Hobbies And Interests
Things You'll Need
Instructions
Collect and focus the light on the input to the vision processing mechanism. This part is very similar in animals and robots, although it may have differences. For example, a robot might be more interested in infrared light or may amplify light to the level that it can see better at night. The human eye focuses light on the retina, and the robot focuses light on an array of light sensing elements, but this first part of the process is remarkably similar.
Collect the light patterns that fall on the Charge-Couple Device (CCD). This is a semiconductor device that lets spots of light build up a charge on spots on the semiconductor. These charges are then shifted off the CCD chip, and into a data stream of impulses for further processing. As in the human eye, this data stream is processed to detect some basic patterns before it is sent to the robot "brain" for more processing.
Start the preliminary processing. This includes edge detection, motion detection and dividing up the scene into objects. It is important to know which objects are near, which are far away and which objects are moving. All of this information is found by the pre-processor. Use an Artificial Neural Network (ANN) to make sense of the basic information from the preprocessor. The ANN has been trained with thousands of examples of visual objects, so the robot eyes can send meaningful information to the robot "brain."