Details

Autor(en) / Beteiligte

• Qu, Yufeng
• Zong, Guanghua

Titel

Automated de-stacking using compact and low-cost robotized system

Ist Teil von

• Industrial robot, 2014-01, Vol.41 (2), p.176-189

Ort / Verlag

Bedford: Emerald Group Publishing Limited

Erscheinungsjahr

2014

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Purpose – This paper aims to introduce a compact and low-cost robotized system and corresponding processing method for automatically identifying and de-stacking circulation boxes under natural stacking status. Design/methodology/approach – The whole system is composed of an industrial robot, a laser scanner and a computer. Automated de-stacking requires comprehensive and accurate status information of each box. To achieve this goal, the robot carries the laser scanner to perform linear scanning to describe a full depth image for the whole working area. Gaussian filter is applied to the image histogram to suppress the undesired noise. Draining and flooding process derived from classic algorithm identifies each box region from an intensity image. After parameters calculation and calibration, the grasping strategy is estimated and transferred to the robot to finish the de-stacking task. Findings – Currently, without pre-defined stack status, there is still manual operated alignment in stacking process in order to enable automatic de-stacking using robot. Complicated multi-sensor system such as video cameras can recognize the stack status but also brings high-cost and poor adaptability. It is meaningful to research on the efficient and low-cost measurement system as well as corresponding common data processing method. Research limitations/implications – This research presents an efficient solution to automated de-stacking task and only tests for three columns stack depending on the actual working condition. It still needs to be developed and tested for more situations. Originality/value – Utilizing only single laser scanner to measure box status instead of multi-sensor is novel and identification method in research can be suitable for different box types and sizes.