Locking Object Position using Camshift Algorithm for Automatic Trolley

Abstract

Today, the development of science and technology produces many innovative devices, that can facilitate people to make their work easier. One of the devices is a trolley, which can help people to bring goods from one place to another place in a modern market. This paper proposed a method to develop an automatic trolley control, which can follow a customer based on the clothes color. The control system is built with Jetson Nano, where a webcam and motors are connected. The webcam has a function as the trolley's eye, which is used to view and retrieve the color data of the customer's clothes. These data are processed by Jetson Nano and the results are sent to the motors for the robot movement. The unique point of this paper is that the trolley has ability to lock the position of a certain customer. To improve the performance of locking object position, the Camshift Algorithm was used. This research conducted three experiments to lock the object position using proposed algorithm. The result of these experiments were: (1) minimum distance between main object and another object to be detected by the camera was 5 cm; (2) the main object was successfully detected by the camera with maximum distance of 360 cm; (3) the camera successfully locked the main object in front of the camera as far as 360 cm while another object stood in front of the camera as far as 20 cm. The advantage of this system is that the trolley can remember the customer that is supposed to be followed even though there are some customer use similar color. This study shows that the trolley can identify the position of customer and follows successfully.