Vending Machine Based on Optical Flow Method
Download as PDF
DOI: 10.23977/amce.2019.002
Author(s)
Xingxing Li, Chao Duan, Yan Zhi, Panpan Yin
Corresponding Author
Xingxing Li
ABSTRACT
Moving target tracking plays a vital role in vending machines. This paper proposes a method of applying the optical flow method to a vending machine, which can detect and extract the corner features of moving objects and identify objects that track motion. Firstly, the image features in the video stream are extracted by Shi-Tomasi corner detection, and then the detected corner features are input to the Kanade-Lucas-Tomasi (KLT) algorithm to estimate the positional relationship of the feature points in the adjacent two frames. Finally, the method can determine the moving trajectory of the object according to the positional relationship, thereby determining whether there is a moving object in the vending machine, and providing the information to the upper computer for automatic judgment.
KEYWORDS
Moving target tracking, Shi-Tomasi corner detection, Kanade-Lucas-Tomasi (KLT), Vending machine