Research On Calibration Method Of Combined Panoramic Camera With Constraints

Panoramic camera is a hotspot in the fields of photogrammetry and computer vision in recent years.Camera calibration is a key link in the measurement process.The accuracy of the calibration results and the stability of the algorithm directly affect the accuracy of the results of the work.The traditional method is to calibrate the internal and external orientation elements of each sub-camera in the same reference coordinate system,then use the external parameters to indirectly solve the relative constraint relationship.This type of method is simple and convenient,but does not utilize the relative orientation relationship,the calibration results are easy to fluctuate.In this paper,a calibration method of combined panoramic camera with constraints is proposed to obtain robust calibration results.An experimental calibration field with multi-markers is built by using a 360 degree composite panoramic camera with eight uniformly distributed network cameras,and an automatic identification method of markers is proposed.Firstly,the least squares ellipse fitting method is used to detect and identify the markers,and then the correlation between the markers is used to encode the markers automatically.The calibration content of the combined panoramic camera includes determining the internal and external orientation elements of the sub-camera and the fixed relative orientation relationship between the sub-cameras.By establishing the calibration model of the self-calibration beam adjustment combined panoramic camera,the intrinsic geometric constraints are introduced into the calibration of the panoramic camera to improve the calibration accuracy and robustness.Experiments have shown that automatic recognition of landmarks can reduce interference and increase speed and overall efficiency.Compared with the traditional methods,the introduction of constraints makes the calibration results less fluctuating,which can effectively improve calibration accuracy and robustness.It has reference value for the construction of similar systems and its application to 3D reconstruction.