scispace - formally typeset
Search or ask a question

Showing papers on "Subpixel rendering published in 1986"


Journal ArticleDOI
TL;DR: A system that takes a gray level image as input, locates edges with subpixel accuracy, and links them into lines and notes that the zero-crossings obtained from the full resolution image using a space constant ¿ for the Gaussian, are very similar, but the processing times are very different.
Abstract: We present a system that takes a gray level image as input, locates edges with subpixel accuracy, and links them into lines. Edges are detected by finding zero-crossings in the convolution of the image with Laplacian-of-Gaussian (LoG) masks. The implementation differs markedly from M.I.T.'s as we decompose our masks exactly into a sum of two separable filters instead of the usual approximation by a difference of two Gaussians (DOG). Subpixel accuracy is obtained through the use of the facet model [1]. We also note that the zero-crossings obtained from the full resolution image using a space constant ? for the Gaussian, and those obtained from the 1/n resolution image with 1/n pixel accuracy and a space constant of ?/n for the Gaussian, are very similar, but the processing times are very different. Finally, these edges are grouped into lines using the technique described in [2].

502 citations


Journal ArticleDOI
TL;DR: In this paper, a series of one-dimensional surfaces are fit to each window and the surface description is accepted, which is adequate in the least square sense and has the fewest parameters.
Abstract: An edge in an image corresponds to a discontinuity in the intensity surface of the underlying scene. It can be approximated by a piecewise straight curve composed of edgels, i.e., short, linear edge-elements, each characterized by a direction and a position. The approach to edgel-detection here, is to fit a series of one-dimensional surfaces to each window (kernel of the operator) and accept the surface-description which is adequate in the least squares sense and has the fewest parameters. (A one-dimensional surface is one which is constant along some direction.) The tanh is an adequate basis for the stepedge and its combinations are adequate for the roofedge and the line-edge. The proposed method of step-edgel detection is robust with respect to noise; for (step-size/?noise) ? 2.5, it has subpixel position localization (?position < ?) and an angular localization better than 10°; further, it is designed to be insensitive to smooth shading. These results are demonstrated by some simple analysis, statistical data, and edgelimages. Also included is a comparison of performance on a real image, with a typical operator (Difference-of-Gaussians). The results indicate that the proposed operator is superior with respect to detection, localization, and resolution.

495 citations


Journal ArticleDOI
TL;DR: To register two images from the same scene, first, the images are segmented and closedboundary regions in the image are extracted, which enables determination of centers of gravity of the regions up to subpixel accuracy.
Abstract: Automatic registration of images with translational, rotational, and scaling differences is discussed. To register two images from the same scene, first, the images are segmented and closedboundary regions in the images are extracted. Next, centers of gravity of closed-boundary regions are taken as control points and correspondence is established between the control points. Using this correspondence, the original images are then revisited and the segmentation process is refined in such a way that the obtained corresponding regions become optimally similar. This enables determination of centers of gravity of the regions up to subpixel accuracy. Finally, registration parameters are determined by the least squares error criterion.

227 citations


Proceedings ArticleDOI
01 Apr 1986
TL;DR: An efficient algorithm to locate corners and at the same time encode curve segments between them using B-Splines is developed, which is a good approximation of the original, and achieves significant data compression.
Abstract: In this paper, we propose to use B-Splines to represent digital curves. We have developed an efficient algorithm to locate corners and at the same time encode curve segments between them using B-Splines. Used in conjunction with our subpixel edge detector, [1], it allows us to obtain accurate position of the corners, as needed in many registration problems such as stereo matching and motion parameter estimation. In addition to corners, we detect points of significant curvature between them. The resulting representation is a good approximation of the original, in the sense that it makes interesting points explicit, and achieves significant data compression.

71 citations


Journal ArticleDOI
TL;DR: An analysis of decrease of accuracy in passive stereoscopic vision shows that in order to obtain absolute depth estimates useful for robotics (Δz < 1%) it is necessary to measure all mechanical parameters with an extremely high precision, to reach subpixel accuracy, and to match features between two rather different images.
Abstract: An analysis of decrease of accuracy in passive stereoscopic vision shows that in order to obtain absolute depth estimates useful for robotics (Δz < 1%) it is necessary to measure all mechanical parameters with an extremely high precision, to reach subpixel accuracy, and to match features between two rather different images.

70 citations


Patent
31 Dec 1986
TL;DR: In this paper, a serial serpentine pattern is used to test all row or all column bus lines at one time, and after testing, the serial connections are broken, and the subpixels also can be formed with common row and column buses lines.
Abstract: Subdivided pixels are provided with interconnected and hence redundant row and column bus lines to reduce fatal defects. The respective redundant row and column lines also can be interconnected between subpixels to further reduce defects. One defective subpixel is generally an acceptable non-fatal defect, since the rest of the subpixels are still operative. The subpixels also can be formed with common row and column bus lines. The pixels or subpixels can be connected in a serial serpentine pattern to test all row or all column bus lines at once. After testing, the serial connections are broken.

56 citations


01 Mar 1986
TL;DR: The results indicate that the proposed operator is superior with respect to detection, localization, and resolution.
Abstract: An edge in an image corresponds to a discontinuity in the intensity surface of the underlying scene. It can be approximated by a piecewise straight curve composed of edgels, i.e., short, linear edge-elements, each characterized by a direction and a position. The approach to edgel-detection here, is to fit a series of one-dimensional surfaces to each window (kernel of the operator) and accept the surface-description which is adequate in the least squares sense and has the fewest parameters. (A one-dimensional surface is one which is constant along some direction.) The tanh is an adequate basis for the stepedge and its combinations are adequate for the roofedge and the line-edge. The proposed method of step-edgel detection is robust with respect to noise; for (step-size/?noise) ? 2.5, it has subpixel position localization (?position < ?) and an angular localization better than 10°; further, it is designed to be insensitive to smooth shading. These results are demonstrated by some simple analysis, statistical data, and edgelimages. Also included is a comparison of performance on a real image, with a typical operator (Difference-of-Gaussians). The results indicate that the proposed operator is superior with respect to detection, localization, and resolution.

34 citations


Patent
01 Jul 1986
TL;DR: In this paper, a real-time polar controlled video processor electronically rotates scene nformation in a real time video processor by dynamically specifying the origin of each desired consecutive output line from subpixel space, and by specifying the delta "X" and delta "Y" inputs (of X-Y Cartesian coordinates) to a realtime coefficient generator.
Abstract: The real-time polar controlled video processor electronically rotates scene nformation in a real-time video processor. By dynamically specifying the origin of each desired consecutive output line from subpixel space, and by specifying the delta "X" and delta "Y" inputs (of X-Y Cartesian coordinates) to a real-time coefficient generator, a subpixel generator can sweep through a stored image database at any arbitrary angle and scale factor. Subpixels which are generated sequentially describe any image size, orientation, and position. This scheme allows dynamic line-by-line accommodation of dynamically rotating sensor outputs.

2 citations


01 Nov 1986
TL;DR: In this article, the authors summarized the various approaches relevant to estimating canopy cover at subpixel resolution, based on physical models of radiative transfer in non-homogeneous canopies and on empirical methods.
Abstract: The present report summarizes the various approaches relevant to estimating canopy cover at subpixel resolution. The approaches are based on physical models of radiative transfer in non-homogeneous canopies and on empirical methods. The effects of vegetation shadows and topography are examined. Simple versions of the model are tested, using the Taos, New Mexico Study Area database. Emphasis has been placed on using relatively simple models requiring only one or two bands. Although most methods require some degree of ground truth, a two-band method is investigated whereby the percent cover can be estimated without ground truth by examining the limits of the data space. Future work is proposed which will incorporate additional surface parameters into the canopy cover algorithm, such as topography, leaf area, or shadows. The method involves deriving a probability density function for the percent canopy cover based on the joint probability density function of the observed radiances.

01 Jan 1986
TL;DR: An asystem that takes agraylevel image input, locates edges with subpixel accuracy, andlinks them into lines, which differs markedly from M.I.T.’s DOG.
Abstract: We present asystem that takes agraylevel imageasinput, locates edges withsubpixel accuracy, andlinks theminto lines. Edges aredetected byfinding zero-crossings intheconvolution oftheimage withLaplacian-of-Gaussian (LoG)masks. Theimplementation differs markedly fromM.I.T. 'saswedecompose ourmasksexactly into asum oftwoseparable filters instead oftheusual approximation byadiffer- enceoftwoGaussians (DOG).Subpixel accuracy isobtained through theuseofthefacet model(1). We alsonotethatthezero-crossings obtained fromthefull resolution imageusing aspaceconstant afor theGaussian, andthose obtained fromthe1/nresolution imagewith llnpixel accuracy andaspaceconstant ofalnfortheGaussian, are verysimilar, buttheprocessing times areverydifferent. Finally, these edges aregrouped into lines using thetechnique described in(2). IndexTerms-Edge operator, imageprocessing, imagesegmenta- tion, subpixel accuracy edgedetection, zero-crossings ofsecond deriv- ative.

01 Jan 1986
TL;DR: Automatic registration of images with translational, ro- tational, and scaling differences is discussed in this paper, where the center of gravity of closed boundary regions are taken as control points and correlation is established between the control points.
Abstract: Automatic registration ofimages withtranslational, ro- tational, andscaling differences isdiscussed. Toregister twoimages fromthesamescene, first, theimages aresegmented andclosed- boundary regions intheimages areextracted. Next, centers ofgravity ofclosed-boundary regions aretakenascontrol points andcorrespon- denceisestablished between thecontrol points. Using this correspon- dence, theoriginal images arethenrevisited andthesegmentation pro- cessisrefined insuchawaythattheobtained corresponding regions becomeoptimally similar. Thisenables determination ofcenters of gravity oftheregions uptosubpixel accuracy. Finally, registration parameters aredetermined bytheleast squares errorcriterion. Keywords-Image registration, control points, least squares error criterion, subpixel accuracy, satellite imagery.