Research On Spherical Multi-Target Precision Real-Time Location Algorithm

Accurate real-time positioning of target objects is a key technology in areas such as precision control and machine navigation.This technology is widely used in industrial manufacturing,military aerospace,agricultural production and clinical medicine.In order to achieve the precise positioning of the target better,some special markers will be used to identify the target feature points in practice.And then use the spatial position relationship between the marker and the target object.In this way the target position can be achieved via the positioning of the marker.When the target is carried out by the marker,the positioning accuracy will be reduced due to factors such as ambient light,background disturbance,camera distortion,mutual occlusion of the marker,and so on.In order to solve these problems,this paper firstly constructs an experimental system that uses infrared reflector as the target feature point marker and shoots with infrared camera under infrared light source,which reduces the influence of ambient light and background disturbance on target location.At the same time,according to the imaging characteristics of the spherical target,the effects of using the circumcircle method(Circumcircle),the least squares fitting circle method(LSFC),the Hough circle method(Hough)and the centroid method(Centroid)to locate the complete spherical target are analyzed in detail.Finally,in order to solve the problem of positioning multiple spherical targets,this paper proposes a least squares fitting circle algorithm based on corner detection optimization(LFSC-corner),this algorithm can accurately and quickly locate the target center when the target is obscured and the target image is missing.In order to verify the accuracy and real-time performance of the algorithm described in this paper,a series of experiments are designed.The experimental results show that the average positioning error of Circumcircle,LSFC,Hough and Centroid are 0.38mm,0.25mm,0.49mm and 0.31mm.A single average positioning time are 14.3ms,13.5ms,52.8ms,11.7ms.In the case of the incomplete spherical target in the case of occlusion,the LFSC-corner algorithm has an error of 0.33mm for positioning a spherical target that is not more than 1/4,and an error of 0.53mm for a spherical target that is not more than 1/2,and a single average positioning time of about 14.8ms.Therefore,in the spherical multi-target location problem,when the target imaging is complete,LSFC has the highest positioning accuracy,and the real-time performance of Centroid is the strongest.When the target occlusion causes the image to be incomplete,LSFC-corner is better than other algorithms,and has high positioning accuracy and real-time performance.