Friday, March 24, 2006

Surveillance Robotics: Using colors to analyse

u0307999 ZHAI NING Among the possible applications which have been foreseen for Service Mobile Robots, surveillance robots (i.e., robots designed to replace human security guards in making rounds) are becoming more and more popular, as witnessed by the many systems commercially available and the growing interest in the research community. In this blog, I just concentrate one area: How is a surveillance robltic can detect unexpected changes in the environment? Well I found out one method from a paper written by Mattia from university of Genova, Italy, which detailed the use of the following mechanism: The robot “looks at” the environment through a TV camera; next, it compares what “it sees” in a specific moment with what “it should see” at that same location. In particular, an approach is proposed for images comparison which requires to find color clusters in the color histograms corresponding to the images to be compared: by analyzing the color clusters in the two images, the system detects similarities or differences between them and consequently deduces if something has changed in the scene. Well that seems like an computer vision problem, and for sure it is. According to another report written by Paolo in the same university, he propose a “ad hoc” algorithms have been implemented for color clusters comparison. The detail is covered in Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation July 1620, 2003, Kobe, Japan. I want to comment on the simply idea: it is only a simplification of the human unconscious way of dection. But it helps a lot for robotic. Can the robotic community do better by consulting more life sciences research result? Maybe it is a good way to start think about.


Assistive said...

u0204714 Chan Hongjiang

I believe emulating the human body as a method of implementation of robotics has been a hot research area for some time now. Indeed it is a very smart way, but the success rate doesn't seem high, mainly because of the complexity and sheer processing power of the human brain.

However, I am not so sure of the link between colour patch matching and image recognition. Perhaps the author of the blog can elaborate further. Also, I believe that the human eye can actually compensate for the loss of colour in images. Colour blind people can recognise images just as well and normally we can also identify black and white images just as easily.

Perhaps we still have a long long way to go before we even come close to making a robot that can emulate a human being.

eng10822 Loh Khai Choon said...

I feel that with the ever increasing processing powers in processors, emulating humans is only a matter of time, but do we really want such a thing to happen?

Right now, I guess, what the processor loses to the human brain is the AI side. Imagine robots who can think like humans. With much tougher frames, and the ability to control much more equipment, if one robot turned out to think like, say Adolf Hitler, it could really turn into a nightmare for humans. I'm sure those who've watched The Matrix will understand.

Implementation of robots to emulate the human body seems to be something that could be achieved in the near future, with the knowledge to robotics being much more advanced now. Robot soccer has shown that robots can co ordinate and work as a team, even though its aims are simple.

Perhaps, it would be better to limit the full abilities that robots could possibly do?