الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 79 |  |  |
| | | from | 79 | | |
Abstract
Visual object tracking is one of the most active research topics nowadays, we can see its applications such as face detection and smile detection in cameras and camera phones, self-driven cars, pick and DROP robots, and a lot more.
The basic concepts of object tracking are recognizing and identifying the object then track that object automatically, for years researches have been conducted to achieve real time object tracking in computer vision and artificial intelligence.
The main problem of real time object tracking is the need to perform all object tracking process simultaneously with the object movement, this problem required that real time object tracking system run the process faster than the object movement in real time.
This work describes the design and implementation of an object tracking system based on the color and shape of the object. first the camera output transformed from RGB to HSV, the object subtracted from the background using the object color and the noise and color plops removed by erosion. For identifying the shape of the object, the number of the points of its contour is counted and connected to form a boundary line. After that, the output used to control a mobile robot.
We have built a mobile robot to track an object optically. Our mobile robot has a camera to detect the needed object. We built a controller algorithm to detect the object by color and shape using the color’s HSV thresholds and object’s approximated contour, after detecting the object, the output from the controller which is the object’s trajectory in the frame is used to drive the robot using an Arduino board.
We have built our data sets to test the robot. our data sets contain different shapes with different colors in different situations. The results of testing our robot show that it can track single or multi objects successfully. Our proposed system is simple, fast, and tracks the objects in real-time.