An experimental comparison of range image segmentation algorithms

Hoover, Adam; Jean-Baptiste, G.; Jiang, Xiaoyi; Flynn, Patrick J.; Bunke, Horst; Goldgof, Dmitry B.; Bowyer, Kevin W.; Eggert, Daniel; Fitzgibbon, Andrew; Fisher, Robert B.

doi:10.1109/34.506791

Cited by 651 publications

(503 citation statements)

References 32 publications

Supporting

Mentioning

487

Contrasting

Unclassified

Order By: Relevance

“…For a literature review of pose determination using range data see Greenspan [31]. Hoover et al [38] did an experimental comparison of range image segmentation algorithms up to 1996. For a discussion on recognizing 3D objects using surface descriptions see Fan et al [23], Liang and Todhunter [59], Huttenlocher and Ullman [43], Wurster et al [88], Johnson, et al [46], Caelli et al [13], Zhang and Hebert [89], Johnson and Hebert [47].…”

Section: Object Identificationmentioning

confidence: 99%

Construction object identification from LADAR scans:

Gilsinn¹,

Witzgall²,

Cheok³

et al. 2005

View full text Add to dashboard Cite

Laser Scanning devices (a.k.a. LADAR for Laser Detection and Ranging) are primarily used in construction projects to capture as-built data for documentation and re-engineering. These systems can rapidly generate unstructured point clouds of scene data within the field of view and line-of-sight of the LADAR. The Construction Metrology and Automation Group (CMAG) at NIST is interested in using these high-resolution scanned data to identify and locate construction objects of interest. The test operation is large-scale pick-and-place of structural steel, where the data from the scans are used to identify the general pose, position and orientation of a target object for initial positioning of a robotic crane. Close-in control of the crane using high framerate flash LADAR is a separate but related area of CMAG research.This initial study describes an experiment in which an I-beam on a concrete floor surface is scanned, and the resulting point cloud data are used to calculate its pose. Multiple scans of two sizes of I-beams were taken with the I-beam placed at various angles relative to the scanner. Two approaches for segmenting potential target objects are described: binning and triangular irregular networks (TINs). Using the method of principal components analysis, the pose of potential target objects is determined. Bounding boxes are then formed around these objects and compared to an ideal bounding box generated from the known geometric specifications of the Ibeam of interest. A measure of best fit is used to determine which scanned object most closely fits the bounding box of the ideal I-beam. A separate laser-based site measurement system (SMS) was used to measure points on the I-beams to form reference data for estimating the closeness of fit of computed pose of the I-beam to measured pose of the I-beam. Three spheres were located in the scan field as a means of registering the scan and SMS axes to one another.

show abstract

Section: Object Identificationmentioning

confidence: 99%

Construction object identification from LADAR scans:

Gilsinn¹,

Witzgall²,

Cheok³

et al. 2005

View full text Add to dashboard Cite

show abstract

“…It is well known that evaluating segmentation results and comparing segmentation algorithms are not simple tasks [4,[36][37][38][39]. However, one of the most widely used criteria for performance evaluation is whether the system can outline the desired or important regions in the image.…”

Section: Comparisons With Other Segmentation Methodsmentioning

confidence: 99%

A clustering fuzzy approach for image segmentation

Cinque

Foresti

Lombardi³

2004

Pattern Recognition

View full text Add to dashboard Cite

Segmentation is a fundamental step in image description or classiÿcation. In recent years, several computational models have been used to implement segmentation methods but without establishing a single analytic solution. However, the intrinsic properties of neural networks make them an interesting approach, despite some measure of ine ciency. This paper presents a clustering approach for image segmentation based on a modiÿed fuzzy approach for image segmentation (ART) model. The goal of the proposed approach is to ÿnd a simple model able to instance a prototype for each cluster avoiding complex post-processing phases. Results and comparisons with other similar models presented in the literature (like self-organizing maps and original fuzzy ART) are also discussed. Qualitative and quantitative evaluations conÿrm the validity of the approach proposed.

show abstract

“…We believe that interactions between agents provide an alternative way for image segmentation to that of approaches based on complicated and costly models. Extensive experiments have been performed using real images from the ABW database [5]. The obtained results show the high potential of the proposed approach for an efficient and accurate segmentation of range images.…”

Section: Introductionmentioning

confidence: 99%

“…Region classification is performed according to a compare tool tolerance T ; 50% < T ≤ 100%, which reflects the strictness of the classification. In our case, four methods, namely USF, WSU, UB and UE, cited in [5] are involved in the comparison.…”

Section: Experimentation and Analysismentioning

confidence: 99%

See 1 more Smart Citation

An Agent-Based Approach for Range Image Segmentation

Mazouzi

Guessoum

Michel

et al.

Massively Multi-Agent Technology

View full text Add to dashboard Cite

Abstract. This paper presents and evaluates a multi-agent approach for range image segmentation. A set of reactive and autonomous agents perform a collective segmentation by partitioning a range image in its different planar regions. The agents move over the image and perform cooperative and competitive actions on the pixels, allowing a robust region extraction, and an accurate edge detection. An artificial potential field, created around the pixels of interest, ensures the agent coordination. It allows the agents to concentrate their actions around the edges and the noise regions. The experimental results show the potential of the proposed approach for scene understanding in range images, regarding both segmentation efficiency, and detection accuracy.

show abstract

An experimental comparison of range image segmentation algorithms

Cited by 651 publications

References 32 publications

Construction object identification from LADAR scans:

Construction object identification from LADAR scans:

A clustering fuzzy approach for image segmentation

An Agent-Based Approach for Range Image Segmentation

Contact Info

Product

Resources

About