Large microphone array fast coordinate calibration method and system based on visual markers

Abstract

The application discloses a large microphone array fast coordinate calibration method based on visual markers, which comprises the following steps: after the microphone array is arranged on site, an array original image containing visual markers and reference points is collected, and then the array original image is sequentially subjected to distortion correction and perspective transformation to obtain a corrected array image; in the corrected array image, the visual markers corresponding to the positions of the microphones are automatically identified to obtain image coordinates of the centers of the visual markers; according to the mapping relationship between the image coordinates and array physical coordinates, the image coordinates are converted into plane coordinates in an actual array coordinate system, so that the actual acoustic center position coordinates of the microphones are obtained; the obtained actual acoustic center position coordinates of the microphones are numbered and matched with design coordinates, real measurement coordinates or reference coordinates, and a complete microphone array geometric model is reconstructed. The application further discloses a large microphone array fast coordinate calibration system based on visual markers.

Description

Technical Field

[0001] This invention relates to the field of microphone array technology, and in particular to a method and system for rapid coordinate calibration of large microphone arrays based on visual markers. Background Technology

[0002] Large microphone arrays are testing systems that use multiple spatially distributed microphones to synchronously acquire sound pressure signals and combine this with array signal processing algorithms to achieve sound source localization, sound field reconstruction, and acoustic parameter inversion. These arrays are widely used in aircraft flyby noise testing, wind tunnel acoustic experiments, mobile sound source localization, and industrial equipment noise source identification. In array acoustic testing, the spatial coordinates of the microphone elements are fundamental parameters of the array's geometric model. If there is a deviation between the actual installation coordinates of the microphones and their design coordinates, it will lead to distortion of the geometric phase relationship between channels, thus affecting the accuracy of beamforming sound source localization results.

[0003] Existing methods for obtaining microphone array coordinates mainly include inputting coordinates according to design drawings, manual measurement with a measuring tape, laser ranging, total station measurement, and 3D scanning measurement. For small microphone arrays with a limited number of channels, regular structure, and stable installation environment, these methods can meet basic usage requirements. However, for large arrays containing dozens to hundreds of microphones, especially circular arrays, nested inner and outer ring arrays, sparse arrays, distributed arrays, or irregular arrays, existing coordinate acquisition methods still have significant shortcomings in engineering applications.

[0004] First, directly using the design coordinates cannot reflect the actual geometric state of the array after on-site installation. During the actual deployment of large microphone arrays, factors such as array bracket manufacturing errors, assembly errors, uneven ground, mounting hole deviations, microphone clamp offsets, cable pulling, deformation during handling, and repeated disassembly and assembly can easily affect the actual microphone positions, causing them to deviate from the theoretical design positions. If acoustic calculations are performed directly using the design coordinates without on-site calibration, the array's geometric model will be inconsistent with the actual array state, thus affecting subsequent sound source localization and acoustic inversion results.

[0005] Secondly, manual measurement methods struggle to balance efficiency and reliability. For large arrays, manual point-by-point measurement requires multiple steps, including point location locating, number verification, distance measurement, coordinate conversion, and data entry. This results in a large workload, long operation cycle, and is easily affected by factors such as personnel experience, measurement posture, reading errors, and recording mistakes. When there are many microphones or the array is irregularly arranged, the risk of mismatch between microphone numbers, actual installation locations, and recorded coordinates increases further.

[0006] Third, large-scale arrays often cannot cover all array elements in a single measurement, and segmented measurements can easily lead to inconsistencies in local coordinate references and error accumulation. Operational differences between different measurement segments, different measurement personnel, or different test batches can also reduce the consistency and repeatability of array geometric data. Traditional measurement methods typically only output coordinate tables, lacking visual evidence corresponding to the actual deployment status on site, making it difficult to quickly verify abnormal coordinates, missed measurement points, mismatched points, or installation offset points.

[0007] Therefore, in response to the needs of acquiring on-site coordinates and geometric calibration of large microphone arrays, those skilled in the art are dedicated to developing a coordinate calibration method and system that can quickly acquire array deployment status in a real test environment, automatically identify microphone positions, convert image information into unified array coordinates, complete the matching of measured coordinates with design coordinates, and generate error statistics and anomaly verification results. Summary of the Invention

[0008] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is how to reduce the dependence on point-by-point manual measurement and high-precision surveying equipment, improve the efficiency of large array field deployment and the reliability of coordinate data, and provide an accurate and verifiable array geometric model for subsequent sound source localization, sound field reconstruction and acoustic inversion.

[0009] To achieve the above objectives, this invention provides a method for rapid coordinate calibration of a large microphone array based on visual markers, comprising the following steps: Step 1: After completing the on-site deployment of the microphone array, the original array image containing visual markers and reference points is acquired, and then subjected to distortion correction and perspective transformation to obtain the corrected array image. The visual markers are set at one or more locations on the microphone body, microphone clamp, and adjacent fixed positions of the microphone. The visual markers are one or more of the following: color markers, reflective markers, coded markers, circular markers, and other markers that can be stably identified by the image. Step 2: In the corrected array image, automatically identify the visual markers corresponding to the microphone positions and obtain the image coordinates of the center of the visual markers; Step 3: Based on the mapping relationship between image coordinates and array physical coordinates, convert the image coordinates into planar coordinates in the actual array coordinate system to obtain the actual acoustic center position coordinates of the microphone. Step 4: Match the obtained actual acoustic center position coordinates of the microphone with the design coordinates, actual measurement coordinates or reference coordinates by numbering, and reconstruct the complete microphone array geometric model.

[0010] Furthermore, step 1 specifically includes the following steps: Step 1.1: After completing the on-site deployment of the microphone array, use an image acquisition device to acquire the original array image containing visual markers and reference points; the image acquisition device includes one of the following: a regular camera, an industrial camera, a mobile phone, and a video camera. Step 1.2: Perform distortion correction on the original image based on the pre-completed camera calibration results; the camera calibration results include camera intrinsic parameters, distortion parameters, and camera pose parameters. Step 1.3: Establish the mapping relationship between the image plane and the array physical plane using reference points at several known positions in the array plane, and perform perspective transformation accordingly; Step 1.4: Through perspective transformation, the tilted image is converted into a unified array planar view, so that the coordinates of the subsequently identified marker points can be converted within the same geometric framework.

[0011] Furthermore, in step 1.1, if video acquisition is used on-site, then image frames with clear images and complete markings are selected from the video as the original images of the array.

[0012] Furthermore, when the visual markers use color markings, step 2 specifically includes the following steps: Step 2.1: Convert the corrected image to a color space suitable for color segmentation, and determine the recognition range based on the color characteristics of the marked points in the on-site image; Step 2.2: Perform color segmentation on the image to obtain candidate marker regions; Step 2.3: Remove noisy regions, shadow interference, and non-target regions through morphological processing, connected component analysis, contour filtering, or shape constraints; Step 2.4: For the retained valid marked regions, calculate their center positions; the center positions are obtained by contour centroid, connected region center, fitted circle center or other center extraction methods, and are used as the positions of the microphone visual markers in the image.

[0013] Furthermore, step 3 specifically includes the following steps: Step 3.1: Establish the mapping relationship between the image plane and the array physical plane by using the position of the reference point in the image and the actual position of the reference point in the array coordinate system; Step 3.2: Based on the mapping relationship between image coordinates and array physical coordinates, convert the center point of each identified visual marker into the X and Y coordinates in the array coordinate system; Step 3.3: If there is a fixed geometric offset between the visual marker center and the microphone acoustic center, then after completing the coordinate transformation, the coordinates are corrected according to the fixed geometric offset relationship between the visual marker center and the microphone acoustic center to obtain the actual acoustic center position coordinates of the microphone.

[0014] Furthermore, step 4 specifically includes the following steps: Step 4.1: Use the global optimal matching method, that is, calculate the distance between all design coordinate points and all measurement coordinate points, construct the distance cost matrix, and combine matching constraints and rejection mechanism to determine the overall optimal one-to-one matching relationship; Step 4.2: After matching is completed, calculate the X-direction error, Y-direction error and planar position error of each microphone, and statistically analyze the average error, maximum error, error dispersion and the ratio of error to the array equivalent aperture. Step 4.3: Determine whether the calibration results meet the array geometric accuracy requirements of the test based on the above calculation results, and identify the points with large errors.

[0015] Furthermore, the microphone array includes a central region array and an external extended array. For the central region array, the original image of the array is acquired using a holistic shooting method, and the microphone coordinates are obtained according to steps 1 to 4. For the external extended array, a holistic shooting method or a partitioned shooting method is selected based on the array coverage area. When partitioned shooting is used, each shooting area is unified to the global array coordinate system through a common reference point. After coordinate calibration is completed, error evaluation is performed on the central region array and the external extended array respectively. The error evaluation of the external extended array is determined by combining the array equivalent aperture and the target analysis frequency.

[0016] This invention also provides a method for rapid coordinate calibration of a large microphone array based on visual markers, comprising the following steps: Step 1: Divide the large microphone array into multiple shooting zones; each shooting zone covers a portion of the microphones in the array; set at least one common reference area or common reference point between adjacent shooting zones to establish coordinate relationships between different zones; Step 2: Acquire images of each shooting zone separately. For each zone image, perform image correction, visual marker recognition, and local coordinate transformation to obtain the measurement coordinates of the microphone within the zone in the local coordinate frame. Step 3: Using the common reference points between adjacent partitions, the local coordinates of each partition are uniformly transformed to the same global array coordinate system; for cases with multiple common reference points, the translation, rotation, and scale relationships between partitions are constrained by multiple reference points to reduce the impact of errors of a single reference point on the overall coordinate unification result; Step 4: Match the measurement coordinates of all microphones in the global array coordinate system with the design coordinates or reference coordinates, and output the overall array coordinate results and error analysis results.

[0017] Furthermore, step 2 specifically includes the following steps: Step 2.1: After completing the on-site deployment of the microphone array, the original array images containing visual markers and reference points in each zone are acquired, and then subjected to distortion correction and perspective transformation to obtain the corrected array images. Step 2.2: In the calibrated array image, automatically identify the visual markers corresponding to the microphone positions in each partition to obtain the image coordinates of the visual marker centers; Step 2.3: Based on the mapping relationship between image coordinates and local coordinates, convert the image coordinates into planar coordinates in the local coordinate system to obtain the local coordinates of the microphone acoustic center position.

[0018] This invention also provides a rapid coordinate calibration system for large microphone arrays based on visual markers, comprising: A microphone array to be calibrated, the microphone array comprising multiple microphones located in the same array plane; Visual markings are provided at one or more locations on the microphone body, microphone clamp, and adjacent fixed positions of the microphone. The visual markings are one or more of the following: color markings, reflective markings, coded markings, circular markings, and other markings that can be stably identified by an image. Reference point, which is set in the array plane; An image acquisition device, comprising one of a conventional camera, an industrial camera, a mobile phone, and a video recording device; the image acquisition device is configured to acquire an array of raw images containing the visual markers and the reference points; and A computing device is configured to acquire the original image of the array from the image acquisition device, and sequentially perform distortion correction and perspective transformation to obtain a corrected array image; in the corrected array image, automatically identify the visual markers corresponding to the microphone positions to obtain the image coordinates of the center of the visual markers; according to the mapping relationship between the image coordinates and the array physical coordinates, convert the image coordinates into planar coordinates in the actual array coordinate system to obtain the actual acoustic center position coordinates of the microphone; and match the obtained actual acoustic center position coordinates of the microphone with design coordinates, actual measurement coordinates, or reference coordinates by numbering to reconstruct a complete microphone array geometric model.

[0019] The beneficial effects of this invention are as follows: 1. This invention transforms point-by-point manual measurement into batch image recognition and coordinate calculation, significantly improving the efficiency of large array coordinate acquisition; reducing reliance on high-precision surveying equipment such as total stations and laser rangefinders; reducing the risk of manual recording and microphone number mismatch, and improving the repeatability of airport field test deployment.

[0020] 2. This invention reduces the impact of camera shooting angle, lens distortion, and viewing angle changes on position calibration results; improves the accuracy of image coordinate to array physical coordinate conversion; enables ordinary image acquisition equipment to meet the microphone array position calibration requirements, and reduces system deployment costs.

[0021] 3. This invention establishes a stable one-to-one correspondence between microphone numbers and visual measurement coordinates; reduces duplicate matching, local mismatches, and missed matching; can quantitatively determine whether the calibration results meet the geometric accuracy requirements of fly-through noise testing, and quickly locate microphone points with large errors, providing a reliable array geometric model for on-site verification and subsequent sound source inversion.

[0022] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0023] Figure 1 This is a flowchart of a microphone array position calibration method according to a preferred embodiment of the present invention; Figure 2 This is a preferred embodiment of the present invention applied in fly-through noise testing. Detailed Implementation

[0024] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0025] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0026] Example 1: Overall Process of Microphone Array Position Calibration Based on Computer Vision

[0027] like Figure 1 As shown, this embodiment provides a computer vision-based microphone array position calibration method, applicable to large ground microphone arrays, especially suitable for inner-ring high-density arrays and outer-ring extended arrays in aircraft fly-through noise testing. The method mainly includes four steps: image correction, feature recognition, coordinate transformation, and number matching.

[0028] Step 1: Image Correction

[0029] After the microphone array is deployed on-site, image acquisition equipment is used to capture array images containing microphone visual markers and reference points. This equipment can be a regular camera, industrial camera, mobile phone, or video camera. If video acquisition is used on-site, clear, minimally occluded, and fully marked image frames are selected for subsequent processing. After image acquisition, distortion correction is first performed on the original images based on pre-calibrated camera parameters. These parameters include camera intrinsics, distortion parameters, and camera pose parameters. Distortion correction reduces the impact of lens distortion on marker location identification. After distortion correction, a mapping relationship between the image plane and the array physical plane is established using reference points at several known locations within the array plane, and perspective transformation is performed accordingly. Through perspective transformation, the tilted images are converted into a unified array plane view, allowing subsequent identification of marker points to be performed with coordinate conversion within the same geometric framework.

[0030] The purpose of this step is to reduce the impact of lens distortion, camera posture, and shooting angle on the microphone position recognition results, and to provide a unified image basis for subsequent feature recognition and coordinate transformation.

[0031] Step 2: Feature Recognition

[0032] In the calibrated array image, visual markers corresponding to the microphone positions are automatically identified. These visual markers can be placed on the microphone body, microphone clamp, or at a fixed position adjacent to the microphone. Visual markers can be color markers, reflective markers, coded markers, circular markers, or other markers that can be stably identified by the image. When using color markers, the calibrated image is converted to a color space suitable for color segmentation, and the recognition range is determined based on the color characteristics of the marker points in the field image. Subsequently, color segmentation is performed on the image to obtain candidate marker regions. Then, morphological processing, connected component analysis, contour filtering, or shape constraints are used to remove noise regions, shadow interference, and non-target regions. For the remaining valid marker regions, their center positions are calculated. This center position can be obtained through contour centroid, connected component center, fitted circle center, or other center extraction methods, and is used as the position of the microphone marker in the image.

[0033] The purpose of this step is to extract the corresponding microphone positions in batches from the calibrated image, reducing the workload of manually reading and recording coordinates point by point, and improving the efficiency of large array coordinate acquisition.

[0034] Step 3: Coordinate Transformation

[0035] After obtaining the image coordinates of the visual marker center, the image coordinates are converted into planar coordinates in the actual array coordinate system based on the mapping relationship between the image coordinates and the array physical coordinates. Specifically, the proportional, affine, or homography mapping relationship between the image plane and the array physical plane can be established using the position of the reference point in the image and its actual position in the array coordinate system. Based on this mapping relationship, the center point of each identified visual marker is converted into X and Y coordinates in the array coordinate system. If there is a fixed geometric offset between the visual marker center and the microphone acoustic center, the coordinates are corrected according to this fixed offset relationship after the coordinate transformation, thereby obtaining the actual acoustic center position coordinates of the microphone.

[0036] The purpose of this step is to convert the image recognition results into array geometric coordinates that can be directly used in subsequent acoustic calculations.

[0037] Step 4: Number Matching

[0038] The microphone measurement coordinates obtained from computer vision are matched with design coordinates, actual measurement coordinates, or reference coordinates to reconstruct a complete microphone array geometric model. A globally optimal matching method can be used, which calculates the distance between all design coordinate points and all measurement coordinate points, constructs a distance cost matrix, and combines matching constraints and discard mechanisms to determine the overall optimal one-to-one matching relationship. This method can reduce duplicate matching or local mismatches that may be caused by local nearest neighbor matching. After matching, the X-direction error, Y-direction error, and planar position error of each microphone are calculated, and the average error, maximum error, error dispersion, and the ratio of error to the array's equivalent aperture are statistically analyzed. The results are used to determine whether the calibration results meet the array geometric accuracy requirements of the experiment and to identify points with large errors.

[0039] The purpose of this step is to establish the correspondence between the microphone number and the measurement coordinates, and to evaluate the quality of the calibration results and verify outliers.

[0040] Example 2: Application Example in Flyover Noise Testing

[0041] like Figure 2 As shown, in one application embodiment, the method in Embodiment 1 is used for ground microphone array calibration in civil large aircraft overflight noise testing. The array includes a high-density array in the central region and an extended array in the outer region. The central region array is mainly used to ensure the accuracy of sound source localization at higher frequencies, while the outer extended array is mainly used to expand the array's equivalent aperture and improve the spatial characterization capability of the low-frequency sound radiation region.

[0042] Before testing, visual markers were placed at the microphone positions of both the central array and the external extended array, and reference points were established within the array plane. For the central array, a holistic imaging approach was used to acquire array images, and microphone coordinates were obtained through image correction, feature recognition, coordinate transformation, and number matching. For the external extended array, either a holistic or segmented imaging approach could be selected based on the array's coverage area. When segmented imaging was used, all imaging areas were unified to the global array coordinate system using a common reference point.

[0043] After coordinate calibration, error evaluations were performed on the central region array and the outer extended array, respectively. The central region array, primarily serving the localization of higher-frequency sound sources, requires higher accuracy in positional accuracy; the outer extended array, primarily serving the spatial representation of lower-frequency sound sources, can have its error evaluation determined by combining the array's equivalent aperture and the target analysis frequency.

[0044] This application example illustrates that the method is applicable to composite ground microphone arrays in real-world fly-through noise testing, meeting both the accuracy requirements of high-density arrays and adapting to the deployment characteristics of large-aperture extended arrays.

[0045] Example 3: Extended Example of Large Array Partition Calibration

[0046] In one embodiment, when the microphone array has a large coverage area and a single image is difficult to simultaneously meet both global coverage and local recognition accuracy, a partition calibration method is adopted.

[0047] First, the large microphone array is divided into multiple shooting zones. Each shooting zone covers a portion of the microphones in the array. At least one common reference area or common reference point is set between adjacent shooting zones to establish coordinate relationships between different zones.

[0048] Next, images of each shooting zone are acquired separately. For each zone image, image correction, visual marker recognition, and local coordinate transformation are performed in accordance with the method in Example 1 to obtain the measurement coordinates of the microphone in that zone within the local coordinate frame.

[0049] Then, using the common reference points between adjacent partitions, the local coordinates of each partition are uniformly transformed to the same global array coordinate system. In cases where multiple common reference points exist, the translation, rotation, and scaling relationships between partitions can be constrained by these multiple reference points to reduce the impact of errors from a single reference point on the overall coordinate unification result.

[0050] Finally, all microphone measurement coordinates in the global array coordinate system are matched with the design coordinates or reference coordinates by number, and the overall array coordinate results and error analysis results are output.

[0051] This embodiment addresses the challenges of insufficient single-image coverage, significant perspective distortion at image edges, and the difficulty in balancing local recognition accuracy with global coverage in ultra-large-scale or distributed arrays. By using partitioned shooting and unified coordinates based on common reference points, this method can adapt to a wider range of array calibration scenarios.

[0052] In real airport field environments, microphone arrays are large in scale, numerous in number, and dispersed across a wide area. Relying on manual measurement or high-precision surveying equipment to obtain coordinates point-by-point suffers from low efficiency, high cost, implementation limitations, and a high risk of mismatched numbering. This invention proposes a computer vision-based batch calibration method for microphone positions. Through image acquisition, visual marker recognition, and coordinate transformation, it quickly obtains the planar position coordinates of each microphone in the array. Visually recognizable markers are placed on the microphone body, reflector, or other fixed locations. Array images or video frames containing multiple microphone markers are acquired using a camera, mobile phone, or industrial camera. Image processing methods are used to automatically identify the center position of the markers and convert the image coordinates into microphone coordinates in the array's physical coordinate system.

[0053] This invention transforms point-by-point manual measurement into batch image recognition and coordinate calculation, significantly improving the efficiency of large array coordinate acquisition; reducing reliance on high-precision surveying equipment such as total stations and laser rangefinders; reducing the risk of manual recording and microphone number mismatch; and improving the repeatability of airport field test deployment.

[0054] To address the issue that outdoor images are easily affected by lens distortion, camera pose, and tilted viewing angles, and that directly determining microphone coordinates based on image pixel positions introduces geometric errors, this invention proposes an array plane correction method that combines distortion correction and perspective transformation. This method maps outdoor images uniformly to an array plane coordinate framework. Camera intrinsic parameters and distortion parameters are obtained through camera calibration to correct lens distortion in the original image. Then, using reference points at known locations within the array plane, a perspective transformation relationship is established to convert the tilted image into an array plane view, enabling the identification and coordinate calculation of marker points within the unified geometric framework.

[0055] This invention reduces the impact of camera shooting angle, lens distortion, and viewing angle changes on position calibration results; improves the accuracy of image coordinate to array physical coordinate conversion; enables ordinary image acquisition equipment to meet the microphone array position calibration requirements, and reduces system deployment costs.

[0056] To address the issue that the set of microphone measurement points obtained through visual recognition must accurately correspond to the design array numbers, otherwise issues such as duplicate matching, local mismatches, and missed detection points may easily occur, affecting the reliability of the final array geometric model. This invention proposes an array geometric model reconstruction method based on number matching and error verification. Complete and reliable microphone array coordinate results are obtained through one-to-one matching and error analysis. The distance between the visually recognized measurement coordinates and the design coordinates or reference coordinates is calculated. One-to-one nearest neighbor matching under threshold constraints is used, or a distance cost matrix is ​​constructed and a global optimal matching method is employed to determine the correspondence between microphone numbers and measurement coordinates. After matching, the X-direction error, Y-direction error, and planar position error of each microphone are calculated, and the average error, maximum error, standard deviation of error, and error ratio relative to the array aperture are statistically analyzed.

[0057] This invention establishes a stable one-to-one correspondence between microphone numbers and visual measurement coordinates; reduces duplicate matching, local mismatches, and missed matching; can quantitatively determine whether the calibration results meet the geometric accuracy requirements of fly-through noise testing, and quickly locate microphone points with large errors, providing a reliable array geometric model for on-site verification and subsequent sound source inversion.

[0058] This invention proposes a rapid coordinate calibration method and system for large microphone arrays based on visual markers, primarily targeting large array acoustic testing scenarios such as aircraft fly-through noise testing, vehicle passing noise testing, train noise testing, wind tunnel acoustic testing, mobile sound source localization, and industrial equipment noise source localization. This solution addresses issues such as low efficiency in acquiring on-site coordinates for large microphone arrays, errors in manual measurement, difficulty in reflecting actual installation conditions with designed coordinates, and lack of verification basis for calibration results. It possesses strong engineering practicality and industrial application value.

[0059] From a technical perspective, this invention transforms the traditional coordinate acquisition method, which relies on manual point-by-point measurement or professional surveying equipment, into an automated calibration process based on image acquisition, visual marker recognition, image geometric correction, coordinate mapping, and number matching. By acquiring images once or in a few batches, the planar position coordinates of multiple microphones in a unified array coordinate system can be obtained in batches, thereby reducing the workload of on-site point-by-point measurement, manual numbering and verification, and data processing. For large arrays containing dozens to hundreds of microphones, this solution can significantly improve the efficiency of array deployment and verification, and reduce the risk of mismatches between microphone numbers, physical locations, and coordinate records.

[0060] From a calibration reliability perspective, this invention maps on-site captured images to a unified array planar coordinate framework through camera calibration, distortion correction, and perspective transformation. This reduces the impact of lens distortion, shooting angle, and perspective on coordinate recognition results. By extracting the visual marker center and mapping the coordinates, the actual planar coordinates of each microphone in the array coordinate system can be obtained. For array areas with stable installation conditions, good image quality, and reasonable reference point placement, this solution can achieve high-precision planar position calibration. For extended arrays with large coverage areas, it can also ensure that the calibration results meet the accuracy requirements of engineering testing for the array geometric model through partitioning, reference point constraints, and error verification.

[0061] From an engineering implementation perspective, the hardware required for this invention mainly includes image acquisition equipment, computing equipment, visual markers, reference points, and a microphone array to be calibrated. The image acquisition equipment can be a regular camera, industrial camera, mobile phone, or video camera. The visual markers can be color-coded, reflective, coded, or other markers that facilitate image recognition. Therefore, this invention does not rely on expensive or complex specialized surveying equipment, and features convenient equipment acquisition, flexible on-site deployment, and low implementation costs, making it suitable for application in complex environments such as airport field sites, road test sites, wind tunnel test sites, and industrial sites.

[0062] From a software implementation perspective, this invention can consist of functional modules such as image acquisition, image correction, feature recognition, coordinate transformation, number matching, error analysis, and result output. Each module can be implemented using conventional image processing, geometric transformation, and numerical calculation methods, facilitating the formation of independent array coordinate calibration software. It can also be integrated into flyby noise testing systems, vehicle pass noise testing systems, wind tunnel acoustic testing systems, or sound source localization systems as an automatic array geometric model generation and quality control module.

[0063] From a test quality control perspective, this invention not only outputs microphone coordinate results but also generates number matching results, coordinate deviations, error statistics, and anomaly alerts. Test personnel can use these results to determine whether the array geometry model meets the test requirements and perform on-site reshoots, manual verification, or reinstallation and adjustment for points with significant errors, thereby preventing erroneous coordinates from directly entering subsequent beamforming, sound source localization, and acoustic inversion calculations.

[0064] In summary, this invention has the advantages of convenient on-site implementation, low hardware cost, high calibration efficiency, quantifiable results, verifiable process, and large applicable array size. It can be widely used in the fields of aviation noise, traffic noise, wind tunnel acoustics, and industrial noise testing, and has good prospects for engineering transformation and industrial application.

[0065] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for rapid coordinate calibration of a large microphone array based on visual markers, characterized in that, Includes the following steps: Step 1: After completing the on-site deployment of the microphone array, the original array image containing visual markers and reference points is acquired, and then subjected to distortion correction and perspective transformation to obtain the corrected array image. The visual markers are set at one or more locations on the microphone body, microphone clamp, and adjacent fixed positions of the microphone. The visual markers are one or more of the following: color markers, reflective markers, coded markers, circular markers, and other markers that can be stably identified by the image. Step 2: In the corrected array image, automatically identify the visual markers corresponding to the microphone positions and obtain the image coordinates of the center of the visual markers; Step 3: Based on the mapping relationship between image coordinates and array physical coordinates, convert the image coordinates into planar coordinates in the actual array coordinate system to obtain the actual acoustic center position coordinates of the microphone. Step 4: Match the obtained actual acoustic center position coordinates of the microphone with the design coordinates, actual measurement coordinates or reference coordinates by numbering, and reconstruct the complete microphone array geometric model.

2. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 1, wherein, Step 1 specifically includes the following steps: Step 1.1: After completing the on-site deployment of the microphone array, use an image acquisition device to acquire the original array image containing visual markers and reference points; the image acquisition device includes one of the following: a regular camera, an industrial camera, a mobile phone, and a video camera. Step 1.2: Perform distortion correction on the original image based on the pre-completed camera calibration results; the camera calibration results include camera intrinsic parameters, distortion parameters, and camera pose parameters. Step 1.3: Establish the mapping relationship between the image plane and the array physical plane using reference points at several known positions in the array plane, and perform perspective transformation accordingly; Step 1.4: Through perspective transformation, the tilted image is converted into a unified array planar view, so that the coordinates of the subsequently identified marker points can be converted within the same geometric framework.

3. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 2, wherein, In step 1.1, if video acquisition is used on site, then image frames with clear images and complete markings are selected from the video as the original images of the array.

4. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 1, wherein, When visual markings use color markings, step 2 specifically includes the following steps: Step 2.1: Convert the corrected image to a color space suitable for color segmentation, and determine the recognition range based on the color characteristics of the marked points in the on-site image; Step 2.2: Perform color segmentation on the image to obtain candidate marker regions; Step 2.3: Remove noisy regions, shadow interference, and non-target regions through morphological processing, connected component analysis, contour filtering, or shape constraints; Step 2.4: For the retained valid marked regions, calculate their center positions; the center positions are obtained by contour centroid, connected region center, fitted circle center or other center extraction methods, and are used as the positions of the microphone visual markers in the image.

5. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 1, wherein, Step 3 specifically includes the following steps: Step 3.1: Establish the mapping relationship between the image plane and the array physical plane by using the position of the reference point in the image and the actual position of the reference point in the array coordinate system; Step 3.2: Based on the mapping relationship between image coordinates and array physical coordinates, convert the center point of each identified visual marker into the X and Y coordinates in the array coordinate system; Step 3.3: If there is a fixed geometric offset between the visual marker center and the microphone acoustic center, then after completing the coordinate transformation, the coordinates are corrected according to the fixed geometric offset relationship between the visual marker center and the microphone acoustic center to obtain the actual acoustic center position coordinates of the microphone.

6. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 1, wherein, Step 4 specifically includes the following steps: Step 4.1: Use the global optimal matching method, that is, calculate the distance between all design coordinate points and all measurement coordinate points, construct the distance cost matrix, and combine matching constraints and rejection mechanism to determine the overall optimal one-to-one matching relationship; Step 4.2: After matching is completed, calculate the X-direction error, Y-direction error and planar position error of each microphone, and statistically analyze the average error, maximum error, error dispersion and the ratio of error to the array equivalent aperture. Step 4.3: Determine whether the calibration results meet the array geometric accuracy requirements of the test based on the above calculation results, and identify the points with large errors.

7. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 1, wherein, The microphone array includes a central area array and an outer extended array. For the central area array, the original image of the array is acquired by overall shooting, and the microphone coordinates are obtained according to steps 1 to 4. For the outer extended array, the overall shooting or partition shooting method is selected according to the array coverage. When partition shooting is used, each shooting area is unified to the global array coordinate system through a common reference point. After coordinate calibration, error evaluations were performed on the central region array and the outer extended array respectively; the error evaluation of the outer extended array was determined by combining the array equivalent aperture and the target analysis frequency.

8. A method for rapid coordinate calibration of a large microphone array based on visual markers, characterized in that, Includes the following steps: Step 1: Divide the large microphone array into multiple shooting zones; each shooting zone covers a portion of the microphones in the array; set at least one common reference area or common reference point between adjacent shooting zones to establish coordinate relationships between different zones; Step 2: Acquire images of each shooting zone separately. For each zone image, perform image correction, visual marker recognition, and local coordinate transformation to obtain the measurement coordinates of the microphone within the zone in the local coordinate frame. Step 3: Using the common reference points between adjacent partitions, transform the local coordinates of each partition to the same global array coordinate system; In cases where there are multiple common reference points, the translation, rotation, and scaling relationships between partitions are constrained by multiple reference points to reduce the impact of errors of a single reference point on the overall coordinate unification result; Step 4: Match the measurement coordinates of all microphones in the global array coordinate system with the design coordinates or reference coordinates, and output the overall array coordinate results and error analysis results.

9. The method for rapid coordinate calibration of a large microphone array based on visual markers according to claim 8, wherein, Step 2 specifically includes the following steps: Step 2.1: After completing the on-site deployment of the microphone array, the original array images containing visual markers and reference points in each zone are acquired, and then subjected to distortion correction and perspective transformation to obtain the corrected array images. Step 2.2: In the calibrated array image, automatically identify the visual markers corresponding to the microphone positions in each partition to obtain the image coordinates of the visual marker centers; Step 2.3: Based on the mapping relationship between image coordinates and local coordinates, convert the image coordinates into planar coordinates in the local coordinate system to obtain the local coordinates of the microphone acoustic center position.

10. A rapid coordinate calibration system for a large microphone array based on visual markers, characterized in that, include: A microphone array to be calibrated, the microphone array comprising multiple microphones located in the same array plane; Visual markings are provided at one or more locations on the microphone body, microphone clamp, and adjacent fixed positions of the microphone. The visual markings are one or more of the following: color markings, reflective markings, coded markings, circular markings, and other markings that can be stably identified by an image. Reference point, which is set in the array plane; Image acquisition device, including one of a conventional camera, an industrial camera, a mobile phone, and a video recording device; The image acquisition device is configured to acquire an array of raw images containing the visual markers and the reference points; as well as A computing device configured to acquire the original array image from the image acquisition device, and sequentially perform distortion correction and perspective transformation to obtain a corrected array image; In the corrected array image, the visual markers corresponding to the microphone positions are automatically identified to obtain the image coordinates of the center of the visual markers; based on the mapping relationship between the image coordinates and the array physical coordinates, the image coordinates are converted into planar coordinates in the actual array coordinate system to obtain the actual acoustic center position coordinates of the microphone. The obtained actual acoustic center position coordinates of the microphone are matched with the design coordinates, actual measurement coordinates or reference coordinates by numbering, and the complete microphone array geometric model is reconstructed.

Images