Research On NI FlexRIO High Speed Vision Measurement System

With the implementation of "Made in China 2025",the visual measurement system will play an important role in the field of industrial automation in China.Due to the faster moving speed of the measured object,the traditional visual measurement system may cause frame loss due to the lower frame rate.This requires increasing the processing speed of the visual measurement system to compensate for its shortcomings in the fast motion field.Therefore,this paper will develop a high-speed vision measurement system based on NI FlexRIO.The specific research work is as follows.Firstly,the overall design of the high-speed vision measurement system is carried out,and the target pose is measured by means of monocular vision and cooperative goals.According to the system design requirements,select the appropriate high-speed camera.The lens focal length is calculated based on the operating range of the system and the field of view,and the lens model is determined.Set up the NI FlexRIO system hardware platform and determine the model of each sub-module according to the needs of the system.The target used is analyzed and designed taking into account factors such as late feature point matching and mutual occlusion.Based on the NI FlexRIO system,LabVIEW was selected as the software platform,and the software scheme was designed.Secondly,research on algorithms related to high-speed vision measurement systems.The image feature point centroid extraction algorithm is implemented in the FPGA module,which is the key to improve the processing speed of the system.Based on the FPGA image feature point centroid extraction algorithm,the single-cycle acquisition pixel is used to identify and mark the connected domain.At the same time,the connected regions between the uplink and the downlink are merged,and finally the centroid coordinates of the connected connected domain are calculated.The image feature point matching algorithm reorders the acquired feature point centroid coordinates to make one-to-one correspondence with the target LED feature points.Because the system adopts the cooperative target method,the PnP algorithm is used to solve the pose of the target.Then,the software design of the high-speed visual measurement system is mainly divided into two parts: FPGA and host controller.The FPGA part first synchronizes its clock frequency with the high-speed camera clock to facilitate image data transmission,and then caches the acquired image data for centroid extraction of image feature points.The data transmission module is mainly used for data transmission between the FPGA part and the host controller part.The host controller data recording module is mainlyused for fast data recording,which is convenient for later analysis.The software interface design is divided into FPGA interface design and host controller interface design.The former is used to display system parameters,threshold range of image processing,system status;the latter is used for image processing threshold setting,system parameter configuration,system status Display and so on.Finally,a high-speed visual measurement system based on NI FlexRIO is built to verify whether the system processing speed,image feature point centroid stability,pose resolution repeatability,measurement error,etc.meet system specifications.When verifying the processing speed of the system,the system frame rate is set to 100 fps and190 fps for pose measurement,and the system frame loss rate is counted.When verifying the image feature point centroid stability and pose rectification repeatability,multiple measurements were made at the measurement distances of 2 m and 5.5 m,respectively,and the measurement data were statistically analyzed.When verifying the measurement error of the system,the measurement error of 2 m and 5.5 m is used respectively,and the biaxial sway-linear lifting motion calibration platform with displacement accuracy of 5 is used as the standard of system displacement to detect the measurement error of the system.