Real-time road surface mapping using stereo matching, v-disparity and machine learning

Azevedo, Vitor Barbirato; Souza, Alberto F. De; Veronese, Lucas P.; Badué, Claudine; Berger, Mariella

doi:10.1109/ijcnn.2013.6707066

Cited by 5 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach allows for a quantitative evaluation of the road surface condition by calculating the area or height of the anomalies represented on the map. Azevedo et al (2013) proposed a road surface mapping system using a stereo camera mounted on a vehicle. In this method, the road surface position relative to the camera is calculated using stereo vision, and a road surface map is acquired using image reprojections and an online-trained road segmentation model.…”

Section: Road Surface Inspection Systemmentioning

confidence: 99%

Three‐dimensionalized feature‐based LiDAR‐visual odometry for online mapping of unpaved road surfaces

Lee,

Kurisu,

Kuriyama

2024

Journal of Field Robotics

View full text Add to dashboard Cite

Automated maintenance and motion planning for unpaved roads are research areas of great interest in the field robotics. Constructing such systems necessitates the development of surface maps for unpaved roads. However, the lack of distinctive features on unpaved roads degrades the performance of light detection and ranging (LiDAR)‐based mapping. To address this problem, this paper proposes three‐dimensionalized feature‐based LiDAR‐visual odometry (TFB odometry) for the online mapping of unpaved road surfaces. TFB odometry introduces a novel interpolation concept to directly estimate the three‐dimensional coordinates of the image features using LiDAR. Furthermore, LiDAR intensity‐weighted motion estimation is proposed to effectively mitigate the effects of dust, which significantly impact the performance of LiDAR. Finally, TFB odometry includes pose graph optimization to efficiently fuse global navigation satellite system data and poses estimated from motion estimation. Through field experiments on unpaved roads, TFB odometry demonstrated successful online full mapping and outperformed other simultaneous localization and mapping methods. Additionally, it demonstrated remarkable performance in accurately mapping road surface anomalies, even in dusty regions.

show abstract

Section: Road Surface Inspection Systemmentioning

confidence: 99%

Three‐dimensionalized feature‐based LiDAR‐visual odometry for online mapping of unpaved road surfaces

Lee,

Kurisu,

Kuriyama

2024

Journal of Field Robotics

View full text Add to dashboard Cite

show abstract

“…Thus, the value of each point u(x, d) in the U-disparity map is the number of scene points at x-coordinate assuming disparity d. The U-disparity and complementary to it the V-disparity maps were proposed and applied in scene depth analysis tasks in [19][20][21]. The U-disparity map appears to be a very efficient representation for localizing scene objects (provided the stereovision camera base is parallel to the ground plane) [22]. An object positioned at a well localized distance features a region of the same value in the disparity map, which results in a unimodal histogram with a strong maximum in the U-disparity map (see Fig.…”

Section: "U-depth" Representationmentioning

confidence: 99%

Interactive sonification of U-depth images in a navigation aid for the visually impaired

Skulimowski

Owczarek

Radecki

et al. 2018

J Multimodal User Interfaces

View full text Add to dashboard Cite

In this paper we propose an electronic travel aid system for the visually impaired that utilizes interactive sonification of U-depth maps of the environment. The system is comprised of a depth sensor connected to a mobile device and a dedicated application for segmenting depth images and converting them into sounds in real time. An important feature of the system is that the user can interactively select the 3D scene region for sonification by simple touch gestures on the mobile device screen. The sonification scheme is using stereo panning for azimuth angle localization of scene objects, loudness for their size and frequency for distance encoding. Such a sonic representation of 3D scenes allows the user to identify the geometric structure of the environment and determine the distances to potential obstacles. The prototype application was tested by three visually impaired users who managed to successfully perform indoor mobility tasks. The system's usefulness was evaluated quantitatively by means of system usability and task-related questionnaires.

show abstract

“…In (5), K is a 3 × 3 matrix of the intrinsic calibration parameters, where f x and f y are the focal lengths, s is the skew coefficient that models the lens distortion and x 0 and y 0 are the coordinates of the camera centre. R is a 3 × 3 rotation matrix ∈ SO (3) and t is a 3 × 1 translation matrix ((x, y, z) T ∈ R 3 ), which are also called the extrinsic calibration matrices. We need to calibrate the cameras to use them for localisation.…”

Section: Localisationmentioning

confidence: 99%

“…Recently, consumers and researchers have exhibited growing interest in multi-rotor micro unmanned aerial vehicles (MUAVs) such as quad-copters and hex-copters because of the many possible applications of MUAVs-from cinema and other entertainment to surveillance and search and rescue. [1][2][3] The attractiveness of these vehicles comes from their small size, relatively low price and capabilities for accessing zones unreachable to humans. A great deal of research has focussed on the capability of MUAVs to navigate autonomously and to perform mapping without any human-in-theloop control while avoiding obstacles.…”

Section: Introductionmentioning

confidence: 99%

A survey on stereo vision-based autonomous navigation for multi-rotor MUAVs

Sanchez-Rodriguez¹,

Aceves-López²

2018

Robotica

View full text Add to dashboard Cite

SUMMARYThis paper presents an overview of the most recent vision-based multi-rotor micro unmanned aerial vehicles (MUAVs) intended for autonomous navigation using a stereoscopic camera. Drone operation is difficult because pilots need the expertise to fly the drones. Pilots have a limited field of view, and unfortunate situations, such as loss of line of sight or collision with objects such as wires and branches, can happen. Autonomous navigation is an even more difficult challenge than remote control navigation because the drones must make decisions on their own in real time and simultaneously build maps of their surroundings if none is available. Moreover, MUAVs are limited in terms of useful payload capability and energy consumption. Therefore, a drone must be equipped with small sensors, and it must carry low weight. In addition, a drone requires a sufficiently powerful onboard computer so that it can understand its surroundings and navigate accordingly to achieve its goal safely. A stereoscopic camera is considered a suitable sensor because of its three-dimensional (3D) capabilities. Hence, a drone can perform vision-based navigation through object recognition and self-localise inside a map if one is available; otherwise, its autonomous navigation creates a simultaneous localisation and mapping problem.

show abstract

Real-time road surface mapping using stereo matching, v-disparity and machine learning

Cited by 5 publications

References 16 publications

Three‐dimensionalized feature‐based LiDAR‐visual odometry for online mapping of unpaved road surfaces

Three‐dimensionalized feature‐based LiDAR‐visual odometry for online mapping of unpaved road surfaces

Interactive sonification of U-depth images in a navigation aid for the visually impaired

A survey on stereo vision-based autonomous navigation for multi-rotor MUAVs

Contact Info

Product

Resources

About