Optical Normal Flow Estimation on Log-polar Images. A Solution for Real-Time Binocular Vision

doi:10.1006/rtim.1996.0053

Real-Time Imaging

Volume 3, Issue 3, June 1997, Pages 213-228

https://doi.org/10.1006/rtim.1996.0053 Get rights and content

Abstract

This paper is concerned with a computational solution for normal optical flow estimation using space-variant image sampling. The article describes one solution for the problem based on log-polar images, including the algorithm implementation on digital signal processing hardware. The log-polar structure achieves high data compressing rates by maintaining a high resolution area—the fovea—and a space-variant resolution—the retinal periphery. This sampling scheme enables the implementation of selective processing, characteristic of attentive vision. The solution was developed to be used with an active vision system to pursue moving objects, assuming that the cameras could also move. Some properties of the log-poplar structure are also described and the main results of the implementation are presented.

References (0)

Cited by (11)

A review of log-polar imaging for visual perception in robotics
2010, Robotics and Autonomous Systems
Citation Excerpt :
In this case, however, care should be taken to perform the log-polar transformation properly. The literature contains several works that provide hints for log-polar mapping computation [71–74], design or resolution considerations [75–77], and alternative fovea designs [78]. Unlike hardware-based sensors, software-based ones have the unavoidable cost of obtaining the log-polar image from the Cartesian image.
Log-polar imaging consists of a type of methods that represent visual information with a space-variant resolution inspired by the visual system of mammals. It has been studied for about three decades and has surpassed conventional approaches in robotics applications, mainly the ones where real-time constraints make it necessary to utilize resource-economic image representations and processing methodologies. This paper surveys the application of log-polar imaging in robotic vision, particularly in visual attention, target tracking, egomotion estimation, and 3D perception. The concise yet comprehensive review offered in this paper is intended to provide novel and experienced roboticists with a quick and gentle overview of log-polar vision and to motivate vision researchers to investigate the many open problems that still need solving. To help readers identify promising research directions, a possible research agenda is outlined. Finally, since log-polar vision is not restricted to robotics, a couple of other areas of application are discussed.
Similarity motion estimation and active tracking through spatial-domain projections on log-polar images
2005, Computer Vision and Image Understanding
Citation Excerpt :
A gradient-descent search is used in [66] for monocular fixation, where a measure of stability of image regions is defined for coping with occlusions and several image deformations. As in uniformly sampled images, methods to compute the optic flow have also been proposed for log-polar images [62,67,63,44]. In [45], optic flow is combined with region-based segmentation and region matching to passively track a couple of targets without occlusions.
To cope with the huge amount of visual data in the environment, foveal sensing is not only an elegant biological solution, but also an appropriate mechanism in computer-based vision of artificial agents such as robots. An extremely important visual behavior in active vision systems is that of actively tracking moving targets, which is a problem closely related with foveal vision. The work described in this article has precisely to do with active tracking with space-variant images. The two main components of the proposed algorithm are the log-polar image format, and a projection-based motion estimator, whose joint use results in an efficient and effective active tracking scheme. For the successful real-time performance of the proposed approach, it is very important to have an appropriate combination of: (i) an efficient motion estimation technique, (ii) an adequate image representation, and (iii) the dynamic nature of purposive vision.
Performance of three recursive algorithms for fast space-variant Gaussian filtering
2003, Real-Time Imaging
Animal visual systems have solved the problem of limited resources by allocating more processing power to central than peripheral vision. Foveation considerably reduces the amount of data per image by progressively decreasing the resolution at the periphery while retaining a sharp center of interest. This strategy has important applications in the design of autonomous systems for navigation, tracking and surveillance. Central to foveation is a space-variant Gaussian filtering scheme that gradually blurs out details as the distance to the image center increases. Unfortunately Gaussian convolution is a computationally expensive operation, which can severely limit the real-time applicability of foveation. In the space-variant case, the problem is even more difficult as traditional techniques such as the fast Fourier transform cannot be employed because the convolution kernel is different at each pixel. We show that recursive filtering, which was introduced to approximate Gaussian convolution, can be extended to the space-variant case and leads to a very simple implementation that makes it ideal for that application. Three main recursive algorithms have emerged, produced by independent derivation methods. We assess and compare their performance in traditional filtering applications and in our specific space-variant case. All three methods drastically cut down the cost of Gaussian filtering to a limited number of operations per pixel that is independent of the scale selected. In addition we show that two of those algorithms have excellent accuracy in that the output they produce differs from the output obtained performing real Gaussian convolution by less than 1%.
Dealing with 2D translation estimation in log-polar imagery
2003, Image and Vision Computing
Log-polar mapping has been proposed as a very appropriate space-variant imaging model in active vision applications. This biologically inspired model has several advantages, and facilitates some visual tasks. For example, it provides an efficient data reduction, and simplifies rotational and scaling image transformations. However, simple translations become a difficult transform due to the log-polar geometry. There is no doubt about the importance of translation estimation in active visual tracking. Therefore, in this work, the problem of translation estimation in log-polar images is tackled. Two different approaches are presented, and their performances are evaluated and compared. One approach uses a gradient descent for minimizing a dissimilarity measure, while the other converts the 2D problem into two simpler 1D problems, by using projections. As the experimental results reveal, this second approach, besides being more efficient, can deal with larger translations than the gradient-based search can.
Fast panoramic image mosaicing using one-dimensional flow estimation
2002, Real-Time Imaging
In this paper, we propose a fast method for panoramic image mosaicing using one-dimensional (1D) flow estimation. First, the 1D flow vectors are estimated at each pixel in x and y directions. Then the global affine parameters are estimated from 1D flows using the iterative least-squares method. The accurate parameters are obtained by iteratively repeated optimization in x and y directions. We have implemented the proposed method on a conventional low-cost PC. The experimental results show that the image mosaicing is possible in near-real time.
Image Analysis and Computer Vision: 1997
1998, Computer Vision and Image Understanding
This paper presents a bibliography of nearly 1700 references related to computer vision and image analysis, arranged by subject matter. The topics covered include computational techniques; feature detection and segmentation; image and scene analysis; two-dimensional shape; pattern; color and texture; matching and stereo; three-dimensional recovery and analysis; three-dimensional shape; and motion. A few references are also given on related topics, including geometry and graphics, compression and processing, sensors and optics, visual perception, neural networks, artificial intelligence and pattern recognition, as well as on applications.

View all citing articles on Scopus

View full text

Regular ArticleOptical Normal Flow Estimation on Log-polar Images. A Solution for Real-Time Binocular Vision

Abstract

Regular Article
Optical Normal Flow Estimation on Log-polar Images. A Solution for Real-Time Binocular Vision