A Real-Time Hough-Based Method for Segment Detection in Complex Multisensor Images

Foresti, Gian Luca

doi:10.1006/rtim.1999.0179

Cited by 9 publications

(3 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some fast implementations exist such as the real-time Hough segment detection proposed by Foresti [9], fast line segment detection by Bandera [10], or linear time line detector [11]. All of them have drawbacks in case of embedded applications: computation time remains too long to get accurate line positions, and the search for correct local maxima in the Hough space (corresponding to the greatest number of edge points belonging to each line) may lead to confusion between almost similar lines.…”

Section: Related Workmentioning

confidence: 99%

A New Fast Multi-Layer Line Detection For Embedded Robotic Application

Costis¹,

Hugel²,

Bonnin³

2006

IECON 2006 - 32nd Annual Conference on IEEE Industrial Electronics

View full text Add to dashboard Cite

This paper presents a new multi-layer line detection algorithm that can be used for embedded robotic applications. First, contour points are detected along equally spaced columns or rows. Points are stored into a multilayer array. Then the line detection algorithm is run layer by layer. It is based on Wall and Danielson algorithm. Experiments have been carried out on real images taken from a robot's built-in camera. The line detection rate and accuracy are evaluated. This new algorithm features several parameters that can be adjusted for various embedded applications. I. IntroductionIn robotic applications, lines bring information on edges and borders of polygonal regions. If the line detection process is part of an embedded vision system, it must meet strong real-time constraints to keep the image treatment rate to the videorate. It must also return reliable information to both localization and behavior modules of the robot. As a matter of fact, lines are widely used for selflocalization.This paper presents a new multi-layer line detection algorithm that can be used for embedded robotic applications. The algorithm has been implemented on a quadruped robot equipped with a CMOS camera of 208×176 resolution. The main constraints come from a limited computing power. Because of leg impacts on the ground, the images can also be bouncing, which keeps from smoothing techniques. The robustness of the method proposed here is a key factor. The robot must be able to detect lines while moving or under varying lighting conditions. The first part of this paper focuses on fast edge detection and explains how edge points are stored into multi-layer arrays. The second part describes the initial line detection algorithm. Then an improved method is introduced in order to avoid the main drawbacks of the previous one. Both methods are then compared in terms of robustness to noisy edge detection as the robot should be able to detect lines even under bad lighting conditions. The processing time of both treatments is also compared.

show abstract

Section: Related Workmentioning

confidence: 99%

A New Fast Multi-Layer Line Detection For Embedded Robotic Application

Costis¹,

Hugel²,

Bonnin³

2006

IECON 2006 - 32nd Annual Conference on IEEE Industrial Electronics

View full text Add to dashboard Cite

show abstract

“…The segments of the image are extracted from the edge pixels of I. Segment detection is an important aspect in many computer vision applications that has been tackled employing different approaches (Chan and Yip, 1996;Cha et al, 2005;Foresti, 2000;Guil et al, 1995). As the development of a new technique for segment detection is out of the scope of this paper, we have employed the approach explained in Foresti (2000).…”

Section: Visual Fuzzy Door-detectionmentioning

confidence: 99%

“…Segment detection is an important aspect in many computer vision applications that has been tackled employing different approaches (Chan and Yip, 1996;Cha et al, 2005;Foresti, 2000;Guil et al, 1995). As the development of a new technique for segment detection is out of the scope of this paper, we have employed the approach explained in Foresti (2000). The main advantages of this approach are that: (i) the computing time is saved by limiting the voting space to an small range given by the gradient direction of each pixel; and (ii) each pixel votes in its corresponding cell of the Hough space but also in a small window around it.…”

Section: Visual Fuzzy Door-detectionmentioning

confidence: 99%

Detection of doors using a genetic visual fuzzy system for mobile robots

2006

View full text Add to dashboard Cite

Doors are common objects in indoor environments and their detection can be used in robotic tasks such as map-building, navigation and positioning. This work presents a new approach to door-detection in indoor environments using computer vision. Doors are found in gray-level images by detecting the borders of their architraves. A variation of the Hough Transform is used in order to extract the segments in the image after applying the Canny edge detector. Features like length, direction, or distance between segments are used by a fuzzy system to analyze whether the relationship between them reveals the existence of doors. The system has been designed to detect rectangular doors typical of many indoor environments by the use of expert knowledge. Besides, a tuning mechanism based on a genetic algorithm is proposed to improve the performance of the system according to the particularities of the environment in which it is going to be employed. A large database of images containing doors of our building, seen from different angles and distances, has been created to test the performance of the system before and after the tuning process. The system has shown the ability to detect rectangular doors under heavy perspective deformations and it is fast enough to be used for real-time applications in a mobile robot.

show abstract

New Hypothesis Distinctiveness Measure for Better Ellipse Extraction

Wang

Newman

Cao

Lecture Notes in Computer Science

View full text Add to dashboard Cite

A Real-Time Hough-Based Method for Segment Detection in Complex Multisensor Images

Cited by 9 publications

References 32 publications

A New Fast Multi-Layer Line Detection For Embedded Robotic Application

A New Fast Multi-Layer Line Detection For Embedded Robotic Application

Detection of doors using a genetic visual fuzzy system for mobile robots

New Hypothesis Distinctiveness Measure for Better Ellipse Extraction

Contact Info

Product

Resources

About