Panoramic video stitching method and device

Abstract

The invention discloses a panoramic video stitching method and device. The panoramic video stitching method is used for stitching video images acquired by front-end devices to be stitched into a panoramic video image through a client and comprises the steps of binding the front-end devices to be stitched into a resource group, acquiring four vertices of a stitched video image part of each front-end device so as to form a cut quadrilateral corresponding to each corresponding front-end device, calculating the deformity correction and conversion matrix of each front-end device according to the acquired vertices and a set standard rectangle, converting the video images acquired by all the front-end devices, and then, laying out the converted video images according to set position information, so as to form and display a panoramic video image. The panoramic video stitching device disclosed by the invention comprises a configuration module, a vertex acquisition and matrix generation module and a stitching module. According to the method and the device, disclosed by the invention, the deformity correction and conversion can be carried out by the front-end devices or the client so as to realize panoramic stitching, and the requirements on the performance of matrix conversion equipment are not high, so that the method and the device are particularly applicable to the upgrade and development of old items.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a panoramic video splicing method and device in video surveillance. Background technique [0002] With the development of video surveillance, customers have higher and higher requirements for the viewing angle and visual range of surveillance video. Promote the gradual emergence and popularity of panoramic video stitching, but limited by the system framework and hardware performance, there are not many products on the market that can meet customer requirements, and the visual range of the stitched video is relatively limited by hardware performance, which affects the panorama Promotion and popularization of video stitching solutions. [0003] In video surveillance, multiple cameras are often used to monitor different angles of the same scene to obtain multiple video images of the same scene at different angles, and then the multiple video images are spliced ...

AI Technical Summary

Problems solved by technology

[0005] However, the first solution relies heavily on the mechanical structure of the front-end equipment, and has relatively high requirements for the installation of the front-end equipment. At the same time, the picture after splicing is relatively single, and it is generally only suitable for the splicing effect of two front-end equipment.

The second solution needs to convert the pixels of the video screen captured by each front-end device because of the splicing pixel conversion matrix. After decoding, the look-up table conversion of pixel points consumes more computer performance. Generally, a splicing server with relatively strong performance is required to complete the panorama. video splicing

And the complexity of the stitching algorithm will increase sharply with the increase of stitching front-end equipment, and there are relatively strict requirements for the presentation mode of the video images after panoramic stitching, generally does not support the stitching of irregular images, such as the panorama stitching of pin-shaped images

At the same time, due to splicing on the client side, when the performance is insufficient, load balancing cannot be performed, and a server with better performance must be used instead. The requirements for splicing servers are high and the cost is relatively high.

Images