Portable System for Box Volume Measurement Based on Line-Structured Light Vision and Deep Learning

Area of Science:

• Robotics and Automation
• Computer Vision
• Metrology

Background:

• Accurate box volume measurement is crucial for the intelligent logistics industry.
• Existing systems may lack portability or require complex calibration.

Purpose of the Study:

• To develop a portable, accurate, and easily calibrated system for box volume measurement.
• To leverage line-structured light vision and deep learning for enhanced measurement capabilities.

Main Methods:

• A novel system integrating a 2x2 laser line grid projector, sensor, and software modules.
• An end-to-end deep learning model utilizing an improved holistically nested edge detection network for edge extraction.
• An automatic one-step calibration method for the line-structured light projector.

Main Results:

• The system achieves a measuring range of 100-1800 mm with errors under ±5.0 mm.
• Theoretical analysis shows measurement uncertainty ranging from ±0.52 mm to ±4.0 mm.
• The device is compact (140x35x35 mm) and lightweight (110 g).

Conclusions:

• The proposed system offers a viable solution for portable automatic box volume measurement.
• The integration of advanced vision and deep learning techniques ensures high accuracy and efficiency.
• The system's portability and ease of calibration make it suitable for diverse logistics environments.