博士论文 - 用于先进表面检测的工业视觉系统

Job Description

Main activities

Fundamental Research on Acquisition Technology

• Investigate and compare 3D acquisition principles (stereophotometry, phase-shifting interferometry, vertical scanning interferometry, structured light, and infrared pattern projection) at a fundamental level, assessing their physical limits in terms of resolution, speed, and robustness for various surface characterizations.
• Conduct an in-depth study of illumination spectra (visible, NIR, UV, multispectral) and their interaction with various surface properties (reflectance, scattering, translucency of active material coatings) to determine optimal spectral configurations for maximizing defect contrast.
• Explore novel image acquisition techniques and optical configurations (multi-angle, multi-wavelength, polarimetric imaging) to capture surface and subsurface information beyond what conventional single-modality systems can achieve.

Development of Optimized Multimodal AI Algorithms for In-Line Inference

• Design and develop new multimodal deep learning architectures capable of jointly processing heterogeneous data streams (3D topography, 2D intensity, multispectral, and thermal) for 3D reconstruction and defect detection.
• Optimize these algorithms specifically for in-line inference: low-latency, high-throughput processing compatible with production line speeds, targeting edge deployment on GPU/FPGA compute platforms.
• Address battery-specific AI challenges: extreme class imbalance (rare defects), scarce labeled data (self-supervised, semi-supervised, and synthetic data generation strategies), and generalization across electrode chemistries and process variations.
• Benchmark fusion strategies (early, mid-level, late fusion) and quantify the detection gain brought by multimodality versus single-sensor approaches.

Multi-Sensor Data Fusion Architecture

• Design a real-time data fusion pipeline combining the developed 3D vision system with complementary sensors on the R&D line (line-scan cameras, IR thermography, thickness gauges, laser profilometers).

R&D Production Line Integration & Cell Quality Impact

• Collaborate with Process and Equipment teams to prototype and integrate the vision system on the R&D lines.
• Build the full acquisition-to-decision pipeline: hardware (cameras, illumination, compute units), software architecture, and data flow (acquisition ??? preprocessing ??? fusion ??? inference ??? MES feedback).

Hardware & Software Market Study

• Evaluate industrial cameras, illumination sources, compute platforms, and software frameworks.
• Provide buy-vs-build recommendations for each subsystem.

Requirements

• MSc or Engineering degree in computer vision, optics, image processing, or a related field.
• Strong programming skills (Python, C/C++) and experience with deep learning frameworks.
• Knowledge of optics, 3D metrology, or acquisition systems is a plus
• Fluent English required, French appreciated

We may use artificial intelligence (AI) tools to support parts of the hiring process, such as reviewing applications, analyzing resumes, or assessing responses. These tools assist our recruitment team but do not replace human judgment. Final hiring decisions are ultimately made by humans. If you would like more information about how your data is processed, please contact us.