Improving Measurement Accuracy Using Image Super-Resolution

Abstract:

A super resolution measurement technique is proposed to improve the accuracy of the automated stereovision measurement systems. Image super resolution is useful to reconstruct a visually enhanced high resolution image from a set of low resolution images. Due to the ill conditioning problem of the super resolution model, a-priori information is augmented into the model. We examined different a-priori and concluded that the Solution norm is the most suitable apriori to be used with the optimization technique described. Experiment also showed that the super resolution technique could perform measurement on small images, which are not possible without the technique. An increase in measurement accuracy from 99.73% to 99.91% is obtained.

• Periodical: Key Engineering Materials (Volumes 295-296)
• Pages: 699-704
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.295-296.699
• Online since: October 2005
• Authors: D.S.H. Ling, H.Y. Hsu, G.C.I. Lin, S.H. Lee
• Keywords: A-Priori Regularization, Image Super-Resolution, Image-Based Measurement, Non-Contact Measurement, Stereovision

References

[1] D.S.H. Ling, G.C.I. Lin and S.H. Lee: Proc of International Symposium on Precision Mechanical Measurements, Anhui, Hefei, Vol. 4 (2002), p.241.

[2] N. Nguyen, P. Milanfar and G. Golub: IEEE Transactions on Image Processing, Vol. 10 (2001), p.573.

[3] D. Rajan and S. Chaudhuri: Journal of Mathematical Imaging and Vision, Vol. 16 (2002), p.5.

[4] R.R. Schultz and R.L. Stevenson: IEEE Transactions on Image Processing, Vol. 5 (1996).

[5] C.R. Vogel and M.E. Oman: IEEE Transactions on Image Processing, Vol. 7 (1998), p.813.

[6] D.S.H. Ling, G.C.I. Lin and S.H. Lee: Proc of International Pacific Conference on Manufacturing and Management, Thailand, 2002, p.769.