main Idea ◽

Top Down ◽

Levels Of Abstraction ◽

Visualization System ◽

Client Server ◽

Novel Approach ◽

Visualization Systems ◽

Explanation System ◽

Server Architecture

This article presents a novel approach for explaining algorithms that aims to overcome various pedagogical limitations of the current visualization systems. The main idea is that at any given time, a learner is able to focus on a single problem. This problem can be explained, studied, understood, and tested, before the learner moves on to study another problem. Toward this end, a visualization system that explains algorithms at various levels of abstraction has been designed and implemented. In this system, each abstraction is focused on a single operation from the algorithm using various media, including text and an associated visualization. The explanations are designed to help the user to understand basic properties of the operation represented by this abstraction, for example its invariants. The explanation system allows the user to traverse the hierarchy graph, using either a top-down (from primitive operations to general operations) approach or a bottom-up approach. Since the system is implemented using a client-server architecture, it can be used both in the classroom setting and through distance education.

Download Full-text

A Novel Approach for the Detection of Developing Thunderstorm Cells

Remote Sensing ◽

10.3390/rs11040443 ◽

2019 ◽

Vol 11 (4) ◽

pp. 443 ◽

Cited By ~ 6

Author(s):

Richard Müller ◽

Stéphane Haussler ◽

Matthias Jerg ◽

Dirk Heizenreder

Keyword(s):

Water Vapor ◽

Main Idea ◽

Probability Of Detection ◽

Base Case ◽

Convective Clouds ◽

Critical Success ◽

Novel Approach ◽

Forecast Lead Time ◽

Spatio Temporal ◽

Atmospheric Motion Vectors

This study presents a novel approach for the early detection of developing thunderstorms. To date, methods for the detection of developing thunderstorms have usually relied on accurate Atmospheric Motion Vectors (AMVs) for the estimation of the cooling rates of convective clouds, which correspond to the updraft strengths of the cloud objects. In this study, we present a method for the estimation of the updraft strength that does not rely on AMVs. The updraft strength is derived directly from the satellite observations in the SEVIRI water vapor channels. For this purpose, the absolute value of the vector product of spatio-temporal gradients of the SEVIRI water vapor channels is calculated for each satellite pixel, referred to as Normalized Updraft Strength (NUS). The main idea of the concept is that vertical updraft leads to NUS values significantly above zero, whereas horizontal cloud movement leads to NUS values close to zero. Thus, NUS is a measure of the strength of the vertical updraft and can be applied to distinguish between advection and convection. The performance of the method has been investigated for two summer periods in 2016 and 2017 by validation with lightning data. Values of the Critical Success Index (CSI) of about 66% for 2016 and 60% for 2017 demonstrate the good performance of the method. The Probability of Detection (POD) values for the base case are 81.8% for 2016 and 89.2% for 2017, respectively. The corresponding False Alarm Ratio (FAR) values are 22.6% (2016) and 36.4% (2017), respectively. In summary, the method has the potential to reduce forecast lead time significantly and can be quite useful in regions without a well-maintained radar network.

Download Full-text

PSO based path planner of an autonomous mobile robot

Open Computer Science ◽

10.2478/s13537-012-0009-5 ◽

2012 ◽

Vol 2 (2) ◽

Cited By ~ 12

Author(s):

B. Deepak ◽

Dayal Parhi

Keyword(s):

Mobile Robot ◽

Robot Navigation ◽

Fitness Function ◽

Optimal Path ◽

Mobile Robot Navigation ◽

Autonomous Mobile Robot ◽

Swarm Optimization ◽

Novel Approach ◽

Fitness Value ◽

Path Planner

AbstractA novel approach based on particle swarm optimization has been presented in this paper for solving mobile robot navigation task. The proposed technique tries to optimize the path generated by an intelligent mobile robot from its source position to destination position in its work space. For solving this problem, a new fitness function has been modelled, which satisfies the obstacle avoidance and optimal path traversal conditions. From the obtained fitness values of each particle in the swarm, the robot moves towards the particle which is having optimal fitness value. Simulation results are provided to validate the feasibility of the developed methodology in various unknown environments.

Download Full-text

Metabolomic Markers for the Early Selection of Coffea canephora Plants with Desirable Cup Quality Traits

Metabolites ◽

10.3390/metabo9100214 ◽

2019 ◽

Vol 9 (10) ◽

pp. 214 ◽

Cited By ~ 3

Author(s):

Gamboa-Becerra ◽

Hernández-Hernández ◽

González-Ríos ◽

Suárez-Quiroz ◽

Gálvez-Ponce ◽

...

Keyword(s):

Genetically Modified Organisms ◽

Main Idea ◽

Coffea Canephora ◽

Development Stage ◽

Quality Traits ◽

Conventional Breeding ◽

Novel Approach ◽

Coffee Plants ◽

Selection Of ◽

Cup Quality

Genetic improvement of coffee plants represents a great challenge for breeders. Conventional breeding takes a too long time for responding timely to market demands, climatic variations and new biological threads. The correlation of genetic markers with the plant phenotype and final product quality is usually poor. Additionally, the creation and use of genetically modified organisms (GMOs) are often legally restricted and rejected by customers that demand natural products. Therefore, we developed a non-targeted metabolomics approach to accelerate conventional breeding. Our main idea was to identify highly heritable metabolites in Coffea canephora seedlings, which are linked to coffee cup quality. We employed a maternal half-sibs approach to estimate the metabolites heritability in open-pollinated plants in both leaves and fruits at an early plant development stage. We evaluated the cup quality of roasted beans and correlated highly heritable metabolites with sensory quality traits of the coffee beverage. Our results provide new insights about the heritability of metabolites of C. canephora plants. Furthermore, we found strong correlations between highly heritable metabolites and sensory traits of coffee beverage. We revealed metabolites that serve as predictive metabolite markers at an early development stage of coffee plants. Informed decisions can be made on plants of six months old, compared to 3.5 to 5 years using conventional selection methods. The metabolome-wide association study (MWAS) drastically accelerates the selection of C. canephora plants with desirable characteristics and represents a novel approach for the focused breeding of crops.

Download Full-text

Moving Object Detection Using an Object Motion Reflection Model of Motion Vectors

Symmetry ◽

10.3390/sym11010034 ◽

2019 ◽

Vol 11 (1) ◽

pp. 34 ◽

Cited By ~ 3

Author(s):

Jisang Yoo ◽

Gyu-cheol Lee

Keyword(s):

Object Detection ◽

Moving Objects ◽

Probability Model ◽

Motion Vector ◽

Moving Object Detection ◽

Object Motion ◽

Moving Object ◽

Disparity Map ◽

Motion Vectors ◽

Reflection Model

Moving object detection task can be solved by the background subtraction algorithm if the camera is fixed. However, because the background moves, detecting moving objects in a moving car is a difficult problem. There were attempts to detect moving objects using LiDAR or stereo cameras, but when the car moved, the detection rate decreased. We propose a moving object detection algorithm using an object motion reflection model of motion vectors. The proposed method first obtains the disparity map by searching the corresponding region between stereo images. Then, we estimate road by applying v-disparity method to the disparity map. The optical flow is used to acquire the motion vectors of symmetric pixels between adjacent frames where the road has been removed. We designed a probability model of how much the local motion is reflected in the motion vector to determine if the object is moving. We have experimented with the proposed method on two datasets, and confirmed that the proposed method detects moving objects with higher accuracy than other methods.

Download Full-text