Active Dual Line-Laser Scanning for Depth Imaging of Piled Agricultural Commodities for Itemized Processing Lines

Area of Science:

• Robotics and Automation
• Computer Vision
• Sensor Technology

Background:

• Accurate depth imaging is crucial for automated selection and processing of piled food items.
• Existing depth imaging techniques like Time-of-Flight and stereoscopy lack resolution for small items.
• Structured light methods, while high-resolution, often require conveyor motion, limiting their use with static items.

Purpose of the Study:

• To introduce an active dual line-laser scanning system for high-resolution 3D imaging of static piled items.
• To overcome limitations of traditional depth imaging for itemized food processing.
• To enable robotic perception for handling individual items from piles.

Main Methods:

• Development of an active dual line-laser scanning system with a unique geometrical configuration.
• Utilizing laser redundancy to overcome occlusions and achieve a large field of view (FOV) from a long working distance.
• Scanning static piled items, such as crabs and mushrooms, without conveyor motion.

Main Results:

• Achieved a depth reconstruction Mean Squared Error (MSE) of 0.3 mm and a Standard Deviation (STD) of 0.5 mm.
• Successfully generated high-resolution 3D images of complex items like piled crabs and mushrooms.
• Demonstrated the system's ability to overcome occlusions and provide detailed depth maps.

Conclusions:

• The proposed dual line-laser scanning system provides accurate depth imaging for static piled products.
• This technology significantly benefits robotic perception in automated food processing, quality inspection, and handling.
• The system offers broad applicability for various industries requiring precise 3D perception of piled goods.