Algoirthm Research Of Single Camera Optical Probe Imaging

Through our eyes to see the world is a complex process. How to use thecomputer to achieve this process is a dream in scientists heart. The computer has theability to understand the scene.Realization of computer vision research is tosimulate the human eye see the world. The traditional understanding of computervisual means according to the camera by accessing the two-dimensional image andpresenting valuable data analysis by the actual measured object and scene.The basicquestion is understanding the real objects of space three-dimensional by themeasured target acquisition of two-dimensional pictures. However, in order to makea value judgment, we need to construct the mathematical model, then carries on theanalysis.The process is completeing through the image which from camera lens.The latest and fastest developed technology is the vision coordinatemeasurement technology. This new technology includes computer vision and threecoordinate measurement field,uses their advanced technology and solves the newproblems in production and life through the reform and innovation.This paper mainly research work is to build a monocular vision measurementsystem based on visual detection and probe coordinates,which is for a target of thelarge object used in aerospace, defense, medical, industrial areas.And it alsointroduces how to solve the measured target spatial position,in order to achievepurpose that a target to be measured from the two-dimensional images intothree-dimensional space spatial position.The main contents include the followings:1. Analysis the development status which of the vision coordinatesmeasurement technology. Choose which method is based on the characteristic pointimage from single camera optical measuring head. Achieve a real sense of visioncoordinates measurement on line.2. Study of single camera vision’s coordinate measurement system, elaboratedthe working principle of the system, and established the mathematical model. Mathematical model describes the measurement system’s problem. According tocamera projection model using triangular geometric principle of sine theorem,mathematical principles listed below according to the equation, which given aspecific algorithm for feature point coordinate.3. Study of the two kinds of feature points under the camera coordinate framealgorithm. One is the value iteration algorithm, it is the traditional calculationproblem of P3P, because its slow calculation speed, low measurement accuracy, andrestrict by initial value selection,poor in noise filtering, we choose a singular valuedecomposition based optimization algorithm. The algorithm has linear, singlecharacteristic, make full use of large amounts of data redundancy, avoid theiterative optimization.4.Introduce a kind of method which is solving transform matrix bycorresponding points, it is the most basic and commonly used method.Its purpose issolving the measure point’s coordinate, after linear constraint, the results meet thesystem requirement.5.Analysis of various error, verify the accuracy of solving transformation matrixalgorithm, through a large number of experiments, verify the measured method isfeasible and the system is stable.