Projector Light Blinding Prevention

The accelerated technological development of the last decades has resulted in widespread use of the projector in lectures at conferences, higher education institutions and schools (smart classrooms)


The accelerated technological development of the last decades has resulted in widespread use of the projector in lectures at conferences, higher education institutions and schools (smart classrooms).
Any person who gave a lecture or presentation noticed the glare of the projector when he stood next to or in front of the screen.
Glare may disturb the concentration of the lecturer, prevent him from seeing the audience, the audience may also lose concentration by trying to understand what is projected on the speaker, and so on.
Therefore, the solution to this problem can improve the use of the projector for both lecturer and audience. Moreover, as the projector is a tool that is used so commonly, a development that will solve this problem may also have financial incentives.
As previously described, the projector is a tool that is commonly used so it can be in different rooms or halls with different environments.
These conditions include:
a) variable lighting – artificial or natural lighting from different directions and at different intensity, and also lighting changes resulting from the projector.
b) position of the projector – the location of the projector may alter the shape of the projected image (but nowadays most projectors can deal with this problem).
c) The quality of the projector – a good projector displays color that is similar to the original image, while a low-level projector will distort the colors.
d) The human figure standing before the screen – color of skin hair and clothes, hair length and clothing.

The main objective of the project was to design such a system, that will answer the problem explained above. The solution to the problem must be robust in order to handle any situation in which the system can be found in.
This system will be complex since it must handle various conditions.

An additional main goal was efficiently implementing the system in order to create a software that would be simple to use would be able to do real-time calculations.

Figure 1: the problem of finding the human figure

The solution
The proposed solution was to find the human figure who stands before the screen. After finding the figure, the system will project a black mask where the character is, thus preventing the action of the projector in that area. To project this mask a merge between the mask and the original input image is made thus creating a new image that excludes the area where the character is.

The system includes connecting a webcam on top of projector. The camera output is transfered to the computer (that sends information to the projector) and sends it real-time video signal. When the information is received from the camera, the computer theoretically has all necessary information for finding the character. That is because at any given moment both projected and captured images exist in the computer.
The difference between the two images is the image location. However as earlier explained, improvements to the image obtained from the camera must be conducted because it can be very different from the source image.



Figure 2: Demonstration of the desired flow


Figure 3: Block diagram of the system

The tools used in the projects are:

  • Projector
  • Webcam
  • PC

The character recognition was the main problem of this project. The project mostly included studying and implementing of various image comparison algorithms, image correction, background detection etc.
The main problem was very complicated and it exceeded expectations. The project required a considerable amount of work and intensive study in image processing. The problems that required most of the work were dealing with mismatch between the captured and projected images (intensity and color differences).
During the project, many methods have been studied to solve the problem. Each method has introduced a different approach to resolve the problem. Some methods have yielded good results, but not good enough.
The project definition included precise identification of the figure, but it was very difficult to acheive. A less strict requirement would simplify the process and might have resulted in a complete product.

Future steps:

  • —Continue time-variations research
  • —Active contouring
  • —Using 3D-camera (e.g. Kinect)

We want to express our deep gratitude to our supervisor Yuval Bahat for his guidance throughout this whole year.