From Manual to Intelligent: Redefining BIW Quality with Computer Vision

ALTEN was contacted by a premier global automotive manufacturer operating a state-of-the-art production facility in India. The facility specializes in Body-in-White (BIW) operations, focusing on manual gluing and welding processes integral to vehicle structural integrity.

The client sought to enhance the quality and consistency of manual gluing and welding tasks within their BIW assembly operations. The objective was to implement a real-time, camera-based inspection system capable of:

Validating the completeness of manual operations.
Detecting missing or inconsistent application patches.
Providing immediate feedback to operators to rectify issues promptly.

Challenge:

Clients were having challenges with Body-in-White automation as BIW, is a critical stage in automotive manufacturing where the vehicle’s sheet metal components are welded together to form its structural shell. At this point, the body has no paint, no trim, and no mechanical or electrical systems — it’s the raw framework that gives the vehicle its shape, strength and dimensional foundation.

The BIW structure includes key elements: the floor pan, side panels, roof, pillars and rear body. These components are welded with tight tolerances to form a rigid, self-supporting shell. This shell determines the vehicle’s crash performance, torsional stiffness and dimensional accuracy all critical to vehicle performance and manufacturing efficiency.

Welding is the backbone of BIW. Hundreds of joints must be created with high accuracy to ensure structural strength and consistency across every unit. These welds directly affect:

Dimensional accuracy (so parts align perfectly in later stages)
Torsional rigidity (crucial for ride, handling and NVH performance)
Crashworthiness (how well the car absorbs impact energy)

Even a slight misalignment or weak weld can affect the whole vehicle manufacturing line, leading to assembly issues, more rework and long-term durability issues.

Solutions

ALTEN developed and deployed an advanced computer vision system integrated with AI/ML algorithms to Enabling Smart Maintenance: An IoT & AI-Driven Transformation for Commercial Equipment Operations. Key components included:

Powerful Compute Platform: An NVIDIA Jetson TX2 based hardware compute platform customized to specific needs of the controller was provided with accessories such as remote controls, interface to audio-visual alarms etc.
High-Resolution Imaging: Installation of cameras to capture detailed visuals of manual operations in real-time.
AI-Based Analysis: Implementation of algorithms to compare actual application patterns against predefined standards, identifying deviations and incomplete work.
Interactive HMI: Development of a user-friendly Human-Machine Interface displaying:
- Live video feeds of ongoing operations.
- 3D models superimposed with tracking data to highlight areas of concern.
Feedback Mechanism: Real-time alerts and visual cues provided to operators, enabling immediate corrective actions.

Benefits/Knowledge where it counts

ALTEN’s teams with its extensive experience in Mechanical- Automotive Manufacturing : BIW designers and engineers work in the OEM automotive industry , tier-one suppliers and technical consulting support to companies that help automakers with engineering design services, Product Lifecycle Management, and plant automation, added values delivered to the clients as mentioned below:

Enhanced Quality Control: Automated detection of inconsistencies ensured higher adherence to quality standards.
Operational Efficiency: Immediate feedback reduced rework and minimized production delays.
Data-Driven Insights: Collection and analysis of operational data facilitated continuous improvement in manual processes.
Scalability: The modular design allowed for easy adaptation and deployment across multiple production lines and facilities.

The Toolbox:

This solution comprises a blend of standardized benchmarks are crucial for majority of computer vision applications incorporating AI Features.

NVIDIA Jetson TX2
Python
HTML5
OpenCV