Designing and constructing an efficient targeting system using image processing method to detect and track object
This thesis is purposed to construct a prototype of an effective targeting system based on image processing method to detect and track a desired moving object which captured by camera device based on shape and color limitation. rnThe problem to be solved is how to create a high accuracy system to track an object as ideal as possible since there is always error exist in a system to be considered. This targeting system needs a mechanical design that could support pan and tilt movement. One image sensor is attached to identify object location and it is connected to mechanical system through image processing program. rnSome conversion formulas between digital image dimension into actual distance in orderrnto increase accuracy level is needed. Object is then pointed by pointing device. By using OpenCV library which has special function in maintaining computer vision, Visual C++ software able to build image processing algorithm tested on a circular red object with diameter 7 cm at distance of 1.5 m. Integrated with PID algorithm, the object tracking test using 5 location samples indicated the maximum error of 1.5 cm from object center point.
B00876 | (wh) | Available |
No other version available