AI-Driven 3D Battery Inspection Software Delivered in 6 Months

Summary

INNERLUXES delivered an AI-powered solution that performs high-precision 3D inspection of batteries for a UK leader in manufacturing quality control. In about six seconds, the system turns X-ray scans into volumetric models and detects even the slightest structural deviations in batteries.

About the Client

The Client is a UK-based leader in test and measurement solutions for manufacturing, with a global network of sales and service branches and a long-standing reputation in precision testing. Its equipment is trusted worldwide for rigorous compliance verification and meets the highest standards of measurement and calibration, with operations following lean-manufacturing and continuous-improvement principles and international quality standards.

To strengthen its position at the forefront of innovation, the Client wanted to enhance its inspection hardware with an intelligent software layer to automate battery quality control. The envisioned solution had to process raw X-ray scans captured by the Client's robotic inspection complex, reconstruct high-fidelity 3D models of each battery, and identify cathode-anode misalignment, overhang issues, and other internal defects. The Client sought a technology partner with deep expertise in 3D imaging, computer vision, and industrial quality control, and chose INNERLUXES.

Engineering a High-Precision, Real-Time Battery Inspection System

INNERLUXES's team of four software engineers and a project manager planned and executed a six-month project to build a fully automated image-analysis system and integrate it into the Client's production equipment.

Speeding up CT image analysis with an added 2D pipeline. The project began with collecting sample data and reviewing documentation to fully understand the production equipment and inspection objectives. INNERLUXES's engineers then conducted in-depth R&D on CT scan data, testing reconstruction approaches and defining optimal processing parameters to maximize accuracy, stability, and speed.

Overcoming battery X-ray image-quality challenges. Early R&D revealed uneven reconstruction quality — a typical issue in radiographic imaging of metallic or composite components like battery cells, where variations in material density, surface reflectivity, and cell positioning affect X-ray projection intensity and cause local artifacts and inconsistent voxel brightness. To stabilize accuracy, the engineers fine-tuned key reconstruction parameters and introduced adaptive noise-reduction filters that adjust to local image contrast and density, reducing beam-hardening and scattering effects and producing consistently clear internal structures across cell types. As a result, quality engineers can detect even subtle internal defects with confidence, reducing false rejects and costly rework or manual re-inspection.

Developing the 3D quality-control software end to end. Once the analytical core was validated, INNERLUXES engineered a complete, production-ready solution with these key modules:

• Real-time data ingestion and preprocessing — connects directly to PLC-controlled scanning equipment (line-scan cameras, CT scanners), orchestrates data capture, validates scan completeness and metadata integrity, and triggers analysis.
• 2D linear analysis — uses CNN-based computer vision to extract control points from 2D X-ray scans, runs quick dimensional checks, and verifies cross-section geometry, symmetry, and positioning against tolerances.
• 3D reconstruction and battery-position localization — combines multiple X-ray slices into a full volumetric model, determines the battery's true position and orientation via its XY and XZ reference planes, and — once the battery is virtually straightened — extracts standardized, correctly oriented 2D cut-plane frames that consistently reveal the anode-cathode layers regardless of original placement.
• Anode-cathode segmentation, tracking, and overhang calculation — a neural network distinguishes anode and cathode textures, density patterns, and shapes to produce clean binary masks, then computes control points, anode overhang, inter-anode distances, and layer spacing; a Vision Transformer (ViT) model compares all structures to reference patterns and tolerance rules to detect misalignment, excessive overhang, spacing irregularities, and other hidden defects.
• Data export and reporting — structures and sends results to visualization interfaces and the client's MES or production-control database for recordkeeping and downstream analytics.
• System monitoring and diagnostics — track processing logs, hardware utilization, and module performance (scan throughput, error rates, latency) for stable, high-speed operation.
• Quality-manager interfaces — display key inspection metrics such as pass/fail rates, average scan duration, and defect distribution.

On-site validation and final tuning. INNERLUXES's engineers worked closely with the Client's technical team during on-site validation, with joint debugging and calibration sessions ensuring the software ran flawlessly with the scanning equipment and production data under real load.

Key Outcomes for the Client

• Expanded, enhanced product portfolio — an automated, highly accurate battery quality-control solution; thanks to hybrid 2D-3D analysis and optimized workflows, the full analysis pipeline runs in just six seconds.
• Rapid value realization — from concept to production in just six months.
• Stronger market position — integrating AI and 3D technology into its equipment further cemented the Client's reputation as a pioneer in precision testing and quality-control systems.
• Future-ready scalability — a modular architecture and flexible tech stack let the Client easily extend the solution to new product lines and inspection scenarios.

Technologies and Tools

Python, OpenCV, C++, .NET, PLC integration.