Wyniki wyszukiwania - BazTech

1

Corner-based keypoints for scale-invariant detection of partially visible objects

Śluzek A.

Machine Graphics and Vision

|

2006

|

Vol. 15, No. 3/4

639-647

EN

Local features (also known as interest points, keypoints, etc.) are a popular and powerful tool for matching images and detecting partially occluded objects. While the problems of photometric distortions of images and rotational invariance of the features have satisfactory solutions, satisfactorily simple scale-invariant algorithms do not exist yet. Generally, either computationally complex methods of scale-space (multi-scale approach) are used, or the correct scale is estimated using additional mechanisms. The paper proposes a new category of keypoints that can be used to develop a simple scale-invariant method for detecting known objects in analyzed images. Keypoints are defined as locations at which selected moment-based parameters are consistent over a wide range of different-size circular patches around the keypoint. While the database of known objects (i.e. the keypoints and their descriptions) is still built using a multi-scale approach, analyzed images are scanned using only a single-scale window and its sub-window. The paper focuses on the keypoint building and keypoint matching principles. Higher-level issues of hypotheses building and verification (regarding the presence of objects in analyzed images) are only briefly discussed.

2

Interial moments of generated TIN models for surface matching

Chinh Ke L.

Geodezja i Kartografia

|

2004

|

T. 53, z. 3-4

159-171

EN

The purpose of surface matching is to determine transformation parameters without known corresponding points for two data sets of spatial point coordinates obtained with use of different sensors. Instead of different features auch as points of interest, lines, surface patches in the TIN (Triangle Irregular Network) or DEM model are used. The paper presents an approach of using interial moments of TIN models generated from two data sets of same terrain for surface matching. The interial moments could easily be calculated for each triangle in the TIN using formulae given. Three moment invariants /2,/max,/min that are used as the features of high level for surface matching are defined.

PL

Istotny problem dopasowania dwóch powierzchni generowanych w postaci TIN lub DEM na podstawie zbiorów punktów przestrzennych, otrzymanych przy pomocy różnych sensorów w różnych układach, polega na wyznaczeniu parametrów transformacji pomiędzy nimi bez wspólnych punktów w obu układach. Dla rozwiązania tego problemu wykorzystane są elementy powierzchniowe w formie trójkątów (model TIN) albo kwadratów (DEM). Praca przedstawia propozycję analitycznego rozwiązania zadania "dopasowania" (matching) dwóch powierzchni na podstawie wykorzystania momentów bezwładności sieci generowanych trójkątów (model TIN). Proces dopasowania dwóch powierzchni zostaje zrealizowany na zasadzie warunku minimalizacji "długości momentów" liczonych pomiędzy i-tym modelem TIn pierwszej powierzchni i j-tym modelem TIN drugiej powierzchni na trzech płaszczyznach OXY, OXZ, OYZ. W ten sposób parametry transformacji zostaną zrealizowane.

3

Moment-based contour segmentation using multiple segmentation primitives

Śluzek A.

Machine Graphics and Vision

|

1998

|

Vol. 7, No. 1/2

269-279

EN

The paper presents a new technique of segmentating digital contours using multiple segmentation primitives. Therefore segmentation results are a combination of line segments, arcs. corners,etc., depending on how well these shapesmath the orginal contour. In yhr first step, the prospective instances of all avaible segmentation primitives are detected within the input digital contours. Then, the detected instances are ranked according to how accurately they fit the corresponding fragment of the contour. Finally, the top-rank segmantation primitives are selected one by one until the whole contour is approximeted. The algorithms has relatively low computational complexity, and it allows paraqllel implementation. Moreover, the algorithm is not sensitive to the performance of edge detectors so that similar results are produced no matter what edge detector has extracted contours from the original image. Therefore, if the quality of camera-captured images is satisfactofily high, the algorithm can analyse them whitout any pre- processing required (except, of course, edge detection).