A method and apparatus for determining that pose of a camera and an electronic device

Abstract

A method and apparatus for determining that pose of a camera and an electronic device are disclosed. The method for determining that pose of a camera may include determining a plurality of first pixelpoints having a preset semantics in a current frame image captured by the camera; Determining a plurality of second pixel points having preset semantics in a previous frame image preceding the current frame image, wherein the plurality of first pixel points correspond to at least a portion of the plurality of second pixel points; Determining an error between the plurality of first pixel points and the plurality of second pixel points, the error comprising a semantic error; Determining a camera pose change value of the current frame image relative to the previous frame image based on the error; And determining the camera posture corresponding to the current frame image based on the camera posture corresponding to the previous frame image and the camera posture change value of the current frame image relative to the previous frame image. The camera pose determination method according to the embodiment of the present application can be calculated directly on the semantic segmentation result, thereby avoiding the influence of individual differences between cameras.

Description

technical field [0001] The present application relates to the technical field of image processing, and more specifically, to a camera pose determining method, determining device and electronic equipment. Background technique [0002] The visual mileage calculation method has been widely used in the navigation field of robots and unmanned vehicles, such as smart cars and drones. The measurement error caused by the vehicle skidding of the type odometer also avoids the measurement error caused by factors such as the decrease of sensor accuracy, and improves the measurement accuracy. [0003] In the visual odometry calculation method, the calculation method of the camera pose is its core module. In the currently commonly used camera pose calculation methods, the method of minimizing the reprojection error is often used to calculate the camera pose. In this process, corner points need to be extracted and feature matching is performed; if the corner points are too sparse and the ...

AI Technical Summary

Problems solved by technology

In the currently commonly used camera pose calculation methods, the method of minimizing the reprojection error is often used to calculate the camera pose. In this process, corner points need to be extracted and feature matching is performed; if the corner points are too sparse and the map is sparse, The error of feature matching is very easy to cause system failure, and it is difficult to determine the threshold for feature matching screening and wrong matching elimination

Images