The Study On The Calibration Mode Of The General Measuring Robot Based On Stereo Vision

The general measuring robot system (for short, the system) based on stereo vision isone of the significant engineering fields of the vision measuring technologyapplications. In this system, the robot is the moving carrier of the stereo vision sensor.This structure type not only extends the working space of the sensor, but also keepsthe vision measuring technology's characteristics of non-contract, quickness, etc. Bycontrolling the pose of the robot in space, the stereo vision sensor can arrive at thegiven points one by one, collect the image signals of the given point, and then obtainthe three dimension coordinate data after computing the image data. The key of thissystem is to establish the relationship between the vision sensor frame and theworkpiece frame by the way of system calibration when the system is at any pose inspace. The main contents are as follows:1. Put forward three calibration schemes: ① unity method;② synchronizationmethod;③subsection method .Fully discusses and compares everyone's technologycharacteristics and confirm that the subsection method is the best.2. Analyse the subsection method and the process in detail.①Uses the hand-eyeequation AX=XB to solve the homogeneous transform relationship between the visionsensor frame and the last robot link frame;②On this base, Utilizing the simple droneclamp which will be designed and manufactured additionally, the robot's outer pose,the homogeneous transform relationship between robot's base frame and theworkpiece frame, can be solved;③Thus, the three homogeneous transform factors,the vision sensor frame ——the robot last link frame——the robot base frame——the workpiece frame, are all known and the system's coordinate operation model canbe established when the robot is at any pose in space.3. Write the programs and take simulative tests for the two abovementionedprocesses. The results of emulation data indicate that the subsection calibrationmethod satisfy the system precision in the condition of non-noise.4. The subsection method has the error delivery, so the error compensation methodto the 42 error elements is discussed in detail in the last section using the robot'sdifferential kinematics and vision measuring technology.5. The thesis discusses and confirms the system's best calibration method atworkshop and solves the error compensation problem, improve the system's precision.Make a theory foundation for the system's popularization and application.