A 3D digital modeling method based on dome video stream

Abstract

The invention relates to a three-dimensional digital modeling method based on a ball screen video stream. The method comprises the following steps of S1, holding a ball screen camera to carry out video shooting according to a route required by a user; S2, acquiring a hand-held ball screen camera video stream; S3, automatically extracting a key frame from the video stream; S4, using super-resolution algorithm to increase the image quality of a key frame image; S5, tracking pixel points on the key frame image to obtain dense point cloud; S6, selecting a tracked stable pixel point and using a PnPalgorithm to estimate a camera motion track so as to determine a camera position and an orientation, wherein the stable pixel point is the point in the middle area of the key frame image and the continuous tracking frame number of the stable pixel point is greater than 10 frames; and S7, carrying out three-dimensional modeling according to the dense point cloud obtained in the step S5 and the camera position and the orientation obtained in the step S6. By using the three-dimensional digital modeling method provided in the invention, fixed point shooting is not needed, three-dimensional modelestablishing stability is high and the fluency of a visual effect is high.

Description

Technical field [0001] The invention relates to the technical field of three-dimensional imaging modeling, and relates to a three-dimensional digital modeling method, in particular to a three-dimensional digital modeling method based on a spherical screen video stream. Background technique [0002] In the field of 3D digital modeling technology, the positioning of the camera is an important basic technical indicator. The traditional camera positioning is to take a picture at a certain point by taking a picture, then change to another point to take the picture, and then use two or more points. The position of the photos taken at each point is compared to establish a three-dimensional model. The video frames obtained in this way are discontinuous, so data offset or errors will occur, which will lead to the introduction of major errors in the modeling process, which becomes A big problem in the field of 3D digital modeling. [0003] In order to solve the above problems, technicians g...

AI Technical Summary

Problems solved by technology

This leads to incomplete feature points, which can only be tracked for the previously extracted 500 feature points, so the tracking stability will be very poor

On the other hand, since the effect presentation of the traditional method is transitioned through point-to-point, there will be a stuck situation; and the scene roaming method of establishing a model through the existing method can only present a guided or structured The effect, not many details, not refined enough

So the customer experience is not good

Images