Holographic Coding Target And Image Backprojection Correction Matching

Coding targets are widely used in visual measurement.For example,the use of coding targets in visual photogrammetry can achieve accurate matching of images taken at multiple angles.If coding targets is used in the coordinate measurement system of the target imaging,the imaging field of view is reduced and high-precision measurement is achieved.The design of the coding target as well as the rapid and accurate matching of the target are of practical significance to the development of visual measurement technology.Moreover,coding targets are also widely used in many fields,such as robot motion control systems,large-scale high-precision detection systems,and three-dimensional surface detection of large-size objects.This paper proposes a holographic coding target and the coding target is applied to pose measurement to improve the range and accuracy of the measurement.A decoding algorithm for the match of image back-projection correction is proposed to solve the problem that affine transformation cannot correctly decode large-angle imaging.The decoding method can also be applied in the decoding and matching in the photogrammetric measurement,and is not limited by the coding rules,which can effectively improve the decoding accuracy.The specific research work is as follows:1.The current situation of pose measurement is analyzed and the coding target is proposed to improve the measurement.Faced with the research status of coding targets,the design and decode of new coding target is of practical significance.2.A planar coding target with holographic property is designed.The imaging position and the overall state of the target can be obtained by regional coding information.The holographic property can be realized by only three primitives through coding combination.The decoding and measurement can be acquired by more than 4 encoding units.The azimuth property is also added to the target,which greatly reduces the difficulty of decoding.3.Aimed at the problem of decoding and matching of encoding targets,the idea of separating rotation parameters and translation parameters is proposed.A decoding algorithm based on a combination of correction by target image back-projection and correlation matching is proposed.In this method,the target image is reconstructed in three-dimensional space by perspective transformation.Subsequently,it is rotated and corrected in the three-dimensional space,and then projected back onto the original image surface.The decoding algorithm solves the problem that affine transformation cannot correctly decode large-angle imaging.4.The correlation matching algorithm is improved when image matching is used to correct translation parameters.Only the marked points on the coding target are searched instead of all the pixel points in the image.The matching efficiency is improved greatly and the matching time is shorten.5.Regarding the matching range,the affine transformation correction is only 15° while the back-projection correction proposed in this paper can reach 60°.Regarding the matching time,the average matching time of the ordinary matching algorithm is 2.1s while the value of the improved method in this paper is 0.4s.