Image decomposition model BL-Hilbert-L2 for dynamic thermal measurements of the printed circuit board with a chip by ESPI

• A new image decomposition model is proposed for ESPI fringe pattern filtering.
• It smoothes low-density fringes and preserves high-density fringes efficiently.
• It is better than that of the previous H1-G-E model.
• It outperforms even the WFF and the MCED method.
• The dynamic thermal measurements of the PCB and the chip are realized.

This paper is concerned with dynamic thermal measurements of a printed circuit board (PCB) with a chip under normal working conditions by electronic speckle pattern interferometry (ESPI). The deformation in the chip is much greater than that in the rest of the board, so the fringe density of the resulting ESPI fringe patterns varies greatly. Denoising of this type of ESPI fringe patterns is particularly challenging. We propose a new image decomposition model BL-Hilbert-L2 for this problem. In our model, low density fringes, high density fringes and noise are described respectively by appropriate norms and processed individually, so our model is able to sufficiently smooth low-density fringes as well as perfectly preserve high-density fringes. The experimental results have demonstrated that our model is better than that of the previous H1-G-E model, and outperforms even the windowed Fourier filtering and the modified coherence-enhancing diffusion method. In addition, based on our BL-Hilbert-L2 model, we employ fringe skeleton method to successfully realize the dynamic thermal measurements of PCB and the chip under normal working conditions.