Automatic shopping cart by using image processing

Abstract

This research presents an application of color detection in image processing to make a shopping cart following the user such as elderly or who do not have the ability to use the shopping cart as normal people can use a shopping cart alone in the mall. The operation of automatic pushcart includes the following major sections. The camera gets the image to track the color-target plate and send digital data to each control unit which controls each motor by microcontroller to track the patch at the back of user. The control units make both wheels of the pushcart rotating in different directions. From experiments, pushcart can move in the desired direction that is forward, backward, turn left, or turn right respectively. Plate tracking uses color-plate tracking that learns color from the plate on user. We use non-stationary Gaussian model for keeping plate color with some image noise including environmental lighting changing. The results showed that the color-plate tracking using non-stationary Gaussian model can be used to track color plate but we might get error tracking in some cases such as rapid changing of environmental lighting. The rapid lighting changing causes of shifting from the model that can be solved by using an optimization technique for the future work.