A Multi-View Stereo Measurement System Based on a Laser Scanner for Fine Workpieces

Area of Science:

• Metrology
• Computer Vision
• Mechanical Engineering

Background:

• Accurate 3D reverse modeling of complex surfaces is crucial for fine workpieces.
• Existing methods face challenges with registration errors, measurement inaccuracies, and deformations.
• Fast and precise machine vision measurement techniques are essential.

Purpose of the Study:

• To develop a convenient, economic, and accurate multi-view stereo (MVS) measurement system.
• To achieve complete and precise reconstruction of measured object surfaces.
• To address the limitations of current 3D reverse modeling approaches.

Main Methods:

• A multi-view stereo (MVS) system was constructed using a linear stage and a rotary stage.
• A laser line-scanning sensor was employed for data acquisition.
• The linear stage synchronized sensor scanning, while the rotary stage enabled multi-view data capture.
• Point cloud data from multiple views were registered and integrated into a 3D model.

Main Results:

• The system achieved a measurement accuracy of 0.075 mm.
• A complete 360° reconstruction was accomplished in just 34 seconds.
• Evaluation experiments confirmed the technique's validity and practicality.

Conclusions:

• The proposed MVS system offers a viable solution for high-quality 3D reverse modeling of complex surfaces.
• The integration of linear and rotary stages with laser scanning provides an efficient and accurate measurement approach.
• The technique demonstrates significant potential for applications in machine vision and precision engineering.