Line structure cursor calibration method based on projection geometry

Abstract

The invention discloses a line structure cursor calibration method based on projection geometry, which comprises the following steps of: acquiring a fringe image which comprises a laser and is projected on a calibration plate, and solving a camera internal parameter by utilizing a tensor calibration method. Information of related straight lines and intersection points is obtained through image processing, a blanking line equation corresponding to a target plane is obtained through homogeneous coordinate expression of parallel straight lines, then blanking point coordinates corresponding to laser stripes are obtained, and a direction vector of the laser stripes in a camera coordinate system is obtained through further conversion; the Euclidean distance between two feature points on a laserstripe is obtained by using a projection property with an invariable cross ratio, three-dimensional coordinates of the feature points in a camera coordinate system are obtained by combining a direction vector and imaging model information and combining an equation set, plane fitting is performed on the coordinates of all the image feature points by using a least square method to obtain a light plane equation, and line structure cursor calibration is completed. According to the method, the characteristics and properties of projection geometry are used for calibration, other measurement devicesare not needed, the precision is high, and the method is simple and feasible.

Description

technical field [0001] The invention relates to the field of visual structured light measurement, in particular to a calibration method for line structured light based on projective geometry. Background technique [0002] With the development of industrial cameras and computers, visual inspection is widely used in industrial inspection. Among them, the structured light vision system is more widely used because it does not need to paste or customize marker points, contains more information and has high precision. The most common one is the line For the measurement system combining structured light and monocular camera, before using the system for detection, the expression of the light plane in the camera coordinate system must be calibrated, and then the camera obtains the image of the measured object carrying the light bar information, and the perspective imaging model combines light Plane equations to realize three-dimensional measurement. [0003] Common calibration metho...

AI Technical Summary

Problems solved by technology

The wire drawing method is based on the fact that bright spots will be formed when the light plane is projected onto thin lines with non-coplanar spatial distribution. The three-dimensional coordinates of the bright spots in space are measured by other instruments, and the light plane equation is obtained by combining the imaging coordinates of the bright spots in the camera. This method requires With the help of other instruments and the diffuse reflection produced by the light spot on the filament will bring a large error; the sawtooth target method uses a sawtooth target and a one-dimensional working platform to achieve higher precision calibration, but this method needs to adjust the workbench Make the light plane perpendicular to the ridgeline of the target, the operation is more complicated, and the requirements for the target are higher; the method based on the cross-ratio invariance is to use the collinear points of the perspective projection to satisfy the property of the cross-ratio invariance, and calculate the coordinates of the points on the laser line on the target The three-dimensional coordinates in the system, and then converted to the camera coordinate system through the relative external parameters of the camera and each image, and finally plane fitting, this method is simple to operate, but the error of the external parameter matrix will cause the marker points to be in the camera coordinate system error, and the obtained light plane points are less, the calibration method based on the neural network is to obtain the three-dimensional coordinates of the feature points with the help of the light point aiming device and the high-precision mobile platform, use the sample data to train the BP neural network, and optimize the search for the light plane equation coefficients. This method has high precision, but requires other high-precision instruments, and the operation is complicated and not suitable for on-site calibration

Images