Binocular Model: A deep learning solution for online melt pool temperature analysis using dual-wavelength Imaging Pyrometry

TL;DR

The paper tackles real-time melt pool temperature analysis in Laser Powder Bed Fusion by introducing a Binocular model that leverages dual-wavelength Imaging Pyrometry. The architecture employs two input channels, an Interaction Branch, and an Inception-based Wide Deep Module to deliver pixel-wise MP temperature maps end-to-end, bypassing labor-intensive preprocessing. It achieves a high fidelity with $R^2$ around $0.95$ and processes frames at roughly $650$–$800$ fps, offering up to ~1000x speed improvements over the baseline that relied on KAZE-based alignment. This work enables scalable, batch-capable temperature monitoring for improved process control and part quality in metal AM, while identifying data-transfer bottlenecks as a practical constraint for further speed gains.

Abstract

In metal Additive Manufacturing (AM), monitoring the temperature of the Melt Pool (MP) is crucial for ensuring part quality, process stability, defect prevention, and overall process optimization. Traditional methods, are slow to converge and require extensive manual effort to translate data into actionable insights, rendering them impractical for real-time monitoring and control. To address this challenge, we propose an Artificial Intelligence (AI)-based solution aimed at reducing manual data processing reliance and improving the efficiency of transitioning from data to insight. In our study, we utilize a dataset comprising dual-wavelength real-time process monitoring data and corresponding temperature maps. We introduce a deep learning model called the "Binocular model," which exploits dual input observations to perform a precise analysis of MP temperature in Laser Powder Bed Fusion (L-PBF). Through advanced deep learning techniques, we seamlessly convert raw data into temperature maps, significantly streamlining the process and enabling batch processing at a rate of up to 750 frames per second, approximately 1000 times faster than conventional methods. Our Binocular model achieves high accuracy in temperature estimation, evidenced by a 0.95 R-squared score, while simultaneously enhancing processing efficiency by a factor of $\sim1000x$ times. This model directly addresses the challenge of real-time MP temperature monitoring and offers insights into the encountered constraints and the benefits of our Deep Learning-based approach. By combining efficiency and precision, our work contributes to the advancement of temperature monitoring in L-PBF, thus driving progress in the field of metal AM.

Figures (9)

• Figure 1: The experimental setup used for single-camera two-wavelength imaging pyrometry (STWIP). Base_line_paper
• Figure 2: In this experiment flowchart, the process initiates with specimen design and fabrication plan generation on the left. Proceeding onward, fabrication and data acquisition phases ensue, culminating in comprehensive data processing. Finally, the acquired data undergoes two analyses, one using the baseline and one using AI, leading to precise estimations of the final MP characteristics.
• Figure 3: The illustration displays two images side by side: the original MP image at 620nm on the left and the transformed image at 550nm on the right. ck_1
• Figure 4: Binocular model structure. On the left, two-channel input images are the inputs transitioned to their corresponding MP representations on the right
• Figure 5: Wide Deep module structure, branching the data into four flows and aggregating them back for a diverse feature development