CN-117395516-B - Non-homologous binocular camera capable of automatically adjusting exposure and payment equipment

CN117395516BCN 117395516 BCN117395516 BCN 117395516BCN-117395516-B

Abstract

The non-homologous binocular camera capable of automatically adjusting exposure comprises a first camera, a second camera and a processor, wherein the first camera is used for shooting a first image of a target object, the second camera is used for shooting a second image of the target object, the first camera and the second camera are non-homologous cameras, the first camera and the second camera are used for shooting at the same time, the processor is used for respectively correcting and aligning the first image and the second image, respectively detecting the first image and the second image by adopting different target detection models, respectively confirming the type and the target object area of the target object, and then performing exposure adjustment according to the target object area. The invention respectively detects target objects for the first image I ir and the second image I rgb by adopting different target detection models, automatically determines the type of the target object, evaluates the target object, and adjusts the exposure of the camera according to the image quality.

Inventors

HUANG LONGXIANG
HU TAO
SONG QI
LI ZHIBIN

Assignees

深圳市光鉴科技有限公司

Dates

Publication Date: 20260512
Application Date: 20220630

Claims (9)

1. A non-homologous binocular camera for automatically adjusting exposure, comprising: a first camera for capturing a first image of a target object; a second camera for capturing a second image of the target object; wherein the first camera and the second camera are non-homologous cameras, and the first camera and the second camera shoot simultaneously; The processor is used for respectively correcting and aligning the first image and the second image, respectively detecting the first image and the second image by adopting different target detection models, confirming the type of a target object and a target object area, and then carrying out exposure adjustment according to the target object area; The processor includes: The correction module is used for respectively carrying out distortion correction on the original first image and the original second image, and carrying out polar correction to obtain a corrected first image And a corrected second image ; A detection module for the first image And the second image Detecting target objects respectively by using different target detection models, executing a comparison module if a first target object region and a second target object region are respectively obtained if the first target object and the second target object are respectively detected on the first image and the second image, and if the first image and the second image are only used for the first image And the second image A fourth target object module is executed if a target object is detected on one image of the plurality of images, a brightness evaluation module is executed if no target object is detected, wherein the first image is a single image And the second image A non-homologous image; The comparison module is used for comparing the areas of the first target object area and the second target object area, marking the area with larger area as a third target object area and executing a fourth target object module; A fourth target object module for, when the first and second target objects are detected on the first and second images, respectively, generating a first image based on the first image And the second image Determining a fourth target object area on the image in which the larger target object is not detected if only the first image is And the second image When a target object is detected on one image of the images, a third target object area is determined on the image of the detected target object, and the third target object area is determined according to the first image And the second image Determining a fourth target object region on the image in which the target object is not detected; a brightness evaluation module for the first image And the second image Evaluating brightness, if the brightness is abnormal, resetting exposure gain is needed, if the brightness is normal, no operation is performed; an image quality module for calculating image quality of the third target object region and the fourth target object region; And the exposure adjustment module is used for adjusting the exposure of the camera according to the image quality of the third target object area and the fourth target object area.
2. The non-homologous binocular camera of claim 1, wherein the image quality module comprises: a luminance calculating unit configured to calculate average luminance of the third target object region and the fourth target object region; A definition calculating unit configured to calculate image definition of the third target object region and the fourth target object region; and the output unit is used for outputting an image if the average brightness and the image definition are both in a reasonable range, and executing an exposure adjustment module if not.
3. The non-homologous binocular camera of claim 1, wherein the brightness assessment module, for the first image And the second image When the brightness is evaluated, only a preset range including the center of the image is evaluated.
4. The non-homologous binocular camera of claim 1, wherein the detection module determines the target object area if a target object is detected and determines that a target object is not detected if the target object area is less than a threshold.
5. The non-homologous binocular camera of claim 1, wherein the brightness assessment module, for the first image And the second image When evaluating the brightness, the first image is processed And the second image The subtraction and use of the new image generated for evaluation.
6. The non-homologous binocular camera of claim 1, wherein the exposure adjustment module comprises: the brightness adjusting unit is used for increasing exposure time, current value or gain value if the average brightness of the third target object area and the fourth target object area is lower, and reducing exposure time, current value or gain value if the average brightness of the third target object area and the fourth target object area is higher; and the definition adjusting unit is used for increasing the exposure time, increasing the current value or reducing the gain value if the definition of the target object is lower.
7. The non-homologous binocular camera automatically adjusting exposure according to claim 6, wherein in the brightness adjustment unit: if the average brightness of the third target object area and the fourth target object area is lower, the adjusting sequence is exposure time, gain value and current value; If the average brightness of the third target object area and the fourth target object area is higher, the adjusting sequence is a current value, a gain value and an exposure time.
8. The non-homogeneous binocular camera of claim 1, wherein a plurality of different object detection models are employed in the detection module, respectively, in the first image And the second image The detection is not repeated in this order.
9. A payment device comprising a non-homologous binocular camera of any of claims 1 to 8 which automatically adjusts exposure.

Description

Non-homologous binocular camera capable of automatically adjusting exposure and payment equipment Technical Field The invention relates to the field of depth cameras, in particular to a non-homologous binocular camera capable of automatically adjusting exposure and payment equipment Background The binocular technique is one of the mainstream techniques for obtaining depth data. The binocular technology utilizes the parallax principle to calculate two images and combines the internal participation and external participation of a depth camera, so that the depth information of a target object can be calculated. The depth data obtained by the binocular technology can be used for three-dimensional reconstruction, so that the depth camera has important application value in the fields of biological recognition and the like. In the prior art, a homologous binocular technique is generally adopted to obtain better target object information and perform matching. The two images obtained by the non-homologous binocular technique are different, so that the two images have differences in various aspects such as definition, feature extraction and the like. When the automatic exposure adjustment is carried out on the non-homologous binocular camera, a general automatic exposure adjustment strategy is adopted, and the automatic exposure adjustment is not carried out by combining the characteristics of the non-homologous binocular camera and the characteristics of the target object, so that the target object is unclear or the automatic exposure efficiency is low. In the existing target recognition technology, the target type is usually required to be determined in advance, so that target object recognition is performed. For example, a common payment device requires a user to select a payment type and then identify the payment type, which causes the problems of more operation steps and complicated process of the user. Disclosure of Invention Therefore, the invention respectively detects the target object by adopting different target detection models for the first image I ir and the second image I rgb, automatically determines the type of the target object, evaluates the target object, adjusts the exposure of the camera according to the image quality, combines the characteristics of non-homologous binocular images, can quickly identify the type of the target object and automatically expose the target object by utilizing the characteristics of the target object under the condition that the specific type of the target object is not determined, reduces the image processing amount, and greatly improves the accuracy of automatic exposure. In a first aspect, the present invention provides a non-homologous binocular camera for automatically adjusting exposure, comprising: a first camera for capturing a first image of a target object; a second camera for capturing a second image of the target object; wherein the first camera and the second camera are non-homologous cameras, and the first camera and the second camera shoot simultaneously; And the processor is used for respectively correcting and aligning the first image and the second image, respectively detecting the first image and the second image by adopting different target detection models, confirming the type of a target object and a target object area, and then carrying out exposure adjustment according to the target object area. Optionally, the non-homologous binocular camera capable of automatically adjusting exposure, wherein the processor comprises: The correction module is used for respectively carrying out distortion correction on the original first image and the original second image, and carrying out polar correction to obtain a corrected first image I ir and a corrected second image I rgb; A detection module for detecting target objects respectively on the first image I ir and the second image I rgb by using different target detection models, a comparison module for executing a comparison module if a first target object area and a second target object area are respectively obtained by detecting the first target object area and the second target object area on the first image and the second image respectively, a fourth target object module for executing a fourth target object module if the target object is detected on only one image of the first image I ir and the second image I rgb, and a brightness evaluation module for executing a brightness evaluation module if the target object is not detected, wherein the first image and the second image are non-homologous images; The comparison module is used for comparing the areas of the first target object area and the second target object area, marking the area with larger area as a third target object area and executing a fourth target object module; A fourth target object module, configured to determine a third target object area on an image in which a target object is detected, and determine a fourth target object area on a