Guided, Fusion-Based, Large Depth-of-field 3D Imaging Using A ...

Trang chủ » Xiaowei Hu Tsinghua » Guided, Fusion-Based, Large Depth-of-field 3D Imaging Using A ...

Có thể bạn quan tâm

Full text links CiteDisplay options Display options Format AbstractPubMedPMID

Abstract

Three dimensional (3D) imaging technology has been widely used for many applications, such as human-computer interactions, making industrial measurements, and dealing with cultural relics. However, existing active methods often require both large apertures of projector and camera to maximize light throughput, resulting in a shallow working volume in which projector and camera are simultaneously in focus. In this paper, we propose a novel method to extend the working range of the structured light 3D imaging system based on the focal stack. Specifically in the case of large depth variation scenes, we first adopted the gray code method for local, 3D shape measurement with multiple focal distance settings. Then we extracted the texture map of each focus position into a focal stack to generate a global coarse depth map. Under the guidance of the global coarse depth map, the high-quality 3D shape measurement of the overall scene was obtained by local, 3D shape-measurement fusion. To validate the method, we developed a prototype system that can perform high-quality measurements in the depth range of 400 mm with a measurement error of 0.08%.

Keywords: 3D imaging; focal stack; guided fusion.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1

Figure 1

The framework of our proposed…

Figure 1

The framework of our proposed method. Image sequences of different focal distances are…

Figure 1 The framework of our proposed method. Image sequences of different focal distances are acquired by the multi-focus illumination-imaging system; then, local, fine 3D measurements and a global coarse depth map are generated by the gray-code method and focus stacking technique, respectively. These local, fine 3D measurements, along with the global coarse depth map, are finally fed into the guided fusion model to get the high-quality 3D measurement of the overall scene.
Figure 2

Figure 2

Optimal focus positions and texture…

Figure 2

Optimal focus positions and texture map extraction. ( a ): optical system; ( …

Figure 2 Optimal focus positions and texture map extraction. (a): optical system; (b): texture map. The optimal sequence of focus positions has no gaps or overlaps between depths of field (DOFs) from consecutive focal planes. The image projected by the finest stripes of the gray code sequence is extracted into the focal stack.
Figure 3

Figure 3

The guided fusion model. The…

Figure 3

The guided fusion model. The global coarse depth map contains the corresponding focus…

Figure 3 The guided fusion model. The global coarse depth map contains the corresponding focus position of each object. Therefore, we can extract the accurate region of each local fine 3D measurements under the guidance of the global coarse depth map to fuse into the global fine 3D measurement.
Figure 4

Figure 4

The prototype system and the…

Figure 4

The prototype system and the test scene. The system includes a MEMS mirror-enabled…

Figure 4 The prototype system and the test scene. The system includes a MEMS mirror-enabled laser projector, a CMOS camera equipped with a liquid lens, and a control platform for synchronization. Three different sizes of dolls are placed in front of the camera, from far to near. The specific scene is shown in the figure.
Figure 5

Figure 5

Comparison of measurement results. ( …

Figure 5

Comparison of measurement results. ( a ) The camera focal distance at f …

Figure 5 Comparison of measurement results. (a) The camera focal distance at f2=200 mm; (b) the camera focal distance at f4=400 mm; (c) the camera focal distance at f6=600 mm; (d) our method.
Figure 6

Figure 6

Comparison of reconstruction results. ( …

Figure 6

Comparison of reconstruction results. ( a ) f 2 = 200 mm; ( …

Figure 6 Comparison of reconstruction results. (a) f2 = 200 mm; (b) f4 = 400 mm; (c) f6 = 600 mm; (d) our method. The first row represents the results of the small sphere (diameter = 30 mm at 200 mm) under different camera settings and methods. The second row represents the corresponding results of the medium sphere(diameter = 50 mm at 400 mm), and the last row represents the corresponding results of the big sphere(diameter = 70 mm at 600 mm). The reconstruction errors are shown in the upper right corner of each figure, which are RMSEs and mean errors.
Figure 7

Figure 7

Photograph of the test scene…

Figure 7

Photograph of the test scene and the global coarse depth map. ( a …

Figure 7 Photograph of the test scene and the global coarse depth map. (a) Test scene; (b) depth map.
Figure 8

Figure 8

Comparison of measurement results. The…

Figure 8

Comparison of measurement results. The scene includes three objects of different shapes and…

Figure 8 Comparison of measurement results. The scene includes three objects of different shapes and materials. (a) The statue in-focus; (b) the box in-focus; (c) the doll is in-focus; (d): our method.
Figure 9

Figure 9

Comparison of measurement results. The…

Figure 9

Comparison of measurement results. The scene includes one object having a large depth…

Figure 9 Comparison of measurement results. The scene includes one object having a large depth range. (a) The global coarse depth map; (b) the left part in-focus; (c) the right part in-focus; (d) our method.
All figures (9) See this image and copyright information in PMC

References

    1. Liberadzki P., Adamczyk M., Witkowski M., Sitnik R. Structured-Light-Based System for Shape Measurement of the Human Body in Motion. Sensors. 2018;18:2827. doi: 10.3390/s18092827. - DOI - PMC - PubMed
    1. Song L., Li X., Yang Y.g., Zhu X., Guo Q., Liu H. Structured-light based 3D reconstruction system for cultural relic packaging. Sensors. 2018;18:2981. doi: 10.3390/s18092981. - DOI
    1. Zhang S. High-speed 3D shape measurement with structured light methods: A review. Opt. Lasers Eng. 2018;106:119–131. doi: 10.1016/j.optlaseng.2018.02.017. -

Từ khóa » Xiaowei Hu Tsinghua