Methodology for analysis of objects shape data, received from video sequence of the camera

Аuthors

Efimov A. I.^*, Il'in V. N.^**

Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia

*e-mail: alhimic23@yandex.ru
**e-mail: vnil2005@yandex.ru

Abstract

In this article we will discuss different ways of analysis of objects shape data, received from webcams. Purpose of this research was development of various areas of augmented reality and computer vision.

In the analysis of the video sequence, we can distinguish several steps:

• Pre-processing image. It include decreasing influence of noise effects and light artifacts on data recognition.

• Receiving the camera position. Program try to find special flat marker, placed on image. Depending on received data, we establish relative position of camera in current frame.

• Find stable points. We find set of cue points, using FAST-ER method. After frame data processing we can declare stable points of global coordinate system.

• Getting silhouette. We can resolve image depth and cue points of image drop. By processing of gradient between such points we can produce approximate silhouette of object on current frame.

• Systematization final result to get accurate data about object shape.

In the future we decide to develop halftone recognition to make object shape restore even more accurate. We can also create variation of described algorithms without using of special marker, but it required bunch of sophisticated changes in it.

Described algorithm can be used to get object’s 3D models for create it copy on 3D printer. Next important area for us is augumented reality, because of potential of depth map to resolve problems of real and virtual objects intersections and increased range of objects for tracking. Depth maps also has an ability to create stereographic image from set of frames of video.

Analysis of objects shapes is important way of development of computer vision and recognition.

Keywords:

computer vision, analysis 3D shape, augmented reality, object recognition

References

1. Andrew Blake, Michael Isard. Active Contours, Luxembourg, Springer, 2000, 306 p.

2. Simon J.D. Prince. Computer Vision: Models, Learning, and Inference, Cambridge, Cambridge University Press, 2012, 600 p.

3. Scott Krig. Computer Vision Metrics: Survey, Taxonomy, and Analysis, Luxembourg, Springer, 2014, 418 p.

4. Asim Bhatti. Current Advancements in Stereo Vision, Rijeka, Croatia, InTech, 2012, 548 p.

5. Aapo Hyvarinen, Jarmo Hurri, Patrik O. Hoyer. Natural Image, Statistics, Luxembourg, Springer, 2009, 448 p.

6. Xiong Zhihui. Computer Vision, Rijeka, Croatia, InTech, 2008, 548 p.

7. Robert Laganière. OpenCV 2 Computer Vision Application Programming Cookbook, Birmingham, Packt Publishing Ltd, 2011, 290 p.

8. Lunev E.M., Neretin E.S. Dyachenko S.A., Dubrovo A.I. Trudy MAI, 2016, no. 86, available at: http://trudymai.ru/eng/published.php?ID=66366

Download