Software And Hardware Implementation Of The Information Processing Method For A 3-D Vision Sensing System

Industrial robots are used largely in modern manufacturing with the development of flexible assembly and intelligent manufacturing system. As the objects being operated by the robots distribut randomly and the target trajectory changes uncertainly in the process of tracking, the sensor system based on machine vision requires high speed, high accuracy, simple data output. Developing real time and universal three-dimensional machine vision sensing system for the industrial applications related to intelligent detection and pattern recognition is of great significance.Machine vision measurement technology is used widely for the non-contact distance measurement, which has the advantage of high precision and high speed, simple operation and maintenance, and so on. Thanks to the improvement of the measurement accuracy and exploration of the application in many fields, active structured light method, a mature method of the machine vision measurement, is being used to develop a variety of industrial applications of 3-D measurement system in recent years.This sensing system is based on the laser triangulation structured point light three-dimensional machine vision measurement. A point laser reflected by a rotating prism scans the object being measured and the reflection spot center on the CCD can be located to get the 3-D position point. The whole three-dimensional contour of the object is constructed by the discrete point one by one, after moving the sensor along the vertical scanning cross section.This article describes a smart sensor measurement system which can be used easily and reduce the amount of PC calculation.Image preprocessing circuit board is designed by the Protel software tool using the core FPGA processor EP2C20. The signal acquisition module, FPGA core processing module and USB communication module are integrated on the circuit board. Image acquisition experiment is done to test the stability and connect condition of the board, and complete test images are obtained successfully by the serial port debugging tool in the PC.FPGA logic program is designed in the Quartus software to make the gray images processed quickly by the image preprocessing circuit board. Median filter module and average filter module are constructed to remove the noises in the images and extreme value module is designed to get the position of spot center. Six bytes effective data including the spot center position, the gray value and the current angle of the scan lasor are transferred to the computer via USB communication control module. A New fast median filter module is designed, which can be used extendly in real-time image processing for the simple algorithm and high speed.Software application is developed for the machine vision measurement sensor in the VC++6.0 development environment. Sensor data received from the USB communication module is transformed to three-dimensional position data of the object, and the real-time data showed on the windows is generously designed and easy to use. In the OpenGL graphics interface, three-dimensional color display model is designed with illumination rendering, scaling and rotation model translation.The pixel coordinate and displacement coordinate of the camera are established to make size measurement easy. Calibration experiment shows that maximum nonlinear error between pixel number and displacement is 0.7428% in the measurement scope of 80mm-210mm, and mean square error is 441.6μm. An easy scanning angle error measurement method of the laser is proposed and the measure accuracy of mean square error along scanning cross section reaches 62.26μm after two experimental calibration in the flat scan experiment. The sensor is fixed on the cartesian robot to scan some typical components and the 3-D outline showed in the host computer is clear.The measurement error of the vision-based robot system is small, which is the foundation for the research related to vision-based pattern recognition and smart robot control applications.