Camera system for head-mounted support device and image generation method thereof

Abstract

This invention discloses a dual-camera recording system for a head-mounted support and its image generation method. The two cameras employ outward-biased geometry to form a controlled central overlap area. The system includes an intermediate bus module that provides a common time base and synchronous encapsulation without image fusion, and transmits the left and right images to a mobile computing device via a single host link. The mobile device performs surface projection, local geometric alignment, and multi-band mixing to output a single horizontal wide-angle image. The method can generate depth information and distance measurements, and supports backtracking, automatic binocular / monocular switching, and separate storage of the fused image and the original left and right images. The system adopts a modular design, allowing mode selection based on the number of cameras and image quality specifications.

Claims

1. A recording system for a head-mounted support device, comprising: Two camera modules are configured to be integrated with a head-mounted support. The optical axes of the two camera modules are each deflected outward by an external angle φ between 10° and 30° relative to the wearer's line of sight in the horizontal plane. The horizontal viewing angle θ of each single lens is between 65° and 95°, and the central overlap angle C = (θ - 2φ) is between 15° and 75°. An intermediate bus module is detachably coupled to the two camera modules. The intermediate bus module is used for data aggregation and clock synchronization of the two camera modules, but not for image fusion. The intermediate bus module is configured to encapsulate and transmit a synchronized left image and a right image to a mobile computing device. One or more processors are located on the mobile computing device and configured to execute a fusion program that projects the left image and the right image onto a parametric coordinate system of a cylindrical or spherical projection model, and performs local geometric alignment based on one or a combination of features or optical flow within the overlap area corresponding to the central overlap angle, generating a single horizontal wide-angle image through multi-band mixing.

2. As in claim 1, a recording system for a head-mounted support, wherein, The headband support is selected from at least one of the following groups: temples or frames of eyeglasses, headbands of goggles, hats, and fastening parts of helmets.

3. The camera system for a head-mounted support as claimed in claim 1 further includes a mounting interface configured to detachably attach the two camera modules to the head-mounted support; the attachment method being selected from one or a combination of the following groups: clamping, slide rail latching, quick-release locking, screws, standard photographic threads, adhesive, overmolding, replacement of a section of an existing component, or magnetic alignment with a mechanical load-bearing structure.

4. The recording system for a head-mounted support as claimed in claim 1 further includes a low-power inertial measurement unit (IMU) rigidly connected to the at least one camera module for attitude compensation and timestamp synchronization; wherein, in the case of a small overlap angle where the central overlap angle C is between 15° and 25°, and under the condition of measurement in the wearing state, the processor of the intermediate bus module or the mobile computing device is configured to: set a minimum working distance greater than or equal to 0.60m; and at least one of the following is performed: (a) enabling local non-rigid distortion compensation for depth / parallax perception; (b) referencing the relative attitude prior of the low-power inertial measurement unit (IMU).

5. The recording system for a head-mounted support as described in claim 1, wherein, The recording path does not include a digital image stabilization (EIS) processing module, while the camera module includes an optical image stabilization (OIS) actuator and its feedback interface.

6. The recording system for a head-mounted support as described in claim 1, wherein, The intermediate bus module includes a first transmission interface and a second transmission interface. The first transmission interface includes one or more physical links carrying image encoding data that are respectively connected to the two camera modules and is configured to: (a) provide power and low-speed control / synchronization signals to each camera module; and (b) receive image encoding data from each camera module. The second transmission interface is configured to transmit the image data encapsulated by the intermediate bus module to the mobile computing device via a single host link.

7. The recording system for a head-mounted support as described in claim 1, wherein, The intermediate confluence module includes a memory configured to provide cyclic buffering to support Retro Capture, with a continuous buffering duration between 10 and 300 seconds. When Retro Capture is triggered, the intermediate confluence module encapsulates the encoded data of the buffered left and right images and transmits them to the mobile computing device via a single host link, without performing image fusion within the intermediate confluence module.

8. The recording system for a head-mounted support as described in claim 1, wherein, One or more processors of the motion computing device are configured to: estimate a depth map based on the stereo parallax between the left image and the right image, and are configured to determine the distance between a reference point of the recording system and an object in the image based on the depth map.

9. The recording system for a head-mounted support as described in claim 1, wherein, The intermediate convergence module is configured to: (a) detect the number of camera modules and image quality metrics associated with the head-mounted support; (b) write the aforementioned number and image quality metrics into synchronization relay data (e.g., camera status flags, source integrity, and timestamp consistency metrics) during encapsulation; and (c) encapsulate one or two video streams based on the detection results and transmit them to the mobile computing device via the single host link; wherein the intermediate convergence module does not perform an image fusion process.

10. The recording system for a head-mounted support as described in claim 9, wherein, The mobile computing device includes a memory, and the processor is configured to: (i) parse the number of camera modules and image quality metrics declared by the intermediate bus module in the synchronous relay data; (ii) when two are declared and the image quality metrics meet the threshold, operate in binocular mode, perform projection, local geometric alignment, and multi-band mixing to generate a single horizontal wide-angle image; (iii) when only one valid camera module is declared, or when the receiver determines that any source is interrupted, the frame timing deviation exceeds the threshold, or the alignment confidence is lower than a preset value, operate in monocular mode, skip the fusion process; and (iv) store the generated single horizontal wide-angle image and the image data of the left image and the right image as separate files; and in case (iii), retain the usable image data without outputting the single horizontal wide-angle image, and indicate the status in the relay data field.

11. An image generation method, applied to a camera system comprising a head-mounted support with two camera modules, the image generation method comprising: Step (a) Capture the left and right images, wherein the optical axes of the two camera modules are each deflected outward by an outward angle φ relative to the wearer's line of sight in the horizontal plane, which is between 10° and 30°, and the horizontal viewing angle θ of the single lens is between 65° and 95°, and the central overlap angle C = (θ - 2φ) is between 15° and 75°; Step (b) Transmit the left and right images to an intermediate bus module, which provides a common time base and encapsulation and does not perform image fusion; Step (c) Transmit the encapsulated data to a mobile computing device via a single host link; Step (d) Execute the following fusion procedure by one or more processors of the mobile computing device; Step (e) Project the left and right images onto a parametric coordinate system of a cylindrical or spherical projection model; Step (f) Perform local geometric alignment based on one or a combination of features or optical flow in the overlap area corresponding to the central overlap angle; Step (g) Generate a single horizontal wide-angle image through a fusion procedure that includes multi-band mixing.

12. The image generation method of claim 11 further includes: (i) Analyze the number of camera modules and image quality indicators declared by the intermediate convergence module in the synchronous relay data, and estimate one or a combination of source integrity and timestamp consistency at the receiver; (ii) when two valid sources are declared or estimated, perform steps (d) to (f) to generate the single horizontal wide-angle image; (iii) when only one valid source is declared or estimated, or when the fusion quality indicator is below the threshold, skip steps (d) to (f) and operate in monocular mode and output the usable one-channel image data; and (iv) store the generated single horizontal wide-angle image, the left image, and the right image as separate files; and in case (iii), retain the usable one-channel image data and indicate the status in the relay data field.

13. The image generation method of claim 11 further includes: (v) acquiring the time series of the inertial measurement unit (IMU) during the acquisition period, and additionally acquiring the actuation feedback of optical image stabilization (OIS); (vi) performing rigid body or affine-level geometric inverse compensation on each path of the image based on the data in (v) to restore the calibration geometry before projecting the left and right images onto the parameterized coordinate system of a curved projection model; and (vii) performing global stabilization on the single path of the image based on the data in (vi) after generating a single wide-angle image.

14. The image generation method of claim 11 further includes, when outputting the single horizontal wide-angle image, writing at least one of the following into a relay data field: the outer angle φ, the single-lens horizontal viewing angle θ, the activation status of EIS / OIS, the inverse compensation parameter, the stabilization parameter, or a combination thereof.

15. The image generation method as described in request item 11, wherein, The parametric coordinate system of the surface projection model in step (d) is the parametric coordinate system of cylindrical projection.

16. The image generation method as described in request item 11, wherein, The parametric coordinate system of the surface projection model in step (d) is the parametric coordinate system of the spherical projection.

17. The image generation method of claim 11 further includes: generating a depth map based on the stereo parallax between the left image and the right image, and determining the distance between the reference point of the head-mounted support and the object in the single horizontal wide-angle image based on the depth map.

18. A non-transitory computer-readable storage medium having instructions stored thereon that, when executed by a computing device, cause the computing device to perform an image generation method as described in any one of claims 11 to 17.

Images