Sensor integration system utilizing fuzzy inference with LED displacement sensor and vision system

The authors deal with a 3-D measurement system applied to a curved-metal-surface carving system, and a multisensor integration system based on fuzzy inference. The measurement system consists of two different sensors, a light-emitting-diode (LED) displacement sensor and a vision system. The LED displacement sensor is used as a part of the vision system, which is based on the active stereo sensing method. In addition, the sensor´s outputs are used for calibrating camera parameters. Therefore, the system can calibrate the camera parameters easily. Neural networks are used to compensate for the output of the image processing system for some errors, such as camera parameter error and lens distortion. By utilizing the neural network, a vision system can be used with as high an accuracy as possible. A sensor integration method based on fuzzy inference is used. The fuzzy inference input consists of information on the change in the sensor output and the position change of the sensor system, together with the environmental data of the measurements. For this integration system, the sensory system can be used accurately. The proposed system has been shown to be effective through extensive experiments