CN-122024303-A - Method and system for identifying image mode of camera module of external equipment of computer

CN122024303ACN 122024303 ACN122024303 ACN 122024303ACN-122024303-A

Abstract

The application relates to the technical field of image mode recognition of a camera module of external equipment of a computer, in particular to a method and a system for recognizing the image mode of the camera module of the external equipment of the computer, wherein the method comprises the following steps of obtaining image data captured by a plurality of camera modules and time information which corresponds to each image data and reflects actual exposure time; the method comprises the steps of obtaining image data of a target object, estimating the motion state of the target object according to the spatial position information of the target object in the image data and combining time information, carrying out time correction on the image data from different camera modules based on the motion state of the target object, and fusing to construct three-dimensional information of the target object so as to identify the defect of the target object. The method realizes the accurate alignment and fusion of the multi-view images, thereby being capable of accurately constructing the three-dimensional information of the target object and identifying the defects, and improving the accuracy of industrial quality inspection and the efficiency of security monitoring.

Inventors

ZENG LI
LIAO HONGHU

Assignees

惠州市合利诚科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. The method for identifying the image mode of the camera module of the external equipment of the computer is characterized by comprising the following steps: Acquiring image data captured by a multi-camera module and time information which corresponds to each image data and reflects actual exposure time; according to the spatial position information of the target object in the image data, and combining time information, deducing the motion state of the target object; based on the motion state of the target object, performing time correction on the image data from different camera modules, and fusing to construct three-dimensional information of the target object so as to identify the defect of the target object.
2. The method for recognizing image patterns of camera module of computer peripheral device according to claim 1, wherein the step of deducing the motion state of the target object based on the spatial position information of the target object in the image data in combination with the time information comprises: The method comprises the steps of carrying out feature point extraction and matching on target objects in image data to form an image observed quantity, wherein the feature point extraction and matching comprises the steps of combining images captured by different camera modules at different time, utilizing parallax produced by the images on the same physical point under different visual angles to determine and explain all observed virtual feature points, establishing a feature point pixel coordinate corresponding relation of the virtual feature points in the images of the camera modules, identifying highlight spots or abnormal brightness change in the image data, aiming at the condition that the feature point pixel coordinate corresponding to the virtual feature points at the shielding time is absent or unreliable when the highlight spots or abnormal brightness change is identified, projecting clear feature points to the shielding time by utilizing undisturbed image data to obtain repaired feature point pixel coordinates, replacing the feature point pixel coordinate of the shielded time by the repaired feature point pixel coordinate to complement the image observed quantity, and the image observed quantity at least comprises a feature point space displacement quantity determined by the pixel coordinate set based on the camera geometric parameters and a feature point space coordinate set and a time stamp determined by the corresponding exposure time stamp. ; defining a state vector containing target object motion information; Predicting the motion state of the target object at the next time step according to the state vector, the current estimated motion state of the target object and a preset motion dynamics model; The predicted motion state is corrected using the image observables to infer the motion state of the target object.
3. The method for recognizing image patterns of camera modules of a computer peripheral device according to claim 1, wherein the step of performing time correction on image data from different camera modules based on a motion state of the target object and fusing to construct three-dimensional information of the target object to recognize defects of the target object comprises: Extracting characteristic points on a target object from images captured by the multi-camera module, and calculating pixel displacement of the characteristic points in different images; based on pixel displacement, predicting theoretical displacement of the feature points in a three-dimensional space by combining camera geometric parameters and a primarily estimated motion state of a target object; Calculating the actual three-dimensional displacement of the feature point at different time points according to the time information which corresponds to the image data and reflects the actual exposure time; comparing the difference between the theoretical displacement and the actual three-dimensional displacement, and identifying the residual time deviation; According to the identified residual time deviation, adjusting time correction parameters of the image data of each camera module; performing time correction on the image data by using the adjusted time correction parameters; And fusing the time corrected image data to construct three-dimensional information of the target object so as to identify the defect of the target object.
4. A method for recognizing an image pattern of a camera module of a computer peripheral device according to claim 3, wherein the step of fusing the time-corrected image data to construct three-dimensional information of the target object comprises: Extracting characteristic points of the internal structure or surface texture of the target object from the image data after time correction to obtain the actual observed characteristic point image projection position; Constructing a light propagation model based on the geometric parameters of the camera and priori knowledge of the optical characteristics of the target object material; comparing the actually observed characteristic point image projection position with the characteristic point image projection position predicted by the light propagation model to obtain a comparison result; according to the comparison result, adjusting the three-dimensional depth information of the feature points and parameters of the light propagation model; And fusing the time corrected image data by utilizing the three-dimensional depth information of the adjusted characteristic points and the parameters of the light propagation model to construct the three-dimensional information of the target object.
5. A method for identifying a camera module image pattern of a computer peripheral device as claimed in claim 3, wherein the step of identifying a defect of the target object comprises: extracting the surface geometric information of the target object from the three-dimensional information of the target object; filtering the surface geometric information to reduce high-frequency noise and local uncertainty; identifying a geometric anomaly region on the surface of the target object according to the filtered surface geometric information; And extracting local texture information of the geometric abnormal region, and judging whether the geometric abnormal region is a defect or not based on the local texture information of the geometric abnormal region.
6. The method for recognizing image patterns of camera module of computer peripheral device according to claim 5, wherein the step of filtering the surface geometry information comprises: extracting geometric features of different scales from the surface geometric information; Judging the local curvature change rate of the geometric features according to the geometric features of different scales; according to the local curvature change rate, candidate defect areas with different scales in the surface geometric information are determined; Selecting a filter kernel of a corresponding scale aiming at candidate defect areas of different scales; and filtering the surface geometric information by using a filtering kernel with a corresponding scale.
7. The method for recognizing image patterns of camera module of computer peripheral device according to claim 6, wherein the step of extracting geometric features of different scales from the surface geometric information comprises: Acquiring surface geometric information of a target object and time data which corresponds to the surface geometric information and reflects actual acquisition time; Identifying the occurrence time and duration of the instantaneous micro-deformation of the surface of the target object according to the time data so as to determine a time window of the instantaneous micro-deformation; continuously sampling surface geometric information in a time window of the instantaneous micro-deformation to obtain a dynamic change sequence of the instantaneous micro-deformation; And according to the dynamic change sequence of the instantaneous micro-deformation, carrying out scale self-adaptive adjustment on the geometric features in the surface geometric information so as to capture the dynamic change geometric features of different scales.
8. The method for recognizing image patterns of camera module of computer peripheral device according to claim 6, wherein the step of determining candidate defect regions of different scales in the surface geometry information according to the local curvature change rate comprises: Acquiring a local curvature change rate sequence of the surface geometric information at a plurality of continuous sampling moments; Analyzing the transient fluctuation characteristics of the local curvature change rate sequence, and identifying a nonlinear fluctuation mode caused by transient deformation; According to the nonlinear fluctuation mode, distinguishing curvature change caused by real defects from curvature change caused by transient deformation; Candidate defect regions of different scales in the surface geometry information are determined based on the distinguished curvature changes.
9. The method for recognizing the image pattern of the camera module of the external equipment of the computer according to claim 1, further comprising the steps of: extracting multi-scale iris features from iris images captured by different camera modules; evaluating the quality of iris images captured by each camera module; According to the image quality and the time information, adjusting the fusion weights of iris characteristics of different camera modules; estimating the movement of eyes or heads of a person wearing the optical glasses in different camera module image acquisition time differences to obtain estimated movement information; according to the estimated motion information, performing space transformation and time correction on the iris characteristics to obtain new iris characteristics; Time alignment is carried out on the new iris features by combining the time information, so that iris features after time alignment are obtained; And carrying out weighted fusion on the iris characteristics after time alignment based on the fusion weight to obtain fusion iris characteristics, and identifying iris characteristics of the personnel wearing the optical glasses based on the fusion iris characteristics.
10. A computer external equipment camera module image pattern recognition system, applying the computer external equipment camera module image pattern recognition method as claimed in claim 1, characterized in that the system comprises: The acquisition module acquires image data captured by the multi-camera module and time information which corresponds to each image data and reflects actual exposure time; the deducing module is used for deducing the motion state of the target object according to the spatial position information of the target object in the image data and the time information; The processing module is used for carrying out time correction on the image data from different camera modules based on the motion state of the target object, and fusing the image data to construct three-dimensional information of the target object so as to identify the defect of the target object.

Description

Method and system for identifying image mode of camera module of external equipment of computer Technical Field The invention relates to the technical field of image mode identification of a camera module of external equipment of a computer, in particular to a method and a system for identifying the image mode of the camera module of the external equipment of the computer. Background In the field of modern industrial production and security monitoring, in order to realize accurate detection of product defects and effective identification of personnel identities, a set of image identification system consisting of a plurality of camera modules of computer external equipment is usually deployed. The system aims at acquiring three-dimensional space information or biological characteristics of a target object by cooperatively acquiring multi-view images and applying an advanced image processing technology. However, in actual operation, due to the influence of various complex factors, such as changes in the production environment, natural aging of equipment components, temporary adjustment of operators, and interference of the external environment, there may be a slight time inconsistency of these camera modules during the image acquisition process, i.e., a so-called "micro-step out" phenomenon. This inconsistency, coupled with the additional challenges to image quality in certain scenarios (e.g., wearing optical glasses), makes it difficult to maintain high accuracy and high reliability with conventional image processing and recognition methods, thereby bringing serious challenges to the accuracy of industrial quality inspection and the efficiency of security monitoring. Disclosure of Invention The invention aims to provide a method and a system for identifying an image mode of a camera module of external equipment of a computer aiming at the defects. The invention adopts the following technical scheme: A method for identifying image modes of a camera module of external equipment of a computer comprises the following steps: Acquiring image data captured by a multi-camera module and time information which corresponds to each image data and reflects actual exposure time; according to the spatial position information of the target object in the image data, and combining time information, deducing the motion state of the target object; based on the motion state of the target object, performing time correction on the image data from different camera modules, and fusing to construct three-dimensional information of the target object so as to identify the defect of the target object. According to the technical scheme, the problem of micro-step loss in multi-camera module image acquisition can be effectively solved, and accurate alignment and fusion of multi-view images are realized through deducing the motion state of the target object and correcting the time of image data, so that three-dimensional information of the target object can be accurately constructed and defects can be identified, and the accuracy of industrial quality inspection and the efficiency of security monitoring are remarkably improved. The application also discloses a system for identifying the image mode of the camera module of the external equipment of the computer, which applies the method for identifying the image mode of the camera module of the external equipment of the computer, and the system comprises the following steps: The acquisition module acquires image data captured by the multi-camera module and time information which corresponds to each image data and reflects actual exposure time; the deducing module is used for deducing the motion state of the target object according to the spatial position information of the target object in the image data and the time information; The processing module is used for carrying out time correction on the image data from different camera modules based on the motion state of the target object, and fusing the image data to construct three-dimensional information of the target object so as to identify the defect of the target object. According to the technical scheme, the system capable of realizing the method is provided, the effective implementation of the method is ensured through the modularized design, and reliable hardware support is provided for industrial quality inspection and security monitoring. The application effectively solves the problems of 'micro-step-out' phenomenon in the prior art of multi-camera module image acquisition, such as inaccurate image alignment, reduced three-dimensional reconstruction precision, missed detection or false alarm of defect identification, and the like. By accurate time correction and motion state inference, microsecond-level and even nanosecond-level time deviation caused by factors such as production environment change, equipment aging, parameter adjustment and the like can be overcome, and the high consistency of the multi-view images in the time dimensi