The Design Of Stereo Vision Control Systembased On LabVIEW

As an important branch of computer vision, stereo vision can sense the objective world very well by simulating the ways of human visual imaging system. It was of great significance to improve the measurement accuracy and the automation degree of the stereo vision system by equipping a control system with it.After investigating the issues of synchronous controlling and acquiring between the two CCDs in binocular stereo vision system and the external sensor, and the stereo vision system’s calibration, one control system of stereo vision based on LabVIEW was designed. This system was developed on LabVIEW platform, through which controlling the data acquisition card PCIE8620 triggering the acquisition of images and force value synchronously, and the frequency of synchronous acquisition by changing the trigger signal’s cycle. In order to realize the calibration plate images’ pseudo-color enhancement and the stereo vision system’s calibration, we used mixed programming technology to call the Matlab in the LabVIEW software. The practical study includes:We made the PCIE8620 DAQ card output three synchronous TTL pulse signals, of which two signals triggered the two CCDs to acquire images, the remain one triggered the AD port in the data acquisition card to collect force value synchronously by using the virtual instrument consisted of LabVIEW and multi-function PCIE8620 data acquisition card. The error analysis shows that the synchronous error between the images and the force value data were both less than 10μs, relative to the system’s largest acquisition frequency 7 HZ, the maximum error of synchronous acquisition is 0.007%. To apply the synchronous acquiring and controlling stereo vision system into prefabricated crack of aluminum alloy plate’s three-point bending experiment, the front panel of LabVIEW’s total control worked well in the experiment, and the maximum error between the force value collected by PCIE8620 and the electronic universal testing machine was 2.89%, which indicate the real-time synchronous acquisition of images and the force value can be completed.The pseudo-color enhancement of calibration plate image was realized by using hybrid programming technology to call the Matlab through Matlab Script in LabVIEW, Then the inside-outside parameters of stereo vision system was achieved by invoking Matlab calibrate toolbox, and the calibration error was less than 0.3 pixel.