Download presentation
Presentation is loading. Please wait.
Published byCorey Poole Modified over 9 years ago
1
Autonomous Vehicle Pursuit of Target Through Optical Recognition Vision & Image Science Laboratory, Department of Electrical Engineering,Technion Daniel Vinson, Sergei Djuvinov Supervised by: Johanan Erez, Ronel Veksler
2
Goal of the Project To control an autonomous vehicle through recognition of a moving target
3
Image Processing A computer can ’ t simply understand a picture the way a human can. Numbers must be used to help the computer isolate shapes and colors such as the red sheet of paper
4
Image Processing RGB One of the easiest ways to process an image is through the colors that they are made of. Red, green and blue. The pictures are stored on the computer as an assortment of R,G and B values from 0 (none) to 255(full). In these RGB filters of the previous image, the lighter the shade is, the more of that color is present at that spot.
5
Image Processing HSV
6
HSV stands for Hue, Saturation, and Value It is a powerful and different way to analyze pictures The Hue is associated with the color of the picture The Saturation is associated with how intense the color is The value is associated with how bright the color is
7
Image Processing HSV Hue The Hue of an image relates to the color of that image. By changing the hue of the blue sign, the blue color changes through different colors
8
Image Processing HSV Saturation The Saturation of an image relates to the intensity of the color. By changing the saturation of the blue sign from the highest (1) to the lowest (0), the blue color changes from very intense to grey.
9
Image Processing HSV Value The Value of an image relates to the brightness of that image. By decreasing the value of the blue sign from 1 to 0, the blue color gets darker, but so does the white sign and background
10
Image Processing HSV The picture on the right is a HSV representation of the picture seen on the left The Hue is shown as red The Saturation is shown as green The Value is shown as blue While this is a difficult way to look at and understand an image for a person, in a numerical sense, it allows you to compare and contrast things that you wouldn’t be able to in RGB color
11
Image Processing HSV HUE SATURATION VALUE Just as any RGB picture can be broken down into the Red, Green, and Blue components, a HSV picture can be split into the components. In these grayscale images, the lighter the color is, the greater the H/S/V value is at that point.
12
Image Processing Use Image processing is critical in order for the computer to be able to locate a selected item in a picture. Using an algorithm to find the target, the computer can isolate important pixels, in this case the red bull ’ s-eye by only coloring pixels with a hue less than.04 and a saturation greater than.5
13
General Program Layout Initialize everything Take one frame out of the streaming video Look for the target in the picture Find where the target is in the picture Turn or go straight accordingly Find about how close the target is If it is close enough, stop
14
Processing Pseudo-code Capture an image Using every 3 rd pixel, find if the saturation is >.5 and the hue is <.04 If it is, remember it, otherwise discard the numbers Find center of target Move, turn or stop according to location and distance
15
The Vehicle ’ s Components
16
Software Used VideoOCX®, A commercial software library was used to capture images onto the computer and for the robot Phantom is a set of functions for the Visual Studio 6.0 that helps us control the LEGO™ vehicle. Microsoft Visual Studios 6.0 ™ Visual C++ 6.0
17
Final Result
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.