After reading the assigned materials and checking in detail Real Life Applications and Art Installations sections in our slides, choose one example of Computer Vision. Research about that example and post on your blog your findings. Be specific about which tools you think were used and which challenges the developers faced. In your response, clarify why you chose it and how does it relate to your own experiences with Computer Vision. Due December 7th.
The example I chose and why I chose it
I chose self-driving cars both because I personally like it so much and because I believe it is a technology that will bring revolutionary change to the future. Imagine that there will be not drivers in the future; imagine there will be much fewer traffic accidents in the future… I think with the development of technology, self-driving will be perfectly achieved in the near future.
Which challenges the developers faced
I think the main challenge the developers are faced with is how to detect the cars and people on the road. In the reading, it is called Detecting presence. It is very hard to achieve considering how many objects may appear on the road: Pedestrians, cars, buses, bicycles, motorcycles, cats and dogs, etc. Pedestrians have different body shapes and cars have different colors. It is almost impossible for the computer vision to detect all these objects. A probable solution may be using a combination of sensors instead of using only one of them.
Another challenge is to detect and predict the movement of the objects on the road. The self-driving cars need to predict how and where the objects are moving so that it can take actions to avoid traffic accidents. In the reading, it is called Detecting motion. It is very hard to achieve, too because the movements of the objects are very irregular. It’s very hard even just to predict the movement of cars. How can we predict the movement of a cat on the road?
which tools you think were used
Various tools should be used. Camera is the best technology to detect different objects on the road. Radar is already a mature technology to be used on the transportations. The computer will of course be used to analyze the data colloected by sensors and make reactions. A tri-axis accelerometer can be used so that the car can tell not only its speed in the direction of forward and back, left and right, but it can also tell whether it is climbing or downhill, which is also very important to a driver. When it is climbing, the engine should work harder to avoid being crashed into by the cars behind, when it is downhill, the speed should be controlled. A light sensor like potentiometer should also be used. When it is dark, the car should know it and automatically turn on the light so that its camera can work well.