Multi-exposure fusion for welding region based on multi-scale transform and hybrid weight
- 96 Downloads
The multi-exposure fusion is an effective image enhancement technique for high dynamic range (HDR) scene. In this paper, a novel multi-scale hybrid weight fusion framework is proposed to overcome the inherent defects of detail loss during the reconstruction process. Firstly, a novel hybrid weight method is developed by employing the local weight of a single image, the global weight between different exposure images, and the saliency weight from spectral residual model. Secondly, a new multi-scale hybrid weight image fusion algorithm based on Laplacian pyramid is proposed by applying the hybrid weight at each scale. The advantages of the proposed fusion algorithm over individual weight are analyzed from a theoretical point of view and then experimentally verified with multi-exposure image in welding region. Furthermore, the guided filter is utilized to smooth the reconstruction image, Laplacian pyramid image, and saliency weight maps for all the low dynamic range (LDR) images, which can effectively keep the edge information and reduce artifacts of weld seam region. Finally, by comparing our results comprehensively with other methods subjectively and objectively, the proposed fusion framework is verified that it can obtain better performance.
KeywordsMulti-exposure fusion High dynamic range Welding image Visual saliency Guided filter Hybrid weight
Unable to display preview. Download preview PDF.
This work was supported in part by National Natural Science Foundation (NNSF) of China under Grant 61873315, 61403119, Natual Science Foundation of Hebei Province under Grant F201402166, F2018202078, Special Correspondent Technology Plan of Tianjin under Grant 15JCTPJC55500, Science and Technology Project of Hebei Province under Grant 17211804D, and Talent Support Project in Hebei Province.
(AVI 24637 kb)
- 3.Mann S, Lo R C H, Ovtcharov K, Gu S, Dai D, Ngan C, Ai T (2012) Realtime HDR (high dynamic range) video for eyetap wearable computers, FPGA-based seeing aids, and glasseyes (eyetaps). In: Electrical & computer engineering (CCECE), 2012 25th IEEE Canadian Conference on (pp. 1–6)Google Scholar
- 4.Nayak, N. R., & Ray, A. (2013) Intelligent seam tracking for robotic welding. Springer-Verlag press, 8-30Google Scholar
- 7.Guido H, Kuhlenkoetter B (2014) A stateful robotic weldment geometry measuring system. Int J Mater Prod Technol 48(1–4):167–178Google Scholar
- 9.Lahdenoja O, Säntti T, Laiho M, Paasio A, Poikonen J K (2015) Seam tracking with adaptive image capture for fine-tuning of a high power laser welding process. In Seventh international conference on machine vision (ICMV 2014) (Vol. 9445), p. 94451VGoogle Scholar
- 20.Wan g, Z. (2015). An imaging and measurement system for robust reconstruction of weld pool during arc welding. IEEE Trans Ind Electron, 62(8), 5109-5118Google Scholar
- 24.Mertens T, Kautz J, Van Reeth F (2009) Exposure fusion: a simple and practical alternative to high dynamic range photography, computer graphics forum. Blackwell Publishing Ltd, 28(1): 161–171Google Scholar
- 34.Fattal R, Agrawala M, Rusinkiewicz S (2007) Multiscale shape and detail enhancement from multi-light image collections. ACM Trans. Graph. 26(3), Article 51Google Scholar
- 35.Hou X, Zhang L (2007) Saliency detection: a spectral residual approach. Computer vision and pattern recognition, 1–8Google Scholar
- 36.Debevec, P. E., & Malik, J. (1997) Recovering high dynamic range radiance maps from photographs. Proceedings of the 24th annual conference on Computer graphics and interactive techniques, 31:369–378Google Scholar
- 40.Li Y, Liu G (2009) Digital images clarity quality evaluation using non-subsampled contourlet transform. International Symposium on Computational Intelligence and Design. IEEE, 318–321Google Scholar