Skip to main content
Have a personal or library account? Click to login
Low cost structured-light based 3D surface reconstruction Cover
Open Access
|Apr 2019

Figures & Tables

Figure 1:

Examples of surface classes: (A) smooth without slope, (B) smooth with some slope, and (C) with both slope and roughness/texture.

Table 1.

Comparison of 3D surface reconstruction techniques.

Time of flightStereo visionOptical fiberStructure light
Working distanceLongMediumShortShort–medium
Vision fieldMediumLenses dependentNarrowMedium
CostHighMediumMediumLow
PowerLowLowLowLow
AccuracyMediumMediumHighMedium
SpeedFastFastSlowMedium
Figure 2:

The principle of data acquisition; working with laser-projected images (left side) to extract a surface profile along x-axis (right side).

Figure 3:

Basic image processing workflow for the proposed system; difference between test and reference images plus binarization (left side), where the extracted laser provides the surface 1D profile.

Figure 4:

The experimental setup of our surface level measurement system.

Figure 5:

2D scanning system with a linear actuator; 2D map is obtained from attaching 1D profiles, where the laser line is highlighted in a dashed rectangle.

Figure 6:

The experimental results for a simple surface (woody wedge), including (A) test image with projected laser line, and (B) obtained levelness H(x) from the wedge sample.

Figure 7:

The experimental results for a complex surface (robot toy), including test image with projected laser line (first row), and obtained levelness H(x) from the sample (second row). Working distance is 25 cm, and the sample is placed horizontally and vertically.

Table 2.

Evaluated cases for complex surface measurement with horizontal layout.

Horizontal layout
Dist.Images1D Profile H(x)MSE
19 cmgraphic/j_ijssis-2019-002_unfig_001.jpggraphic/j_ijssis-2019-002_unfig_002.jpg0.427
22 cmgraphic/j_ijssis-2019-002_unfig_003.jpggraphic/j_ijssis-2019-002_unfig_004.jpg0.457
25 cmgraphic/j_ijssis-2019-002_unfig_005.jpggraphic/j_ijssis-2019-002_unfig_006.jpg0.366
28 cmgraphic/j_ijssis-2019-002_unfig_007.jpggraphic/j_ijssis-2019-002_unfig_008.jpg0.375
31 cmgraphic/j_ijssis-2019-002_unfig_009.jpggraphic/j_ijssis-2019-002_unfig_010.jpg0.014
Mean square error (MSE)0.328
Table 3.

Evaluated cases for complex surface measurement with vertical layout.

Vertical layout
Dist.Images1D Profile H(x)MSE
19 cmgraphic/j_ijssis-2019-002_unfig_011.jpggraphic/j_ijssis-2019-002_unfig_012.jpg0.366
22 cmgraphic/j_ijssis-2019-002_unfig_013.jpggraphic/j_ijssis-2019-002_unfig_014.jpg0.209
25 cmgraphic/j_ijssis-2019-002_unfig_015.jpggraphic/j_ijssis-2019-002_unfig_016.jpg0.220
28 cmgraphic/j_ijssis-2019-002_unfig_017.jpggraphic/j_ijssis-2019-002_unfig_018.jpg0.073
31 cmgraphic/j_ijssis-2019-002_unfig_019.jpggraphic/j_ijssis-2019-002_unfig_020.jpg0.161
Mean square error (MSE)0.206
Figure 8:

Real image (top) and obtained 2D levelness maps (bottom) from the surface of two samples (cap A and cap B). The levelness H(x,y) is measured by the color bar in mm.

Language: English
Page range: 1 - 11
Submitted on: Jan 28, 2019
Published on: Apr 23, 2019
Published by: International Journal on Smart Sensing and Intelligent Systems
In partnership with: Paradigm Publishing Services
Publication frequency: 1 issue per year

© 2019 Yijun Yan, Maher Assaad, Jaime Zabalza, Jinchang Ren, Huimin Zhao, published by International Journal on Smart Sensing and Intelligent Systems
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.