Dual-lens-based image processing method, image processing system, and visual device

Abstract

This application relates to an image processing method, system, and visual instrument based on dual lenses. It includes: simultaneously acquiring a first image and a second image; preprocessing and extracting features from the two images to obtain a first score value and a second score value; dynamically switching between panoramic detail mode, intelligent clarity mode, low-light enhancement mode, or dual-channel independent output mode via a finite state machine based on the comparison results of the two score values ​​with two preset thresholds, while simultaneously acquiring images in real time; and merging and stitching the two images before outputting and displaying them. The system includes an image acquisition module, an image processing module, a transmission module, a display module, and a power supply module. The visual instrument uses a dual-lens assembly and a transmission module. Image quality is quantitatively evaluated through clarity scoring, information entropy, and peak signal-to-noise ratio, adaptively adjusting the processing mode to achieve simultaneous dual-lens acquisition and intelligent fusion, supporting stable dual-mode transmission, effectively improving image quality, and is suitable for clinical observation in otolaryngology, dentistry, dermatology, and other fields.

Description

Technical Field

[0001] This application relates to the technical field of image processing methods and visual instruments, and particularly to image processing methods, image processing systems and visual instruments based on dual lenses. Background Technology

[0002] Currently, medical visual examination equipment is widely used in clinical departments such as otolaryngology, dentistry, and dermatology. Existing equipment is mainly divided into two categories: single-department-specific visual instruments and general-purpose visual instruments. Single-department-specific visual instruments (such as otolaryngological endoscopes and oral endoscopes) are usually equipped with only a single lens, which can only observe specific areas, has limited functionality, and poor cross-department versatility. Although general-purpose visual instruments can be adapted to multiple departments, most of them have problems such as cumbersome lens switching and inability to simultaneously acquire and display multiple images, making it difficult to meet the needs of comparative observation of multiple areas of the same patient or multi-angle observation of complex images.

[0003] In terms of data transmission, existing devices either only support wired connections, resulting in poor portability and limiting their use in clinical mobile examination scenarios; or they only support wireless connections, which are prone to image transmission stuttering and loss in unstable signal environments, affecting diagnostic accuracy. At the same time, existing dual-lens medical devices are mostly high-precision professional instruments with complex structures, high manufacturing costs, and bulky bodies with insufficient wear and tear resistance, making it difficult to popularize their application in primary healthcare institutions.

[0004] Furthermore, existing dual-lens medical devices have significant deficiencies in image processing, lacking dynamic adaptive processing capabilities: most devices simply overlay two images, failing to automatically switch operating modes based on real-time image quality (such as sharpness, information entropy, and peak signal-to-noise ratio). When one lens image is blurry, poorly lit, or has excessive noise, the system still outputs in a fixed manner, resulting in unstable output image quality. The fusion method is simplistic: traditional dual-lens devices typically only perform simple picture-in-picture or side-by-side display, failing to achieve intelligent fusion based on feature registration (such as the organic overlay of wide-angle panoramas and telephoto details), and cannot simultaneously present a large field of view and local high-resolution details. There is no quantitative quality evaluation mechanism: the lack of real-time quantitative scoring for image sharpness, information content, and fidelity makes it difficult for operators to determine whether the current image meets observation standards and prevents the triggering of automatic optimization.

[0005] In summary, existing technologies suffer from poor versatility, limited image acquisition and display capabilities, single connection method, an imbalance between cost and practicality, incomplete image recording and transmission functions, and a lack of intelligent adaptive image processing. Therefore, there is an urgent need to provide an image processing method, image processing system, and visual instrument that can achieve simultaneous dual-lens acquisition, dynamic mode switching, intelligent fusion enhancement, stable dual-mode transmission, and a lightweight and durable structure. Summary of the Invention

[0006] The technical problems this application aims to solve are the poor versatility of existing medical visual instruments, the inability of dual lenses to intelligently synchronize and fuse, the limited connection methods, and the insufficient image adaptive processing capabilities. To address these shortcomings of the prior art, this application provides an image processing method, an image processing system, and a visual instrument based on dual lenses.

[0007] To solve the above-mentioned technical problems, the technical solution adopted in this application is: A dual-lens image processing method is provided, characterized by the following steps: The system acquires the first and second images in real time simultaneously upon startup. A first score value is obtained by preprocessing and feature extraction of the real-time first image, and a second score value is obtained by preprocessing and feature extraction of the real-time second image; The image acquisition dynamic mode is adjusted based on the first and second score values, and the first and second images are acquired in real time according to the dynamically adjusted mode. The first and second images, acquired in real time, are merged and stitched together to obtain the final merged image, which is then output and displayed.

[0008] Preferably, the process of preprocessing and feature extraction of the real-time first image to obtain a first score value, and preprocessing and feature extraction of the real-time second image to obtain a second score value, further includes: After obtaining the features of the real-time first image and the second image, the real-time first image and the second image are quantitatively evaluated based on the extracted features, and a first score and a second score are obtained based on the quantitative quality evaluation. The first score corresponds to the real-time first image, and the second score corresponds to the real-time second image.

[0009] Preferably, the quantitative quality evaluation includes a first feature, a second feature, and a third feature, and the quantitative quality evaluation is calculated from the first feature, the second feature, and the third feature and their preset weights.

[0010] Preferably, the process of dynamically adjusting the image acquisition mode based on the first and second score values, and synchronously acquiring the first and second images in real time according to the dynamically adjusted mode, further includes: The first score and the second score are compared with the first threshold and the second threshold, respectively, and the system enters different processing modes based on the comparison results.

[0011] Preferably, the process of comparing the first score and the second score with the first threshold and the second threshold further includes: The system has a built-in finite state machine, and compares the first score and the second score with the first threshold and the second threshold within a preset time of the state machine.

[0012] Preferably, the mode includes a first mode, a second mode, a third mode, and a fourth mode; When the first score is greater than the first threshold and the second score is less than the second threshold, the first mode is executed. The first mode is the panoramic detail mode, which uses Alpha fusion. When the second score is greater than the first threshold and the first score is less than the second threshold, the second mode is executed. The second mode is the intelligent clarity mode, which increases the exposure gain of the second lens and enhances the edges. When both the first score and the second score are less than the second threshold, the third mode is executed. The third mode is the low light enhancement mode, which uses a histogram equalization algorithm to adaptively improve contrast and uses bilateral filtering to denoise and preserve edges. When both the first score and the second score are greater than the first threshold, the fourth mode is executed, which is the default mode.

[0013] Preferably, in the process of fusing and stitching the first and second images acquired in real time to obtain the final fused image and outputting and displaying it, the method further includes: First, image alignment is performed to ensure that the error between the first and second images is within a set pixel range. Then, the large field of view of the first image is used as the base, and the second image is superimposed on the base and the boundary is merged to eliminate the stitching gaps to obtain the merged image.

[0014] A dual-lens image processing system is provided, characterized in that it includes an image acquisition module, an image processing module, a transmission module, a display module, and a power supply module; The image acquisition module is used to simultaneously acquire real-time images of two different locations or the same location from different angles; The image processing module is used to preprocess the two images and extract features to obtain a first score and a second score. It also performs dynamic mode switching based on the first score and the second score, and acquires the first image and the second image in real time according to the dynamically adjusted mode. Finally, it fuses the first image and the second image to obtain a fused image or a dual-channel image. The transmission module is used to transmit the processed image data to external terminal devices; The display module is used to display the fused image or dual original images in real time in split-screen or single-screen mode. The power supply module provides a stable power supply to the image acquisition module, image processing module, transmission module, and display module.

[0015] A dual-lens visual device is provided, comprising a first lens and a second lens. The first lens acquires a real-time first image, and the second lens acquires a real-time second image. The device is characterized by: the first lens acquiring the real-time first image while the second lens simultaneously acquires the real-time second image; preprocessing and feature extraction of the real-time first image to obtain a first score value; preprocessing and feature extraction of the real-time second image to obtain a second score value; dynamically adjusting the image acquisition mode based on the first and second score values; and synchronously acquiring the first and second images in real-time according to the dynamically adjusted mode; and fusing and stitching the synchronously acquired first and second images to obtain a final fused image, which is then output and displayed.

[0016] Preferably, the first and second lenses are detachably connected, and the visual instrument is used for image acquisition and observation in otolaryngology, dentistry, and dermatology.

[0017] The beneficial effects of this application are as follows: Through dual-lens synchronous acquisition and display technology, images of two different body parts or the same body part from different angles can be acquired simultaneously and output synchronously in real time, meeting the comparative observation needs of multiple departments; through dual-channel independent output mode, operators can freely choose split-screen comparison or single-screen focused observation. Based on dynamic mode switching with quality scoring, the system has a built-in finite state machine that calculates the sharpness score, information entropy, and peak signal-to-noise ratio of the two images in real time, automatically comparing them with preset thresholds, and dynamically switching to intelligent sharpness mode, panoramic detail mode, low-light enhancement mode, or dual-channel independent output mode every 5 frames, with smooth switching and uninterrupted display. Compared with existing fixed processing methods, this solution can adapt to different lighting, distance, and focus conditions, always outputting the optimal diagnostic image. Intelligent fusion enhancement provides richer observation information and offers four different modes. Image quality is quantifiable, meeting clinical observation requirements. Image quality is monitored in real time through indicators such as PSNR (≥30dB) and information entropy to ensure that the output image meets observation standards and avoids observation deviations due to image degradation.

[0018] Featuring a dual-mode wired and wireless connectivity design, this device balances the stability of wired transmission (USB-Type-C, 100Mbps) with the portability of wireless transmission (WiFi / Bluetooth dual-mode, 10m / 5m range), adapting to various clinical mobile examination scenarios. It supports image fidelity processing during transmission. The lightweight, high-drag-resistant structural design utilizes a unibody high-strength ABS engineering plastic body, measuring 14cm × 7.5cm × 2.8cm and weighing 280g. The lens is covered with scratch-resistant and wear-resistant optical glass, providing waterproof and dustproof capabilities. This simplified structure reduces costs while enhancing device durability, meeting the high-frequency usage needs of primary healthcare facilities. Operation is convenient, with comprehensive functions. The high-definition touchscreen supports split-screen / single-screen switching, touch-sensitive photography, storage, and transmission. A built-in 1500mAh lithium battery provides 8 hours of continuous use and supports charging while in use. Physical buttons on the body cater to the operating habits of different medical personnel. With strong applicability to multiple departments, it can be seamlessly adapted to the observation needs of different parts of the body, such as otolaryngology (ear canal, nasal cavity), stomatology (teeth, mucosa), and dermatology (superficial images, wounds), through manual fine adjustment of lens spacing (1-3cm), focal length adjustment (0.5cm-5cm) and multiple image processing modes, without the need to change equipment, thus reducing costs. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the present application will be further described below in conjunction with the accompanying drawings and embodiments. The drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic flowchart of an image processing method according to a preferred embodiment of this application; Figure 2 This is a schematic diagram of the system structure of the image processing system according to a preferred embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, a clear and complete description will be provided below in conjunction with the technical solutions in the embodiments of this application. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this application.

[0021] A preferred embodiment of the present invention provides an image processing method based on dual lenses; such as... Figure 1 The diagram shown is a flowchart of an image processing method based on dual lenses according to an embodiment of the present invention. The method can be executed by a device, which can be implemented by software and / or hardware.

[0022] Specifically, in this embodiment, the dual-lens image processing method includes: S1: The system synchronously acquires the first and second images in real time upon startup; In a preferred embodiment of the present invention, the system includes two sets of lenses, which can simultaneously acquire a real-time first image and a real-time second image after the system is powered on. The two sets of lenses are two independently adjustable high-definition lenses made of scratch-resistant and wear-resistant optical glass. The first lens can be a wide-angle lens or a panoramic lens, responsible for acquiring a large field of view, suitable for observing the overall layout of the image area; the second lens can be a telephoto lens or a close-up lens, responsible for acquiring high-resolution details, suitable for observing fine structures such as local areas and junctions in the image. Both the first and second lenses are high-definition CMOS lenses with a focal length adjustment range of 0.5cm-5cm, supporting autofocus, and an outer diameter of 8mm, suitable for observing and acquiring images of narrow areas, such as the ear canal and between teeth. The lens surface is covered with scratch-resistant and wear-resistant optical glass, providing waterproof and dustproof functions, allowing direct contact with skin and mucous membranes, and is easy to clean. The two lenses are symmetrically arranged at the front of the device, and the distance between them can be manually fine-tuned (adjustment range 1cm-3cm) to adapt to different observation needs.

[0023] S2: Preprocess and extract features from the real-time first image to obtain a first score value, and preprocess and extract features from the real-time second image to obtain a second score value; In a preferred embodiment of the present invention, since the real-time images acquired by the two lenses are streamed through video, the access can be via two HDMI / SDI or USB 3.0 signals. Since there are hardware differences between the first lens and the second lens, it is necessary to first standardize the real-time first image and the real-time second image to eliminate the hardware differences.

[0024] Furthermore, during the standardization process, the two image signals, the real-time first image and the real-time second image, are first decoded to separate the YUV or RGB channel data. Then, Gaussian blur or median filtering algorithms are used to perform preliminary denoising on the two images to smooth out electronic noise caused by lens differences. After standardization, sharpness features are extracted from both sets of images, and the frequency domain energy of the two images is calculated using Fourier transform or Laplace variance to obtain the wide-angle score. and telephoto rating Because higher image resolution means stronger high-frequency components (edges or textures), higher image resolution results in a higher wide-angle score. Or long-range scoring The larger the resolution feature extraction, the better. After completing the resolution feature extraction, the system further performs quantitative quality evaluation on the two images, which serves as the basis for mode switching and dynamic scheduling.

[0025] Furthermore, in the quantitative quality evaluation of images, sharpness score, image entropy, and peak signal-to-noise ratio are used. The sharpness score uses the Laplacian variance to calculate the image edge energy; a higher score indicates a sharper image, and it is used to evaluate image sharpness. Image entropy measures the amount of information in the image, and the formula for calculating image entropy is: H represents the information entropy of the image, which measures the richness or randomness of information in the image. The larger the entropy value, the more dispersed the gray-level distribution and the richer the details in the image; the smaller the entropy value, the more monotonous the image, such as a large area of ​​uniformity.

[0026] Indicates the first in the image The probability of each gray level occurring is typically, for an 8-grayscale image, the gray levels... The value range is from 0 to 255.

[0027] And satisfy If the image is in color, it can be converted to grayscale first, or the entropy of each channel can be calculated separately and then the average value can be taken.

[0028] Peak signal-to-noise ratio (PSNR) is used to evaluate the fidelity of an image after compression or transmission; ≥30dB is considered acceptable image quality.

[0029] Finally, the combined quality score of the two images is output to obtain the first score value. (Wide-angle rating) and second rating (Telephoto score): in, , , The preset weights can be set to the default value. , , ,at the same time , and The sum is 1.

[0030] S3: Adjust the dynamic mode of image acquisition based on the first and second score values, and acquire the first and second images in real time according to the dynamically adjusted mode. In a preferred embodiment of the present invention, based on the first score value of the two images... Second rating The system enters different processing modes to change the image acquisition modes of the two sets of lenses, thereby improving image quality. This results in images that retain detail while significantly improving visibility in dark areas, producing images with clear dark details and controllable noise.

[0031] Furthermore, the system incorporates a finite state machine to score the quality of the two images. and With the first threshold Second threshold The system performs a comparison and automatically switches to the corresponding working mode based on the comparison result. The specific state machine, triggering conditions, and execution modes are shown in the table below: Among them, the first threshold Second threshold Settings can be configured; default settings are available. , (Out of 100) The state machine calculates the first score every 5 frames. Second rating and with the first threshold Second threshold The system compares different execution modes to evaluate the most suitable one, avoiding screen jitter caused by frequent switching. Additional frames can be used for evaluation as needed. During state transitions, the system provides a smooth transition (e.g., a 0.2-second fade-in duration) without interrupting real-time display.

[0032] Furthermore, Mode 1 is Smart Sharpness mode. Smart Sharpness mode requires that the telephoto lens image score is low (i.e., the telephoto lens image is blurry), but the wide-angle lens image score is high (i.e., the wide-angle lens image is sharp). At this time, the system automatically locks the telephoto lens to increase the exposure gain of the telephoto lens (not exceeding +1EV) and strengthens the edges of the telephoto image through unsharp masking. The radius of the unsharp masking is 1.5 pixels and the intensity is 0.8. At the same time, the resolution of the wide-angle image is dynamically reduced to half of the original resolution to save computing power.

[0033] Furthermore, Mode 2 is the Panorama + Detail mode. This mode requires the wide-angle lens to have a high score (meaning a wide and clear field of view), while the telephoto lens has a low score due to its distance (meaning insufficient detail). In this case, the system uses the wide-angle image as the underlying background and extracts a high-resolution portion (1 / 3 of the image width) from the center of the wide-angle image, directly replacing the blurred center of the telephoto image. Alpha fusion is used when fusing the images from the two lenses, with a fusion coefficient... (Central area), with an edge feathering radius of 10 pixels, thus outputting a fused image that has both a wide field of view and a high center resolution, achieving the effect of "large field of view and clear center".

[0034] Furthermore, Mode 3 is Low-light Enhancement mode. Low-light enhancement mode requires low ambient light and significant noise in both images. In this mode, the system does not increase ISO to prevent noise explosion, but instead directly uses the Histogram Equalization (CLAHE) algorithm to adaptively enhance contrast. Simultaneously, it prioritizes the image with less noise for output. When using histogram equalization to limit the contrast, the image is first divided into 8x8 local blocks, and then histogram equalization is performed on each block to limit the maximum contrast gain (default 3.0) and prevent noise amplification. Then, denoising preprocessing is performed using a bilateral filter to preserve edges while denoising (spatial variance = 5, grayscale variance = 30). During adaptive brightness adjustment, the average brightness of the entire image is calculated. (Range 0-255), if The overall brightness is increased by 20%, and the final output image significantly improves the visibility of dark areas while maintaining the observation details, resulting in images with clear dark details and controllable noise.

[0035] Furthermore, mode four is a dual-channel independent output mode, where both images are clear and do not need to be merged. The two original images can be output directly, or after basic noise reduction and color correction, they can be displayed in split screen by the display module or manually switched by the user.

[0036] S4: Merge and stitch the first and second images acquired in real time to obtain the final merged image, and then output and display it. In a preferred embodiment of the present invention, since there is a physical distance (adjustable from 1-3cm) between the telephoto lens and the wide-angle lens, it is necessary to find feature points in the image, such as the tip of a surgical instrument, the bifurcation point of a blood vessel, etc., as a reference. The Accelerated Robust Feature (SURF) algorithm is used for image alignment, and the registration error is controlled within 2 pixels. Then, the first image and the second image are fused and stitched together. The wide field of view of the wide-angle lens is used as the base, and the high-resolution area of ​​the telephoto lens or the area processed in the aforementioned steps is superimposed on the base through Alpha fusion. For the fusion boundary, Gaussian pyramid multi-band fusion is used to eliminate the stitching seam, and finally a multifunctional image with both a wide field of view and a high-resolution close-up in the center is output. The output image format is YUV422 or RGB888, the frame rate is 30fps, and the output resolution is 1920×1080 (full HD) or 1280×720 (configurable). It is simultaneously transmitted to the display module (touch screen) and the dual-mode connection module (wired / wireless).

[0037] Furthermore, the dual-mode connection module includes a wired interface unit (USB-Type-C interface) and a wireless transmission unit (WiFi / Bluetooth dual-mode), supporting integrated data transmission and charging with a transmission rate of up to 100Mbps. The wireless transmission unit supports dual-mode switching between WiFi (802.11 b / g / n protocol) and Bluetooth (5.0 protocol), with a WiFi transmission distance of up to 10 meters and a Bluetooth transmission distance of up to 5 meters. It can connect to terminal devices such as computers, tablets, and mobile phones to achieve real-time image transmission. It can selectively achieve stable wired or convenient wireless transmission of image data and supports image fidelity processing during transmission. The display module uses a high-definition touch screen, supporting dual-screen split-screen display and single-screen switching display modes. Image capture, transmission, and storage commands can be issued via touch operation. The 3.5-inch high-definition touch screen has a resolution of 1280×720, supports split-screen display (dual screens equally divided left and right) and single-screen switching (clicking the screen to switch to full-screen display of a single lens image), with a touch response speed of ≤0.5 seconds. The screen surface is covered with an anti-fingerprint coating for easy cleaning. It also features a built-in power supply module with a 1500mAh lithium battery, supporting fast charging (fully charged in 2 hours) and a battery life of up to 8 hours (continuous operation). It charges via a USB-Type-C interface and also supports external power supply, making it suitable for long-term image acquisition and display work.

[0038] Through the above methods, this application achieves intelligent acquisition, optimization, and fusion of dual-lens images. Using dual-lens synchronous acquisition and display technology, it can simultaneously acquire images from two different locations or the same location from different angles and output them synchronously in real time, meeting the comparative observation needs of multiple departments. With dual-channel independent output mode, operators can freely choose split-screen comparison or single-screen focused observation. Based on dynamic mode switching using quality scoring, the system has a built-in finite state machine that calculates the sharpness score, information entropy, and peak signal-to-noise ratio of the two images in real time, automatically comparing them with preset thresholds. Every 5 frames, it dynamically switches to intelligent sharpness mode, panoramic detail mode, low-light enhancement mode, or dual-channel independent output mode, with smooth switching and uninterrupted display. Compared to existing fixed processing methods, this solution can adapt to different lighting, distance, and focus conditions, always outputting the optimal observation image. Intelligent fusion enhancement provides richer observation information and offers four different modes. Image quality is quantifiable, meeting clinical diagnostic requirements. Image quality is monitored in real time using indicators such as PSNR (≥30dB) and information entropy to ensure that the output image meets observation standards and avoids observational deviations due to image degradation.

[0039] Corresponding to a dual-lens-based image processing method, the present invention also provides a dual-lens-based image processing system, specifically, as follows: Figure 2 As shown, a dual-lens image processing system according to this application includes: an image acquisition module 100, an image processing module 200, a transmission module 300, a display module 400, and a power supply module 500.

[0040] Image acquisition module 100: used to simultaneously acquire real-time images of two different parts or the same part from different angles, providing raw image data for subsequent processing.

[0041] Specifically, the image acquisition module 100 includes two sets of high-definition CMOS lenses, namely a first lens and a second lens. The first lens is a wide-angle lens or a panoramic lens, responsible for acquiring images with a large field of view, suitable for observing the overall layout of the image area; the second lens is a telephoto lens or a close-up lens, responsible for acquiring high-resolution detailed images, suitable for observing local fine structures. The two sets of lenses are symmetrically arranged at the front of the camera body. The outer diameter of the lenses is 8mm, the focal length adjustment range is 0.5cm-5cm, and they support autofocus. The spacing can be manually fine-tuned (adjustment range 1cm-3cm) to adapt to the observation needs of different parts. The lens surface is covered with scratch-resistant and wear-resistant optical glass, and has waterproof and dustproof functions. After the system is powered on, the image acquisition module 100 simultaneously acquires a real-time first image and a real-time second image, and transmits the two image signals (such as through two HDMI / SDI or USB3.0 signals) to the image processing module 200.

[0042] Image processing module 200: It is used to preprocess the two images, extract features to obtain a first score and a second score, and perform dynamic mode switching based on the first score and the second score. It also acquires the first image and the second image in real time according to the dynamically adjusted mode, and fuses the first image and the second image to obtain a fused image or a dual-channel image.

[0043] Specifically, the image processing module 200 employs a low-power ARM processor and integrates a dual-channel synchronous acquisition unit, an image preprocessing unit, a quality evaluation unit, a state machine control unit, and an image enhancement and fusion unit. The dual-channel synchronous acquisition unit simultaneously receives two image signals transmitted from the image acquisition module 100 at a frame rate of 30 frames per second and separates the YUV or RGB channel data. The image preprocessing unit performs preliminary denoising on the two images using Gaussian blur or median filtering algorithms. The quality evaluation unit calculates the image edge energy using Laplacian variance to obtain a sharpness score, and then uses the information entropy formula... Calculate the image information content and peak signal-to-noise ratio (PSNR), and finally assign them to preset weights. Output the first score value Second rating The state machine control unit compares the two scores with a preset threshold (default) every 5 frames. The image enhancement and fusion unit automatically switches to intelligent clear mode, panoramic detail mode, low light enhancement mode, or dual-channel independent output mode, and controls the smooth transition. The image enhancement and fusion unit performs corresponding processing according to the current mode (such as unsharpened mask, alpha fusion, CLAHE histogram equalization, bilateral filtering, etc.), and performs image registration and Gaussian pyramid multi-band fusion based on the SURF algorithm on the two images. The final output is a fused image with a resolution of 1920×1080 or 1280×720 and a format of YUV422 or RGB888, which is then transmitted to the transmission module 300 and the display module 400.

[0044] Transmission module 300: Used to stably and faithfully transmit the processed image data to external terminal devices via wired or wireless means.

[0045] Specifically, the transmission module 300 is a dual-mode connection module, including a wired interface unit and a wireless transmission unit. The wired interface unit uses a USB-Type-C interface, supporting integrated data transmission and charging, with a transmission rate of up to 100Mbps. The wireless transmission unit supports dual-mode switching between WiFi (802.11 b / g / n protocol) and Bluetooth (5.0 protocol), with a WiFi transmission distance of up to 10 meters and a Bluetooth transmission distance of up to 5 meters, and can connect to terminal devices such as computers, tablets, and mobile phones. This module receives image data output from the image processing module 200, supports image fidelity processing during transmission, and can selectively achieve stable wired transmission or convenient wireless transmission depending on the clinical scenario.

[0046] Display module 400: Used to display the fused image or dual original images in real time in split screen or single screen, and to receive touch commands to complete operations such as image capture, transmission, and storage.

[0047] Specifically, the display module 400 uses a high-definition touchscreen, preferably a 3.5-inch screen with a resolution of 1280×720, and its surface is covered with an anti-fingerprint coating. This module supports dual-screen split-screen display (evenly divided left and right) and single-screen switching display modes (clicking the screen switches to full-screen display of a single camera view), with a touch response speed of ≤0.5 seconds. Medical personnel can issue commands for image capture, transmission, and storage via touch operation. This module also receives and displays real-time images output from the image processing module 200, and can also play back stored images.

[0048] Power supply module 500: Used to provide a stable power supply for the image acquisition module, image processing module, transmission module and display module.

[0049] Specifically, the power supply module 500 includes a built-in 1500mAh lithium battery and a power management unit, supporting fast charging (fully charged in 2 hours) and a battery life of up to 8 hours (continuous operation). This module charges via a USB-Type-C interface and also supports external power supply, providing continuous power to the image acquisition module 100, image processing module 200, transmission module 300, and display module 400. It also supports simultaneous charging and use, making it suitable for scenarios involving long-term image acquisition and motion inspection.

[0050] Based on the above embodiments, the present invention also provides a visual device, which includes two sets of lenses: a telephoto lens and a wide-angle lens. The telephoto lens and the wide-angle lens acquire two sets of images using the aforementioned image processing method, and then fuse them for display, facilitating subsequent use. The visual device is integrally molded from high-strength ABS engineering plastic with a non-slip textured surface. Its dimensions are 14cm × 7.5cm × 2.8cm, and it weighs 280g, making it easy to hold and carry. A power button, lens adjustment button, and photo button are located on the side of the device for convenient operation. To accommodate lenses used in different departments, the telephoto lens and wide-angle lens are detachably connected to the main unit. A charging dock is connected to the bottom of the main unit to charge the visual device. The visual device can be used in otolaryngology, dentistry, and dermatology to acquire images of corresponding areas for doctors to observe and analyze later.

[0051] After the device is powered on, the dual-lens assembly starts simultaneously, acquiring real-time image signals from both lenses through the dual-channel synchronous acquisition unit of the image processing module. After image optimization processing, the signals are synchronously transmitted to the display module, enabling dual-screen split-screen display. Operators can fine-tune the lens spacing and focal length according to image acquisition needs, clearly observe target areas, and capture key images using the shutter button. Images are automatically stored in the device's built-in storage (expandable up to 32GB). For data transmission, wired connection is available: connecting to the terminal device via a USB-Type-C cable to directly transfer image files; or wireless connection: pairing with the terminal device via WiFi / Bluetooth to complete wireless image transmission. Image fidelity processing is performed during transmission to ensure observation accuracy. The power supply module continuously powers all modules, supporting simultaneous charging and use to meet the needs of mobile inspection scenarios.

[0052] In ENT applications, medical staff hold the device, aligning the dual-lens assembly with both ear canals of the patient. They fine-tune the lens spacing to match the distance between the ear canals (approximately 2cm) and adjust the focus to 0.5cm-1cm. The display module shows a dual-screen view of the internal conditions of both ear canals, allowing for real-time comparison of the internal conditions and the presence of foreign objects or inflammation. If an abnormality is detected, the camera button is pressed to capture the image, which is then transmitted to a computer via WiFi for archiving in the electronic medical record system. Simultaneously, the image can be transmitted to a consultation terminal for remote expert guidance. After use, cleaning the lens surface with an alcohol swab is sufficient; the operation is convenient.

[0053] In dental applications, the distance between the two lenses is adjusted to 1cm, and they are inserted into the patient's mouth, aimed at the corresponding teeth on the upper and lower jaws respectively. The focal length is adjusted to 1cm-2cm, and the occlusal surfaces of the upper and lower teeth and the condition of the gums are observed through split-screen display. The redness and swelling of the gums on both sides or the degree of tooth decay can be quickly compared. After capturing key images, they are transmitted to the dental chair terminal via wired connection. Combined with the medical record system, an examination report is generated, and the images can be printed for the patient to view the dental images.

[0054] In dermatological applications, for superficial skin images, the dual lenses are positioned 3cm apart, aiming at the target area from both the front and side, with the focal length adjusted to 2cm-5cm. This allows for simultaneous observation of the target area's size, shape, and surrounding skin condition. Image magnification provides clear detail. After capturing images, the device connects to a mobile phone via Bluetooth and sends the images to the patient's phone, while simultaneously storing them in the hospital's information system for easy comparison during follow-up visits. The device is waterproof and dustproof, allowing for direct contact observation after cleaning, making it suitable for examining skin wounds.

[0055] It should be understood that this application has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of this application. Furthermore, based on the teachings of this application, these features and embodiments can be modified to suit specific circumstances and materials without departing from the spirit and scope of this application. Therefore, this application is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of this application.

Claims

1. A dual-lens image processing method, characterized in that, Includes the following steps: The system acquires the first and second images in real time simultaneously upon startup. A first score value is obtained by preprocessing and feature extraction of the real-time first image, and a second score value is obtained by preprocessing and feature extraction of the real-time second image; The image acquisition dynamic mode is adjusted based on the first and second score values, and the first and second images are acquired in real time according to the dynamically adjusted mode. The first and second images, acquired in real time, are merged and stitched together to obtain the final merged image, which is then output and displayed.

2. The image processing method according to claim 1, characterized in that: The process of preprocessing and feature extraction of the real-time first image to obtain a first score value, and preprocessing and feature extraction of the real-time second image to obtain a second score value, further includes: After obtaining the features of the real-time first image and the second image, the real-time first image and the second image are quantitatively evaluated based on the extracted features, and a first score and a second score are obtained based on the quantitative quality evaluation. The first score corresponds to the real-time first image, and the second score corresponds to the real-time second image.

3. The image processing method according to claim 2, characterized in that: The quantitative quality evaluation includes a first feature, a second feature, and a third feature, and the quantitative quality evaluation is calculated from the first feature, the second feature, and the third feature and their preset weights.

4. The image processing method according to claim 1, characterized in that: The process of dynamically adjusting the image acquisition mode based on the first and second score values, and synchronously acquiring the first and second images in real time according to the dynamically adjusted mode, further includes: The first score and the second score are compared with the first threshold and the second threshold, respectively, and the system enters different processing modes based on the comparison results.

5. The image processing method according to claim 4, characterized in that: The process of comparing the first and second scores with the first and second thresholds also includes: The system has a built-in finite state machine, and compares the first score and the second score with the first threshold and the second threshold within a preset time of the state machine.

6. The image processing method according to claim 4, characterized in that: The modes include a first mode, a second mode, a third mode, and a fourth mode. When the first score is greater than the first threshold and the second score is less than the second threshold, the first mode is executed. The first mode is the panoramic detail mode, which uses Alpha fusion. When the second score is greater than the first threshold and the first score is less than the second threshold, the second mode is executed. The second mode is the intelligent clarity mode, which increases the exposure gain of the second lens and enhances the edges. When both the first score and the second score are less than the second threshold, the third mode is executed. The third mode is the low light enhancement mode, which uses a histogram equalization algorithm to adaptively improve contrast and uses bilateral filtering to denoise and preserve edges. When both the first score and the second score are greater than the first threshold, the fourth mode is executed, which is the default mode.

7. The image processing method according to claim 1, characterized in that: The process of fusing and stitching the first and second images acquired in real time to obtain the final fused image and then outputting and displaying it also includes: First, image alignment is performed to ensure that the error between the first and second images is within a set pixel range. Then, the large field of view of the first image is used as the base, and the second image is superimposed on the base and the boundary is merged to eliminate the stitching gaps to obtain the merged image.

8. A dual-lens image processing system, characterized in that, It includes an image acquisition module, an image processing module, a transmission module, a display module, and a power supply module; The image acquisition module is used to simultaneously acquire real-time images of two different locations or the same location from different angles; The image processing module is used to preprocess the two images and extract features to obtain a first score and a second score. It also performs dynamic mode switching based on the first score and the second score, and acquires the first image and the second image in real time according to the dynamically adjusted mode. Finally, it fuses the first image and the second image to obtain a fused image or a dual-channel image. The transmission module is used to transmit the processed image data to external terminal devices; The display module is used to display the fused image or dual original images in real time in split-screen or single-screen mode. The power supply module provides a stable power supply to the image acquisition module, image processing module, transmission module, and display module.

9. A dual-lens visual device, comprising a first lens and a second lens, wherein the first lens acquires a real-time first image and the second lens acquires a real-time second image, characterized in that: While the first lens acquires a real-time first image, the second lens simultaneously acquires a real-time second image. The real-time first image is preprocessed and features are extracted to obtain a first score value. The real-time second image is preprocessed and features are extracted to obtain a second score value. The image acquisition dynamic mode is adjusted according to the first score value and the second score value. The first image and the second image are acquired in real-time and synchronously according to the dynamically adjusted mode. The real-time synchronously acquired first image and the second image are fused and stitched together to obtain the final fused image, which is then output and displayed.

10. The visual device according to claim 9, characterized in that: The first and second lenses are detachably connected, and the visual instrument is used for image acquisition and observation in otolaryngology, dentistry, and dermatology.

Images