CN-115830141-B - Method and device for acquiring calibration parameters and electronic equipment

CN115830141BCN 115830141 BCN115830141 BCN 115830141BCN-115830141-B

Abstract

The embodiment of the invention discloses a calibration parameter acquisition method and device and electronic equipment. The method comprises the steps of obtaining motion tracks of a plurality of targets in a first mode and a second mode, roughly matching the motion tracks of the targets in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion tracks of the targets, obtaining point pair information of the targets according to the preliminary matching result and time information corresponding to the motion tracks of the targets, and finely matching the motion tracks of the targets in the first mode and the second mode according to the point pair information of the targets to obtain a final matching result corresponding to the motion tracks of the targets. The invention solves the technical problem of poor multi-mode information fusion effect caused by the fact that the multi-mode equipment cannot timely and accurately realize registration correction, and achieves the technical effects of improving the real-time performance and the accuracy of the multi-mode equipment for realizing registration correction and ensuring the validity of multi-mode information fusion.

Inventors

WANG HUI
LU HAIXIAN

Assignees

杭州海康威视数字技术股份有限公司

Dates

Publication Date: 20260512
Application Date: 20221213

Claims (9)

1. The method for acquiring the calibration parameters is characterized by comprising the following steps: acquiring motion tracks of a plurality of targets in a first mode and a second mode; performing rough matching on the motion trail of each target in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion trail of each target; Acquiring point pair information of each target according to a preliminary matching result and time information corresponding to the motion trail of each target, wherein the point pair information is the center point coordinates of a target detection frame which are simultaneously generated in the first mode and the second mode; Carrying out fine matching on the motion trail of each target in the first mode and the second mode according to the point pair information of each target to obtain a final matching result corresponding to the motion trail of each target, wherein the final matching result corresponding to the motion trail of each target is used as a calibration parameter between the first mode and the second mode; Performing rough matching on the motion trail of each target in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion trail of each target, wherein the rough matching comprises the following steps: Acquiring track line characteristics and/or characteristic descriptors corresponding to motion tracks of the targets in the first mode and the second mode, wherein the track line characteristics comprise target pose information and a moving direction, the target pose information is the center point coordinates of a target detection frame, the moving direction is the tangential direction of the motion tracks, and the characteristic descriptors are used for representing regional characteristic information formed by interleaving a plurality of motion tracks; and carrying out feature matching on track line features and/or feature descriptors corresponding to the motion tracks of the targets in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion tracks of the targets.
2. The method for obtaining calibration parameters according to claim 1, wherein obtaining motion trajectories of a plurality of targets in a first mode and a second mode comprises: Acquiring characteristic point information of each target in the first mode and the second mode, wherein the characteristic point information comprises center point coordinates and time information of a target detection frame; And processing the characteristic point information of each target in the first mode and the second mode to obtain the motion trail of each target in the first mode and the second mode.
3. The method for obtaining calibration parameters according to claim 2, wherein processing the feature point information of each target in the first mode and the second mode to obtain the motion trail of each target in the first mode and the second mode includes: Performing target tracking on each target according to the characteristic point information of each target in the first mode and the second mode to acquire initial motion tracks of each target in the first mode and the second mode; smoothing the initial motion trail of each target in the first mode and the second mode to obtain the motion trail of each target in the first mode and the second mode.
4. The method for obtaining calibration parameters according to claim 1, wherein performing feature matching on track line features corresponding to motion tracks of the targets in the first mode and the second mode to obtain preliminary matching results corresponding to the motion tracks of the targets, includes: Pairing object pose information corresponding to the motion trail of each object in the first mode and the second mode to obtain a transformation relation corresponding to the motion trail of each object in the first mode and the second mode; According to the moving directions of the movement tracks of the targets in the first mode and the second mode, the transformation relation of the movement tracks of the targets in the first mode and the second mode is adjusted, and the target transformation relation of the movement tracks of the targets in the first mode and the second mode is obtained; According to the target transformation relation corresponding to the motion trail of each target in the first mode and the second mode, calculating the coincidence degree corresponding to the motion trail of each target in the first mode and the second mode respectively; and screening the target transformation relation corresponding to the motion trail of each target in the first mode and the second mode according to the coincidence degree corresponding to the motion trail of each target in the first mode and the second mode, and determining a preliminary matching result corresponding to the motion trail of each target.
5. The method for obtaining calibration parameters according to claim 4, wherein the step of screening the target transformation relationship corresponding to the motion trail of each target in the first mode and the second mode according to the coincidence ratio corresponding to the motion trail of each target in the first mode and the second mode, and determining the preliminary matching result corresponding to the motion trail of each target comprises: Screening the coincidence degrees corresponding to the motion trail of each target in the first mode and the second mode, and determining the optimal coincidence degree of each target; And taking the target transformation relation corresponding to the optimal contact ratio of each target as a preliminary matching result corresponding to the motion trail of each target.
6. The method for obtaining calibration parameters according to claim 1, wherein performing feature matching according to feature descriptors corresponding to motion trajectories of the targets in the first mode and the second mode to obtain preliminary matching results corresponding to the motion trajectories of the targets, comprises: Registering feature descriptors corresponding to the motion trail of each target in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion trail of each target.
7. An apparatus for obtaining calibration parameters, comprising: the first acquisition module is used for acquiring motion tracks of a plurality of targets in a first mode and a second mode; the rough matching module is used for rough matching the motion trail of each target in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion trail of each target; The second acquisition module is used for acquiring point pair information of each target according to a preliminary matching result and time information corresponding to the motion trail of each target, wherein the point pair information is the center point coordinates of a target detection frame which are simultaneously generated in the first mode and the second mode; The fine matching module is used for carrying out fine matching on the motion trail of each target in the first mode and the second mode according to the point pair information of each target to obtain a final matching result corresponding to the motion trail of each target, wherein the final matching result corresponding to the motion trail of each target is used as a calibration parameter between the first mode and the second mode; wherein, the coarse matching module includes: The second acquisition unit is used for acquiring track line characteristics and/or characteristic descriptors corresponding to the motion tracks of the targets in the first mode and the second mode, wherein the track line characteristics comprise target pose information and a moving direction, the target pose information is the center point coordinate of a target detection frame, the moving direction is the tangential direction of the motion tracks, and the characteristic descriptors are used for representing regional characteristic information formed by interleaving a plurality of motion tracks; And the characteristic matching unit is used for carrying out characteristic matching on track line characteristics and/or characteristic descriptors corresponding to the motion tracks of the targets in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion tracks of the targets.
8. An electronic device, comprising: A processor; A memory for storing processor-executable instructions; wherein the processor is configured to perform the method of obtaining calibration parameters of any one of claims 1 to 6.
9. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored program, wherein the program when run controls a device in which the computer readable storage medium is located to perform the method of obtaining the calibration parameter according to any one of claims 1 to 6.

Description

Method and device for acquiring calibration parameters and electronic equipment Technical Field The present invention relates to the field of computer vision, and in particular, to a method and an apparatus for obtaining calibration parameters, and an electronic device. Background The multi-mode equipment is usually subjected to joint calibration before delivery, usually information of the same group of data under each mode is obtained based on a calibration plate, and then mapping calculation is performed to obtain a mapping matrix, so that registration among modes is realized. However, the practical problem is that the transportation and installation process and the like often cause physical deviation, for example, the multi-mode equipment (generally calibrated when shipped) may cause fine deviation of respective mode positions during the installation and the like, which may cause registration deviation of the multi-mode on some remote object detection, and larger mode deviation may also be caused along with the increase of the running time of the equipment, for example, the registration accuracy of the multi-mode equipment may be slowly reduced due to the aging of mechanical firmware and the like during the use process, for example, RGB/Thermal is better aligned, the two-mode viewing angles coincide, and after the deviation, the two-mode deviation, ghost and the like occur. If the operation is required to be maintained, the calibration and even the factory return calibration are often required to be carried out with the aid of human beings. In theory, under two RGB modes, two images with stronger characteristic textures are respectively obtained, so that the registration of the images can be realized, namely the on-line calibration is realized, but if the equipment is installed in a relatively open area, the better registration correction cannot be realized if the texture features are absent. In addition, because the existing multi-mode is often cross-modal, such as RGB and Thermal, RGB and Lidar, the traditional graph-based registration is basically invalid, and the correction cannot be realized by the existing single-mode registration mode due to the large difference between modes. For the above-described problems, no effective solution has been proposed. Disclosure of Invention The embodiment of the invention provides a calibration parameter acquisition method, a calibration parameter acquisition device and electronic equipment, which are used for at least solving the technical problem of poor multi-mode information fusion effect caused by the fact that the multi-mode equipment cannot timely and accurately realize registration correction. According to one aspect of the embodiment of the invention, an acquisition method of calibration parameters is provided, which comprises the steps of acquiring motion tracks of a plurality of targets in a first mode and a second mode, performing rough matching on the motion tracks of the targets in the first mode and the second mode to obtain a preliminary matching result corresponding to the motion tracks of the targets, acquiring point pair information of the targets according to the preliminary matching result corresponding to the motion tracks of the targets and time information, wherein the point pair information is center point coordinates of a target detection frame which simultaneously appears in the first mode and the second mode, performing fine matching on the motion tracks of the targets in the first mode and the second mode according to the point pair information of the targets to obtain a final matching result corresponding to the motion tracks of the targets, and taking the final matching result corresponding to the motion tracks of the targets as the calibration parameters between the first mode and the second mode. Optionally, acquiring motion trajectories of a plurality of targets in a first mode and a second mode comprises acquiring characteristic point information of each target in the first mode and the second mode, wherein the characteristic point information comprises center point coordinates and time information of a target detection frame, and processing the characteristic point information of each target in the first mode and the second mode to obtain the motion trajectories of each target in the first mode and the second mode. Optionally, processing the characteristic point information of each target in the first mode and the second mode to obtain the motion trail of each target in the first mode and the second mode, wherein the processing comprises the steps of carrying out target tracking on each target according to the characteristic point information of each target in the first mode and the second mode to obtain the initial motion trail of each target in the first mode and the second mode, and carrying out smoothing processing on the initial motion trail of each target in the first mode and the second mode to obtain the motion trail of