A high-definition fluorescence endoscopic imaging system

Abstract

This application relates to the field of imaging technology and provides a high-definition fluorescence endoscopic imaging system. Through the close collaboration and organic integration of a composite light source module, an optical imaging module, an image processing module, and a control feedback module, the imaging system can acquire high-resolution, high-contrast, color-accurate, and detail-rich images in various complex scenarios, providing doctors with clearer and more accurate tissue information, which helps in the early detection and accurate diagnosis of diseases. The information interaction and collaborative control between the various modules of the imaging system realizes the intelligent and automated process from light source excitation, image acquisition, processing to feedback control. The intelligent processing of the imaging system also reduces the impact of human factors on the imaging results and improves the consistency and repeatability of imaging.

Description

Technical Field

[0001] This application relates to the field of imaging technology, and in particular to a high-definition fluorescence endoscopic imaging system. Background Technology

[0002] Traditional endoscopic light sources cannot flexibly adjust wavelength combinations according to different imaging modes, making it difficult to accurately adapt to the imaging needs of different target tissues. This results in poor irradiation of the target tissue by the excitation light and insufficient or unsatisfactory fluorescence signal generation. Moreover, the integration time of the sensor is usually fixed and cannot be dynamically adjusted according to the actual situation of the fluorescence image. This makes it impossible to adapt to fluorescence signals of different intensities and distributions, reducing the adaptability of the imaging system. Existing endoscopic systems may have problems with low registration accuracy when registering white light images and fluorescence images, resulting in poor image fusion and an inability to clearly display the morphology and fluorescence characteristics of the target tissue. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this application provides a high-definition fluorescence endoscope imaging system, which includes: a composite light source module, an optical imaging module, an image processing module, and a control feedback module;

[0004] The composite light source module is used to select wavelength combinations according to the imaging mode, identify the edge of the target tissue to determine the shape of the excitation light spot and generate excitation light to irradiate the target tissue to generate fluorescence signals, and dynamically adjust the power of the excitation light based on the feedback signal.

[0005] The optical imaging module is used to collect mixed light signals through an endoscope probe, separate the mixed light signals into white light signals and fluorescence signals, convert the light signals into electrical signals through a sensor, and simultaneously output white light images and fluorescence images. The integration time of the sensor is dynamically adjusted based on the feedback signal.

[0006] The image processing module is used to register the white light image and the fluorescence image based on matching feature point pairs, and to fuse the registered white light image and the fluorescence image in real time based on the boundary of the target tissue to generate an overlay image.

[0007] The control feedback module is used to extract fluorescence images to generate feedback signals, monitor the power of the excitation light and the temperature of the target tissue in real time, and determine whether to trigger an early warning process.

[0008] As an optional implementation, the excitation light generation logic includes:

[0009] Pre-scan the target tissue to identify its edges and determine the shape of the excitation light spot;

[0010] The wavelength combination is invoked according to the imaging mode, and the timing driving parameters of the wavelength combination are loaded synchronously.

[0011] A driving signal is sent to the laser source according to the wavelength combination, and the laser source is driven to emit pulse-modulated excitation light by combining the timing driving parameters of the wavelength combination.

[0012] The pulsed excitation light is transmitted to the endoscope probe and the target tissue is irradiated according to the shape of the excitation light spot to generate a fluorescence signal.

[0013] As an optional implementation, the edge recognition logic of the target organization includes:

[0014] A fluorescence image is obtained by pre-scanning the target tissue, and then the fluorescence image is processed into a grayscale image to obtain a fluorescence grayscale image;

[0015] Calculate the gradient magnitude and gradient direction of each pixel in the fluorescence grayscale image, identify edge pixels, and connect the edge pixels to obtain the initial target tissue edge;

[0016] Morphological processing is performed on the initial target tissue edges to obtain the target tissue edges.

[0017] As an optional implementation, the logic for generating the feedback signal includes:

[0018] Receive fluorescence images and update the location region of the target tissue based on a target tracking algorithm;

[0019] The average pixel intensity and pixel intensity distribution of pixels in the location region of the target tissue are obtained by pixel traversal, and the fluorescence intensity and distribution uniformity of the fluorescence image are obtained.

[0020] The feedback signal is determined based on the fluorescence intensity and distribution uniformity of the fluorescence image.

[0021] As an optional implementation, the logic for dynamically adjusting the integration time of the sensor includes:

[0022] Extract fluorescence intensity, distribution uniformity, and target tissue type from fluorescence images;

[0023] The integral time adjustment is calculated based on fluorescence intensity, distribution uniformity, and target tissue type, and an adjustment signal is generated.

[0024] Adjust the sensor's integration time based on the adjustment signal and verify the adjustment effect.

[0025] As an optional implementation, the registration process logic includes:

[0026] Structural feature points are extracted from white light images and the edges of target tissues in fluorescence images. The structural feature points are then coarsely matched with the edges of target tissues to obtain matching feature point pairs.

[0027] Based on the matching feature point pairs, the fluorescence image is mapped to the coordinate system of the white light image;

[0028] The transformed fluorescence image is resampled and spatially aligned using bilinear interpolation to obtain the registered white light image and fluorescence image.

[0029] As an optional implementation, the coarse matching logic includes:

[0030] Calculate the distance between structural feature points and the edge of the target tissue, and configure an error threshold to remove mismatched points;

[0031] The weights of matching feature points are dynamically adjusted based on the uniformity of distribution.

[0032] The number of iterations and error threshold for coarse matching are optimized based on the type of the target organization.

[0033] As an optional implementation, the logic for generating the overlay image includes:

[0034] The registered white light image and fluorescence image are fused based on the boundary of the target tissue.

[0035] Perform color adjustments and optimizations on the merged image;

[0036] The color-adjusted and optimized image is output as the final overlay image.

[0037] As an optional implementation, the logic for fusion based on the boundary of the target organization includes:

[0038] The boundaries of the target tissue were extracted from the registered white light image and fluorescence image, respectively;

[0039] The fusion weights of white light and fluorescence images are determined based on the type and distribution uniformity of the target tissue.

[0040] Based on the fusion weights, the white light image and the fluorescence image are fused pixel by pixel to obtain the fused image.

[0041] As an optional implementation, the triggering logic for the early warning processing includes:

[0042] Real-time monitoring of excitation light power and target tissue temperature;

[0043] The power of the excitation light and the temperature of the target tissue are compared with the configured power threshold and temperature threshold, respectively, to determine whether they exceed the limits.

[0044] If the limit is exceeded, an early warning will be triggered. The warning level will be determined and a response will be initiated based on the degree and duration of the exceedance.

[0045] Compared with existing technologies, the beneficial effects of this application are as follows: Through the close collaboration and organic integration of the composite light source module, optical imaging module, image processing module, and control feedback module, the imaging system can acquire high-resolution, high-contrast, color-accurate, and detail-rich images in various complex scenarios, providing doctors with clearer and more accurate tissue information, which helps in the early detection and accurate diagnosis of diseases; through the collaborative processing of each module, the imaging system can more accurately capture the morphological and fluorescence characteristics of the target tissue, reduce image interference and errors, further improve the reliability of diagnostic results, help doctors make more accurate diagnostic decisions, and provide patients with more effective treatment plans; the information interaction and collaborative control between the modules of the imaging system realizes the intelligent and automated process from light source excitation, image acquisition, processing to feedback control. The intelligent processing of the imaging system also reduces the impact of human factors on imaging results and improves the consistency and repeatability of imaging.

[0046] The composite light source module can select wavelength combinations according to the imaging mode and identify the edge of the target tissue to determine the shape of the excitation light spot. This enables precise irradiation of the target tissue, effectively reduces interference from irrelevant tissues, and improves the generation efficiency and quality of fluorescence signals. At the same time, the function of dynamically adjusting the excitation light power ensures that the excitation light is always in the optimal state, guaranteeing the stability and intensity of the fluorescence signal.

[0047] The optical imaging module can effectively separate white light signals and fluorescence signals, and can accurately convert light signals into electrical signals through sensors, simultaneously outputting high-quality white light images and fluorescence images. The function of dynamically adjusting the sensor integration time enables the system to adaptively optimize the imaging effect according to different fluorescence intensities, distribution uniformity and target tissue types, and obtain clear and accurate images in various complex imaging scenarios.

[0048] The image processing module, through precise registration and fusion logic, can accurately register and fuse white light images and fluorescence images based on matching feature point pairs. The resulting overlay image can more comprehensively and clearly display the morphology and fluorescence characteristics of the target tissue. Color adjustment and optimization further enhance the visual effect of the image, helping doctors to observe and diagnose more accurately.

[0049] The control feedback module can monitor the power of the excitation light and the temperature of the target tissue in real time. When the power or temperature exceeds the standard, it can trigger an early warning process in a timely manner. The warning level is determined and responded according to the degree and duration of the exceedance, which improves the safety and reliability of the system, effectively protects the target tissue, and reduces potential risks. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, 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. Wherein:

[0051] Figure 1 A system flowchart of a high-definition fluorescence endoscopic imaging system provided in this application embodiment;

[0052] Figure 2 An edge recognition logic diagram of a target tissue provided in an embodiment of this application for a high-definition fluorescence endoscopic imaging system;

[0053] Figure 3 This application provides a registration processing logic diagram for a high-definition fluorescence endoscopic imaging system.

[0054] Figure 4 This is a coarse matching logic diagram of a high-definition fluorescence endoscopic imaging system provided in an embodiment of this application. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of this application more apparent and understandable, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0056] Example

[0057] like Figure 1 The diagram shown is a system flowchart of a high-definition fluorescence endoscopic imaging system provided in this application embodiment. The system includes a composite light source module, an optical imaging module, an image processing module, and a control feedback module.

[0058] The composite light source module is used to select wavelength combinations according to the imaging mode, generate excitation light, irradiate the target tissue to produce fluorescence signals, and dynamically adjust the power of the excitation light based on the feedback signal.

[0059] Specifically, the logic for generating excitation light includes:

[0060] Pre-scan the target tissue to identify its edges and determine the shape of the excitation light spot;

[0061] The wavelength combination is invoked according to the imaging mode, and the timing driving parameters of the wavelength combination are loaded synchronously.

[0062] A driving signal is sent to the laser source according to the wavelength combination, and the laser source is driven to emit pulse-modulated excitation light by combining the timing driving parameters of the wavelength combination.

[0063] The pulsed excitation light is transmitted to the endoscope probe and the target tissue is irradiated according to the shape of the excitation light spot to generate a fluorescence signal.

[0064] Furthermore, such as Figure 2 As shown, the edge detection logic of the target organization includes:

[0065] A fluorescence image is obtained by pre-scanning the target tissue, and then the fluorescence image is processed into a grayscale image to obtain a fluorescence grayscale image;

[0066] Calculate the gradient magnitude and gradient direction of each pixel in the fluorescence grayscale image, identify edge pixels, and connect the edge pixels to obtain the initial target tissue edge;

[0067] Morphological processing is performed on the initial target tissue edges to obtain the target tissue edges.

[0068] Different target tissues have different morphologies. In order to accurately excite fluorescence signals and improve imaging quality and efficiency, it is necessary to adjust the shape of the excitation light spot according to the shape of the target tissue so that the excitation light can be concentrated and uniformly irradiated on the target tissue. After the imaging system is started, the control feedback module sends a pre-scan command to the endoscope probe. The composite light source module receives the pre-scan command and drives the endoscope probe to emit low-power probe light to perform a preliminary scan of the target tissue. At this time, the optical imaging module converts the received reflected light signal into an image and transmits it to the image processing module.

[0069] Before performing edge recognition, the image processing module needs to convert the color fluorescence image into a fluorescence grayscale image so that the subsequent edge detection algorithm can process the fluorescence grayscale image more effectively. This is because most edge detection algorithms are based on grayscale images. In the image processing module, a weighted average method is used to convert the color fluorescence image into a grayscale image, thereby simplifying the image data volume, removing the interference of color information on edge detection, and improving the computational efficiency and accuracy of the edge detection algorithm. The resulting fluorescence grayscale image provides a suitable input image for subsequent edge detection, enabling the edge detection algorithm to more accurately calculate the gradient magnitude and direction of each pixel, which helps to identify edge pixels.

[0070] Edge pixels in a fluorescence grayscale image exhibit abrupt changes in grayscale value. By calculating the gradient magnitude and direction of each pixel, edge pixels can be identified, and these edge pixels can be connected to obtain the initial edge of the target tissue. The Canny algorithm is used for edge detection. First, the fluorescence grayscale image is smoothed using a Gaussian filter to reduce the impact of noise on gradient calculation. Then, the gradient magnitude and direction of each pixel are calculated. For example, the Sobel operator is used to calculate the gradient components in the x and y directions, and the gradient magnitude and direction are calculated using the corresponding formulas. Next, non-maximum suppression is performed, retaining the pixel with the largest gradient magnitude as an edge pixel and suppressing other non-edge pixels. Finally, strong and weak edges are identified through dual threshold detection, and weak edges are connected to strong edges using an edge tracking algorithm to obtain the initial edge of the target tissue. This allows for accurate identification of edge pixels in the target tissue and connection of them into continuous edges, providing a foundation for subsequent morphological processing.

[0071] The initial target tissue edges may contain noise, burrs, and discontinuities. Morphological processing can remove these defects, making the target tissue edges smoother, more continuous, and more accurate. Erosion removes small burrs and isolated noise points on the edges. Erosion involves sliding a structuring element (such as a 3×3 square matrix) across the fluorescent grayscale image; if all pixels within the area covered by the structuring element are edge pixels, the center pixel is retained as an edge pixel; otherwise, it is removed. Then, dilation connects the broken edges. Dilation involves sliding a structuring element across the fluorescent grayscale image; if at least one pixel within the area covered by the structuring element is an edge pixel... The center pixel is set as the edge pixel. By performing alternating erosion and dilation operations multiple times, a smooth, continuous, and accurate edge of the target tissue is obtained, thereby optimizing the edge of the target tissue, improving the quality and accuracy of the edge, and providing a more reliable basis for determining the shape of the excitation light spot. Accurately identifying the edge of the target tissue provides a precise contour for determining the shape of the excitation light spot, enabling the excitation light to more accurately irradiate the target tissue, improving the efficiency of fluorescence excitation and imaging quality, reducing unnecessary energy loss, and making the subsequently acquired fluorescence signal clearer and more accurate. This helps to generate high-quality fluorescence signals and lays a good foundation for subsequent imaging and signal processing.

[0072] Different imaging modes are suitable for different diagnostic needs and target tissue characteristics. Each imaging mode requires a specific wavelength combination to achieve the best fluorescence excitation effect. At the same time, in order to accurately control the emission of the laser source, the corresponding timing drive parameters need to be loaded synchronously. The imaging system has a built-in database of multiple imaging modes, such as ordinary fluorescence imaging mode and specific tissue fluorescence imaging mode. When the user selects an imaging mode, the composite light source module retrieves the wavelength combination information corresponding to that mode from the database. For example, for ordinary fluorescence imaging mode, the wavelength combination of 405nm and 488nm is selected, and the timing drive parameters corresponding to that wavelength combination are loaded. These timing drive parameters include pulse frequency, duty cycle, and pulse width, which are used to accurately control the emission timing of the laser source. This allows for the rapid and accurate configuration of light source parameters according to different imaging needs, achieving adaptability to various target tissues and diagnostic scenarios, and improving the flexibility and practicality of the imaging system.

[0073] To irradiate target tissue and generate fluorescence signals, a laser source needs to be driven according to specific wavelength combinations and timing parameters to emit excitation light with a specific frequency, duty cycle, and pulse width, thereby achieving efficient fluorescence excitation. The control feedback module sends a driving signal to the corresponding laser source based on the determined wavelength combination, and simultaneously transmits the loaded timing parameters to the laser source driving circuit in the composite light source module. The driving circuit pulses the laser source according to these timing parameters, adjusting the pulse frequency, duty cycle, and pulse width to make the laser source emit pulse-modulated excitation light that meets the requirements, thus generating excitation light with specific wavelength and modulation characteristics, effectively exciting the target tissue to generate fluorescence signals, and improving the quality and contrast of fluorescence imaging.

[0074] Only by accurately transmitting the excitation light to the target tissue and irradiating it according to the previously determined excitation light spot shape can we ensure that the target tissue is uniformly excited and generates a high-quality fluorescence signal. The pulsed excitation light is transmitted to the endoscope probe, the tip of which is equipped with special optical elements, such as deformable microlens arrays or aperture devices. According to the previously determined excitation light spot shape, the deformation or opening and closing state of these optical elements is controlled so that the excitation light irradiates the target tissue according to the predetermined spot shape. For example, for a circular spot, the curvature of the microlens array is adjusted to focus the excitation light into a circle that irradiates the target tissue. This allows for precise control of the irradiation range and shape of the excitation light, ensuring that the target tissue is uniformly excited, improving the intensity and uniformity of the fluorescence signal, thereby enhancing the clarity and accuracy of the imaging and providing reliable data support for subsequent image processing and diagnosis of the target tissue.

[0075] Specifically, the logic for generating the feedback signal includes:

[0076] Receive fluorescence images and update the location region of the target tissue based on a target tracking algorithm;

[0077] The average pixel intensity and pixel intensity distribution of pixels in the location region of the target tissue are obtained by pixel traversal, and the fluorescence intensity and distribution uniformity of the fluorescence image are obtained.

[0078] The feedback signal is determined based on the fluorescence intensity and distribution uniformity of the fluorescence image.

[0079] During imaging, the target tissue may move or deform. To accurately analyze the characteristics of the fluorescence image, the location region of the target tissue needs to be updated in real time so that the fluorescence intensity and distribution uniformity can be calculated accurately. The control feedback module receives real-time fluorescence images from the optical imaging module. Through the target tracking algorithm, it predicts the position of the target tissue in the current frame based on the position and motion state of the target tissue in the previous frame image. Then, through template matching, it finds the actual position of the target tissue in the fluorescence image of the current frame and updates the location region of the target tissue. This allows for real-time tracking of the position changes of the target tissue, ensuring that the analysis of the target tissue is always based on accurate position information and improving the accuracy of fluorescence intensity and distribution uniformity calculations.

[0080] Template matching determines the location region of the target tissue by calculating the correlation between the template image T and the current frame fluorescence image I. The functional expression for the correlation between the template image T and the current frame fluorescence image I is shown below:

[0081]

[0082] In the formula, R(x,y) represents the correlation value of the pixel at position (x,y), I(x+i,y+j) represents the pixel value of the pixel at position (x+i,y+j) in the current frame fluorescence image I, T(i,j) represents the pixel value of the pixel at position (i,j) in the template image T, and μ I μ represents the mean pixel value in the current frame fluorescence image I. T This represents the mean value of the pixels in the template image T.

[0083] It should be noted that when the correlation value R(x,y) of the pixel at position (x,y) is close to 1, it indicates that the matching degree is very high, and the actual position of the target tissue is at position (x,y).

[0084] Fluorescence intensity and distribution uniformity are important indicators for evaluating the quality of fluorescence imaging and the state of target tissue. Calculating these parameters provides a basis for adjusting the excitation light power to optimize imaging results. Within the defined target tissue location area, the grayscale value of each pixel is read sequentially by pixel traversal. For fluorescence images, the grayscale value reflects the fluorescence intensity. The average grayscale value of all pixels is calculated to obtain the average pixel intensity of the target tissue, i.e., the fluorescence intensity. At the same time, the number of pixels with different grayscale values ​​is counted, and a pixel intensity distribution histogram is plotted to obtain the pixel intensity distribution of the fluorescence image, i.e., the distribution uniformity. This accurately obtains the fluorescence intensity and distribution uniformity of the fluorescence image, providing quantitative data support for the generation of subsequent feedback signals.

[0085] To dynamically adjust the excitation light power based on the actual fluorescence image, feedback signals are generated based on fluorescence intensity and distribution uniformity. These signals feed image feature information back to the composite light source module to optimize excitation light emission. The calculated fluorescence intensity and distribution uniformity are compared with preset intensity and uniformity threshold ranges. If both fluorescence intensity and distribution uniformity are below the lower limit of the intensity threshold and the lower limit of the uniformity threshold, it indicates insufficient excitation light power, and a feedback signal is generated to increase the excitation light power. If fluorescence intensity is above the upper limit of the intensity threshold and distribution uniformity is above the lower limit of the uniformity threshold, the excitation light power is insufficient, and a feedback signal is generated to increase the excitation light power. When the value reaches its upper limit, it indicates that the excitation light power is too high and fails to meet the standard. A feedback signal is generated to indicate that the excitation light power should be reduced. The feedback signal can be in the form of digital encoding, clearly containing the direction and amount of power adjustment. This enables dynamic adjustment of the excitation light power, allowing the imaging system to automatically optimize the emission of the excitation light according to the actual fluorescence characteristics of the target tissue, thereby improving the imaging quality. The feedback signal is transmitted to the composite light source module, which adjusts the excitation light power according to the feedback signal, thereby changing the fluorescence excitation effect and affecting the quality of the subsequently acquired fluorescence image, forming a closed-loop adaptive adjustment system.

[0086] The feedback signal in the control feedback module is identified, and its information is extracted. The power adjustment is calculated based on fluorescence intensity, distribution uniformity, and the type of target tissue. The type of target tissue is obtained through a target tissue recognition model built on a convolutional neural network. This model is trained using a large dataset of labeled fluorescence and white light images of different target tissue types. During training, the model learns the feature representations and classification rules for images of different target tissue types. For real-time acquired white light and fluorescence images, the images are input into the trained target tissue recognition model, which outputs a prediction of the target tissue type. The reliability of the prediction is evaluated using the classification probability of the target tissue recognition model. The formula for calculating the power adjustment is shown below:

[0087]

[0088] In the formula, ΔP represents the power adjustment amount, k represents the adjustment influence coefficient, α represents the power increase magnitude coefficient, and F min F represents the lower limit of the intensity threshold, F represents fluorescence intensity, and U represents distribution uniformity. min F represents the lower limit of the uniformity threshold, β represents the magnitude coefficient of power reduction, and F max U represents the upper limit of the intensity threshold. max This represents the upper limit of the uniformity threshold.

[0089] It should be noted that: when the fluorescence intensity is less than the lower limit of the intensity threshold and the distribution uniformity is less than the lower limit of the uniformity threshold, the excitation light power needs to be increased, and the value of the power increase coefficient α ranges from 0 to 1; when the fluorescence intensity is greater than the upper limit of the intensity threshold and the distribution uniformity is greater than the upper limit of the uniformity threshold, the excitation light power needs to be decreased, and the value of the power decrease coefficient β ranges from 0 to 1; the adjustment influence coefficient k is a coefficient determined according to the influence of the target tissue type on the power adjustment. For example, for target tissues that are more sensitive to fluorescence excitation, k is taken as a larger value, so that the power adjustment range is larger, while for target tissues that are relatively insensitive to fluorescence excitation, k is taken as a smaller value, so as to adjust the power adjustment range under different target tissue types; when neither of the above two conditions is met, that is, when the fluorescence intensity and distribution uniformity are within the range of the intensity threshold and the uniformity threshold respectively, ΔP = 0, that is, no adjustment of the excitation light power is required.

[0090] The driving current is adjusted according to the calculated power adjustment amount to dynamically adjust the power of the excitation light, and the shape of the excitation light spot is adjusted in real time according to the edge of the target tissue. Then, the target tissue is re-irradiated according to the adjusted excitation light power to generate a fluorescence signal, thereby generating a new fluorescence image. The fluorescence intensity and distribution uniformity of the fluorescence image are verified again by the image processing module and the control feedback module. If they are not up to standard, the above steps are repeated to dynamically adjust the power of the excitation light. If they are up to standard, there is no need to dynamically adjust the power of the excitation light.

[0091] The optical imaging module is used to collect mixed light signals through an endoscope probe, separate white light signals and fluorescence signals based on a switchable filter, convert the light signals into electrical signals through a sensor, output white light images and fluorescence images simultaneously, and dynamically adjust the integration time of the sensor based on feedback signals.

[0092] The mixed light signal is collected by an endoscopic probe. The mixed light signal includes the white light signal reflected back from the excitation light and the fluorescence signal generated by the target tissue. This ensures the comprehensiveness and accuracy of the light signal collection and provides sufficient data for subsequent signal processing. The mixed light signal is initially separated into two paths according to wavelength range by a beam splitter, including a short wavelength light path and a long wavelength light path. For example, the white light signal (with a wide wavelength range) and the fluorescence signal (with a relatively concentrated wavelength) are initially separated. The complex mixed light signal is initially separated, which simplifies the work of subsequent switchable filters and improves the overall separation efficiency.

[0093] Different imaging modes have different requirements for light signals. By controlling the position of the switchable filter and selecting an appropriate filter, precise separation of white light and fluorescence signals can be achieved to meet different imaging needs. The position of the switchable filter is controlled according to the selected imaging mode. A broadband filter is inserted in the white light channel to allow only the white light signal reflected back from the excitation light to be transmitted, while effectively blocking the interference of the fluorescence signal. A narrowband filter is inserted in the fluorescence channel to allow only the fluorescence signal generated by the excitation of the target tissue to be transmitted, while effectively blocking other wavelengths of light, including white light signals and fluorescence signals from non-target tissues. This completes the separation of mixed light signals. The separated white light signal and fluorescence signal enter independent optical paths, namely the white light channel and the fluorescence channel, respectively, ensuring the purity of the white light signal and fluorescence signal and providing a guarantee for subsequent high-quality imaging.

[0094] After being separated by a switchable filter, the white light and fluorescence signals are guided to the sensor. Before reaching the sensor, the white light and fluorescence signals are focused to ensure that they accurately illuminate the photosensitive area of ​​the sensor, reducing signal loss during transmission and improving transmission efficiency. The white light signal is captured by a CMOS sensor in the white light channel, and the fluorescence signal is captured by a CMOS sensor in the fluorescence channel, thus converting them into electrical signals. These signals are then amplified to improve the signal-to-noise ratio, with the amplifier gain dynamically adjusted based on the fluorescence intensity. Preliminary noise suppression and signal shaping are also performed to improve the quality and stability of the electrical signals, providing high-quality electrical signals for subsequent analog-to-digital conversion and image reconstruction.

[0095] The amplified and pre-processed electrical signal is transmitted to an analog-to-digital converter (ADC) to convert the analog signal into a digital signal. The ADC then discretizes the signal according to a preset sampling precision and frequency, converting it into digital code. This code enables the computer system to perform subsequent digital signal processing and image reconstruction, outputting white light and fluorescence images via digital-to-analog conversion. The digitized white light and fluorescence images are first stored in a buffer area within the sensor for further processing and transmission. This buffer utilizes high-speed storage technology to ensure rapid writing and reading of the images, meeting the requirements of real-time imaging. A synchronization signal ensures the output time alignment of the white light and fluorescence images, eliminating timing deviations caused by filter switching. The buffered white light and fluorescence images are then transmitted to the image processing module via a data bus, thus realizing the conversion from analog to digital signals. This provides digital data for subsequent digital signal processing and image reconstruction, ensuring the synchronous output and transmission of the white light and fluorescence images.

[0096] Specifically, the logic for dynamically adjusting the sensor's integration time includes:

[0097] Extract fluorescence intensity, distribution uniformity, and target tissue type from fluorescence images;

[0098] The integral time adjustment is calculated based on fluorescence intensity, distribution uniformity, and target tissue type, and an adjustment signal is generated.

[0099] Adjust the sensor's integration time based on the adjustment signal and verify the adjustment effect.

[0100] Fluorescence intensity, distribution uniformity, and target tissue type are key factors affecting image quality. Extracting this information provides a basis for reasonably adjusting the sensor integration time to adapt to different imaging scenarios and target tissue characteristics. Fluorescence intensity and distribution uniformity are determined according to the method in the feedback signal generation logic. Based on a pre-trained target tissue recognition model, such as a deep learning-based convolutional neural network model, the white light image is analyzed to identify the type of target tissue, thereby accurately obtaining the key parameters affecting image quality and providing quantitative data support for subsequent integration time adjustment.

[0101] Different fluorescence intensities, distribution uniformity, and target tissue types require different sensor integration times to optimize imaging results. By calculating the integration time adjustment and generating an adjustment signal, precise control of the sensor's integration time can be achieved. The formula for calculating the integration time adjustment is shown below:

[0102]

[0103] In the formula, ΔT represents the adjustment amount of the integration time, λ1 represents the weight of increasing intensity, λ2 represents the weight of increasing uniformly, λ3 represents the weight of decreasing intensity, and λ4 represents the weight of decreasing uniformly.

[0104] It should be noted that the values ​​of λ1, λ2, λ3, and λ4 are between 0 and 1. The adjustment weights measure the relative importance of fluorescence intensity and distribution uniformity in calculating the integration time adjustment. When the fluorescence intensity is below the lower limit of the intensity threshold and the distribution uniformity is below the lower limit of the uniformity threshold, the integration time needs to be increased. The intensity increase weight λ1 refers to the weight of the influence of fluorescence intensity on the integration time adjustment when the integration time is increased, and the uniformity increase weight λ2 refers to the weight of the influence of distribution uniformity on the integration time adjustment when the integration time is increased. The values ​​of λ1 and λ2 differ for different target tissue types. For example, a larger λ1 value means that for that type of target tissue, insufficient fluorescence intensity has a more critical impact on the integration time adjustment, requiring... Prioritize increasing the integration time of the sensor to improve fluorescence intensity, with λ1+λ2=1. When the fluorescence intensity exceeds the upper limit of the intensity threshold and the distribution uniformity exceeds the upper limit of the uniformity threshold, the integration time needs to be reduced. The intensity reduction weight λ3 refers to the influence weight of fluorescence intensity on the integration time adjustment when the integration time is reduced, and the uniformity reduction weight λ4 refers to the influence weight of distribution uniformity on the integration time adjustment when the integration time is reduced. The values ​​of λ3 and λ4 are different for different target tissue types. For example, a larger λ3 value means that for that type of target tissue, excessively high fluorescence intensity has a more critical impact on the integration time adjustment, and it is necessary to prioritize reducing the fluorescence intensity by reducing the integration time of the sensor, with λ3+λ4=1.

[0105] Meanwhile, for target tissues with fluorescence intensity less than the lower limit of the intensity threshold and distribution uniformity less than the lower limit of the uniformity threshold, the integration time is appropriately increased; while for target tissues with fluorescence intensity greater than the upper limit of the intensity threshold and distribution uniformity greater than the upper limit of the uniformity threshold, the integration time is appropriately reduced. The integration time adjustment is calculated according to the above formula, and an adjustment signal containing the adjustment direction and amplitude is generated, thereby realizing intelligent adjustment of the sensor's integration time and improving the adaptability of the imaging system to different imaging scenarios.

[0106] By adjusting the sensor integration time, the acquisition of light signals is optimized, improving imaging quality. Verification of the adjustment effect ensures that the integration time adjustment achieves the expected target. If not, further optimization can be performed. The control feedback module adjusts the sensor's integration time control circuit parameters according to the adjustment signal to adjust the sensor's integration time. After adjustment, the adjusted fluorescence image is acquired again, and the fluorescence intensity and distribution uniformity before and after adjustment are compared to verify the adjustment effect. If not, the above steps are iteratively executed to dynamically adjust the sensor's integration time. If it meets the target, there is no need to adjust the sensor's integration time. This optimizes the sensor's ability to acquire light signals, improves imaging quality, and ensures that the imaging system can stably and accurately acquire high-quality fluorescence images.

[0107] The image processing module is used to register white light images and fluorescence images, and then superimposes the registered white light images and fluorescence images in real time to generate a superimposed image.

[0108] Specifically, such as Figure 3 As shown, the registration process logic includes:

[0109] Structural feature points are extracted from white light images and the edges of target tissues in fluorescence images. The structural feature points are then coarsely matched with the edges of target tissues to obtain matching feature point pairs.

[0110] Based on the matching feature point pairs, the fluorescence image is mapped to the coordinate system of the white light image;

[0111] The transformed fluorescence image is resampled and spatially aligned using bilinear interpolation to obtain the registered white light image and fluorescence image.

[0112] White light images primarily reflect the morphological structure of the target tissue, while fluorescence images highlight its fluorescence characteristics. By extracting and matching key features from both, a foundation can be laid for accurately mapping the fluorescence image to the coordinate system of the white light image, thus achieving image registration. For the white light image, a scale-invariant feature transform algorithm is used to extract structural feature points. This algorithm constructs an image scale space, detects extreme points in the scale space, and calculates the orientation and descriptors of the feature points, thereby obtaining a series of scale- and rotation-invariant structural feature points. For the fluorescence image, the Canny algorithm is used to extract the edges of the target tissue. First, the fluorescence image is grayscaled and denoised. Then, the gradient magnitude and gradient direction of the fluorescence grayscale image are calculated. Non-maximum suppression and double threshold detection are used to determine the edges of the target tissue. When coarsely matching the extracted structural feature points with the edges of the target tissue, the distance from the structural feature points to each point on the edge of the target tissue is calculated. The nearest edge point is identified as a potential matching point, thus obtaining a preliminary matching feature point pair. This provides a preliminary correspondence between the white light image and the fluorescence image, laying the foundation for subsequent precise registration and improving the efficiency and accuracy of image registration.

[0113] Only by unifying the fluorescence image and the white light image into the same coordinate system can their accurate superposition and analysis be achieved. This allows doctors to comprehensively observe tissue morphology and fluorescence characteristics. Based on the matching feature point pairs obtained from coarse matching, the transformation matrix is ​​calculated using an affine transformation model. The linear equations are then solved using the least squares method to determine the translation, rotation, scaling, and shearing parameters of the affine transformation, resulting in the transformation matrix. This transformation matrix is ​​then used to transform the coordinates of all pixels in the fluorescence image, mapping the fluorescence image to the coordinate system of the white light image. This achieves preliminary spatial alignment between the fluorescence image and the white light image, providing a spatial basis for subsequent pixel resampling and precise registration.

[0114] After mapping the fluorescence image to the coordinate system of the white light image, the coordinate transformation causes changes in pixel positions. Therefore, pixel resampling and spatial alignment are necessary to ensure the continuity and accuracy of the white light and fluorescence images, achieving precise registration. For each new pixel in the transformed fluorescence image, the pixel value is calculated using bilinear interpolation. Centered on the new pixel position, four neighboring pixels are found in the original fluorescence image. Based on the positions and values ​​of these four neighboring pixels, the pixel value of the new pixel is calculated using the bilinear interpolation formula. After resampling, the fluorescence and white light images achieve spatial alignment at the pixel level, resulting in registered white light and fluorescence images. This improves registration accuracy, ensures spatial continuity and accuracy between the white light and fluorescence images, and makes the registered white light and fluorescence images more suitable for subsequent overlay and analysis.

[0115] Furthermore, such as Figure 4 As shown, the logic for coarse matching includes:

[0116] Calculate the distance between structural feature points and the edge of the target tissue, and configure an error threshold to remove mismatched points;

[0117] The weights of matching feature points are dynamically adjusted based on the uniformity of distribution.

[0118] The number of iterations and error threshold for coarse matching are optimized based on the type of the target organization.

[0119] In the initial coarse matching process, there may be some mismatches due to noise and local image similarity. By calculating the distance between the structural feature point and the edge of the target tissue and setting an error threshold, these mismatches can be eliminated, improving the accuracy of matching. For each structural feature point, the Euclidean distance to all points on the edge of the target tissue is calculated, and an error threshold is set. If the distance from the structural feature point to the nearest edge point is greater than the error threshold, it is determined to be a mismatch and eliminated. This effectively reduces the number of mismatches, improves the quality of matched feature point pairs, and enhances the reliability of image registration.

[0120] The contribution of matching feature points in different regions to image registration varies. Matching feature points corresponding to target tissues with good distribution uniformity are more representative of the overall structure of white light and fluorescence images. By dynamically adjusting the weights of matching feature points, registration can be made more accurate and stable. The white light and fluorescence images are divided into multiple sub-regions, and the distribution density of matching feature points in each sub-region is calculated. For sub-regions with high distribution density and good distribution uniformity, the weights of their matching feature points are increased; for sub-regions with low distribution density and poor distribution uniformity, the weights of their matching feature points are decreased. High distribution density refers to the distribution density of matching feature points compared to a preset density threshold; a density greater than the threshold is considered high, and a density less than or equal to the threshold is considered low. Good distribution uniformity means that the distribution uniformity is within the uniformity threshold range, and poor distribution uniformity means that the distribution uniformity is less than the lower limit of the uniformity threshold or greater than the upper limit of the uniformity threshold. This optimizes the weight allocation of matching feature points, improves the accuracy and stability of registration, and makes the registration results more reflective of the true structure of the white light and fluorescence images.

[0121] Different types of target tissues have different structures and fluorescence properties, and therefore require different registration accuracy. By optimizing the number of iterations and the error threshold, the efficiency and accuracy of image registration for different types of target tissues can be improved. For example, for target tissues with simple structures and obvious fluorescence properties, the number of iterations can be appropriately reduced and the error threshold increased to improve registration efficiency. Conversely, for target tissues with complex structures and inconspicuous fluorescence properties, the number of iterations can be increased and the error threshold decreased to improve registration accuracy. In the actual registration process, the corresponding number of iterations and error threshold are determined based on the type of target tissue identified, and the coarse matching process is optimized. This improves the adaptability of the image processing module to different types of target tissues, making the coarse matching process more intelligent and efficient, and further enhancing the quality of image registration.

[0122] Specifically, the logic for generating the overlay images includes:

[0123] The registered white light image and fluorescence image are fused based on the boundary of the target tissue.

[0124] Perform color adjustments and optimizations on the merged image;

[0125] The color-adjusted and optimized image is output as the final overlay image.

[0126] Furthermore, the logic for fusion based on the boundaries of the target organization includes:

[0127] The boundaries of the target tissue were extracted from the registered white light image and fluorescence image, respectively;

[0128] The fusion weights of white light and fluorescence images are determined based on the type and distribution uniformity of the target tissue.

[0129] Based on the fusion weights, the white light image and the fluorescence image are fused pixel by pixel to obtain the fused image.

[0130] The boundary of the target organization is a key basis for fusion. By extracting the boundary of the target organization, the area to be fused can be clearly identified, making the fusion more accurate and highlighting the characteristics of the target organization. The Canny algorithm is used to perform edge detection on the registered white light image and fluorescence image respectively. First, the white light image and fluorescence image are grayscaled to remove the interference of color information on edge detection. Then, the white light image and fluorescence image are smoothed by a Gaussian filter to reduce the impact of noise on edge detection. Next, the gradient magnitude and gradient direction of each pixel in the white light image and fluorescence image are calculated. Non-maximum suppression is used to retain the true edge pixels. Then, double threshold detection is used to determine strong edges and weak edges. Finally, edge tracking is used to connect weak edges and strong edges to obtain the boundary of the target organization. This accurately extracts the boundary of the target organization and provides an accurate area range for subsequent determination of fusion weights and pixel fusion.

[0131] Different types of target tissues and their distribution uniformity vary, resulting in different information requirements for white light and fluorescence images. By determining reasonable fusion weights, the characteristics of the target tissue can be better displayed in the superimposed images. The mapping relationship between the type and distribution uniformity of the target tissue and the fusion weights can be determined. For example, for tumor tissue, if the distribution uniformity is good and the fluorescence intensity is greater than the upper limit of the intensity threshold, it indicates that the fluorescence information is more critical for diagnosis. In this case, the fusion weight of the fluorescence image is set to 0.7, and the fusion weight of the white light image is set to 0.3. If the distribution uniformity is poor, it is necessary to observe the tissue morphology in conjunction with the white light image. Therefore, the fusion weight of the white light image is appropriately increased. Through the analysis and experimentation of a large number of white light and fluorescence images of different types of target tissues, the range of values ​​for the fusion weights under different conditions is determined. This enables intelligent adjustment of the fusion weights according to the type of target tissue, making the fused image more prominent in terms of the key information of the target tissue and improving the diagnostic value of the image.

[0132] By fusing pixels, the information from white light and fluorescence images is organically combined to generate an image that comprehensively reflects tissue morphology and fluorescence characteristics, providing a foundation for subsequent color adjustment and diagnostic analysis. For each corresponding pixel in the registered white light and fluorescence images, weighted fusion is performed according to the determined fusion weight. By fusing all pixels, a fused image is obtained, thus realizing the organic fusion of white light and fluorescence images and generating an image containing rich information, providing strong support for subsequent image processing and diagnosis.

[0133] The fused image may suffer from color inconsistencies and insufficient contrast. Color adjustment and optimization can enhance the visual effect of the image, improve its readability and diagnostic accuracy. Histogram equalization algorithm is used to enhance the contrast of the fused image. By stretching the grayscale histogram of the fused image, the grayscale distribution of the fused image is made more uniform, enhancing the contrast. Gamma correction is used to adjust the brightness and color saturation of the fused image. According to the characteristics of the fused image and the doctor's observation needs, the gamma value is adjusted to make the colors of the fused image more vivid and natural. Finally, a superimposed image is obtained, which improves the visual quality of the superimposed image, making the details in the superimposed image clearer and more distinguishable, helping doctors to more accurately observe and diagnose lesions in target tissues.

[0134] After the preceding processing steps, the overlaid image has completed registration, fusion, and color optimization. The output of the overlaid image can provide doctors with an image that comprehensively reflects the tissue morphology and fluorescence characteristics for disease diagnosis and analysis. The color-adjusted and optimized overlaid image is saved in common image formats, such as JPEG and PNG, for storage and transmission. At the same time, the overlaid image is displayed on the system's screen in real time for doctors to observe and analyze, thus providing doctors with intuitive and accurate diagnostic information and meeting the needs of clinical diagnosis and research.

[0135] The control feedback module is used to synchronously display and store superimposed images, extract fluorescence images to generate feedback signals, automatically adjust the power of the excitation light and the focusing depth of the endoscope probe according to the real-time distance between the endoscope probe and the target tissue, and monitor the power of the excitation light and the temperature of the target tissue in real time to determine whether to trigger the early warning process.

[0136] Doctors need to observe the overlay images in real time to understand the morphology and fluorescence characteristics of the target tissue and make accurate diagnoses. The control feedback module receives the overlay image after registration, fusion, and color optimization from the image processing module and transmits it to the display screen of the imaging system, which can clearly present image details. At the same time, the control feedback module sends a synchronization signal to ensure that the display of the overlay image is synchronized with other operations of the imaging system, avoiding overlay image lag or delay. This provides doctors with a real-time and clear image observation interface, which helps doctors to quickly and accurately analyze the condition of the target tissue.

[0137] The storage of overlaid images can provide data support for subsequent diagnostic review, case studies, and medical data statistical analysis. The control feedback module stores the overlaid images in the imaging system and adds detailed metadata to the overlaid images, including patient information, imaging time, and imaging mode, to facilitate subsequent data retrieval and management. This achieves secure and efficient storage of overlaid images, laying the foundation for the long-term preservation and utilization of medical data.

[0138] The distance between the endoscope probe and the target tissue affects the fluorescence signal intensity. By adjusting the excitation light power in real time, clear and stable fluorescence images can be obtained at different distances. The built-in distance sensor acquires the real-time distance between the endoscope probe and the target tissue. The distance sensor converts the measured physical distance into a digital signal and transmits it to the control feedback module for further processing. The module determines the excitation light power based on the real-time distance between the endoscope probe and the target tissue. For example, when the real-time distance between the endoscope probe and the target tissue increases, the excitation light power is appropriately increased to ensure sufficient fluorescence intensity. When the real-time distance between the endoscope probe and the target tissue decreases, the excitation light power is reduced to avoid over-excitation and damage to the target tissue. Then, a power adjustment command is sent to the composite light source module. The composite light source module dynamically adjusts the excitation light power by adjusting the drive current. During the adjustment process, the power change is monitored in real time to ensure the accuracy and stability of the adjustment.

[0139] Based on the real-time distance changes between the endoscope probe and the target tissue, the adjustment amount of the endoscope probe's focusing depth is calculated using optical imaging principles and conventional formulas for focusing depth. An adjustment signal is then sent to the endoscope probe to precisely adjust its position, achieving dynamic adjustment of the focusing depth. During the adjustment process, the focusing state of the endoscope probe is monitored in real time to ensure that the target tissue is always within the range of clear imaging. This enables real-time and precise adjustment of the endoscope probe's focusing depth, ensuring clear imaging of the target tissue at different distances and improving the adaptability and accuracy of the imaging system.

[0140] Specifically, the triggering logic for early warning processing includes:

[0141] Real-time monitoring of excitation light power and target tissue temperature;

[0142] The power of the excitation light and the temperature of the target tissue are compared with the configured power threshold and temperature threshold, respectively, to determine whether they exceed the limits.

[0143] If the limit is exceeded, an early warning will be triggered. The warning level will be determined and a response will be initiated based on the degree and duration of the exceedance.

[0144] The excitation light power and the temperature of the target tissue are monitored in real time. The excitation light power can be obtained from the composite light source module, while the temperature data of the target tissue is measured by a miniature temperature sensor, ensuring the real-time nature and accuracy of both data.

[0145] The power of the excitation light and the temperature of the target tissue are compared with preset power thresholds and temperature thresholds, respectively. The power threshold is mainly to avoid tissue damage and phototoxicity caused by excessive excitation light power, while excessively low excitation light power will affect the imaging effect. The temperature threshold is determined based on the physiological tolerance range of the target tissue to prevent burns or other physiological dysfunctions caused by excessive temperature.

[0146] If the power of the excitation light exceeds the power threshold range (i.e., below the lower limit or above the upper limit), or if the temperature of the target tissue exceeds the temperature threshold range (i.e., below the lower limit or above the upper limit), an early warning is immediately triggered. Simultaneously, based on the degree to which the excitation light power exceeds the power threshold, the range by which the target tissue temperature exceeds the temperature threshold, and the duration of the exceedance, the warning level is determined, such as a mild warning, a moderate warning, or a severe warning, in order to take appropriate response measures. In a mild warning, the exceedance duration is relatively short; in a moderate warning, the exceedance duration is moderate; and in a severe warning, the exceedance duration is relatively long.

[0147] Once an early warning is triggered, warning signals are issued through various means, such as displaying prominent warning information on the imaging system's screen, issuing audible alarms, and sending warning notifications to remote monitoring terminals. At the same time, depending on the level of the warning, corresponding response measures are taken, such as reducing the power of the excitation light, suspending imaging operations, and cooling the target tissue, to ensure the safety and effectiveness of the imaging process.

Claims

1. A high-definition fluorescence endoscopic imaging system, characterized in that, include: Composite light source module, optical imaging module, image processing module, and control feedback module; The composite light source module is used to select wavelength combinations according to the imaging mode, identify the edge of the target tissue to determine the shape of the excitation light spot and generate excitation light to irradiate the target tissue to generate fluorescence signals, and dynamically adjust the power of the excitation light based on the feedback signal. The optical imaging module is used to collect mixed light signals through an endoscope probe, separate the mixed light signals into white light signals and fluorescence signals, convert the light signals into electrical signals through a sensor, and simultaneously output white light images and fluorescence images. The integration time of the sensor is dynamically adjusted based on the feedback signal. The dynamic adjustment of the sensor's integration time includes: Extract fluorescence intensity, distribution uniformity, and target tissue type from fluorescence images; The integral time adjustment is calculated based on fluorescence intensity, distribution uniformity, and target tissue type, and an adjustment signal is generated. Adjust the sensor's integration time based on the adjustment signal and verify the adjustment effect; The image processing module is used to register the white light image and the fluorescence image based on matching feature point pairs, and then fuse the registered white light image and the fluorescence image in real time based on the boundary of the target tissue to generate an overlay image; the registration process includes: Structural feature points are extracted from white light images and the edges of target tissues in fluorescence images. The structural feature points are then coarsely matched with the edges of target tissues to obtain matching feature point pairs. Based on the matching feature point pairs, the fluorescence image is mapped to the coordinate system of the white light image; The transformed fluorescence image is resampled and spatially aligned by bilinear interpolation to obtain the registered white light image and fluorescence image. The fusion based on the boundary of the target organization includes: The boundaries of the target tissue were extracted from the registered white light image and fluorescence image, respectively; The fusion weights of white light and fluorescence images are determined based on the type and distribution uniformity of the target tissue. Pixel fusion is performed on white light image and fluorescence image based on fusion weight to obtain fused image; The control feedback module is used to extract the fluorescence intensity and distribution uniformity of the fluorescence image to generate a feedback signal, monitor the power of the excitation light and the temperature of the target tissue in real time, and determine whether to trigger an early warning process.

2. The high-definition fluorescence endoscopic imaging system as described in claim 1, characterized in that, The generated excitation light includes: Pre-scan the target tissue to identify its edges and determine the shape of the excitation light spot; The wavelength combination is invoked according to the imaging mode, and the timing driving parameters of the wavelength combination are loaded synchronously. A driving signal is sent to the laser source according to the wavelength combination, and the laser source is driven to emit pulse-modulated excitation light by combining the timing driving parameters of the wavelength combination. The pulsed excitation light is transmitted to the endoscope probe and the target tissue is irradiated according to the shape of the excitation light spot to generate a fluorescence signal.

3. The high-definition fluorescence endoscopic imaging system as described in claim 2, characterized in that, The edges of the identified target tissue include: A fluorescence image is obtained by pre-scanning the target tissue, and then the fluorescence image is processed into a grayscale image to obtain a fluorescence grayscale image; Calculate the gradient magnitude and gradient direction of each pixel in the fluorescence grayscale image, identify edge pixels, and connect the edge pixels to obtain the initial target tissue edge; Morphological processing is performed on the initial target tissue edges to obtain the target tissue edges.

4. The high-definition fluorescence endoscopic imaging system as described in claim 3, characterized in that, The generated feedback signal includes: Receive fluorescence images and update the location region of the target tissue based on a target tracking algorithm; The average pixel intensity and pixel intensity distribution of pixels in the location region of the target tissue are obtained by pixel traversal, and the fluorescence intensity and distribution uniformity of the fluorescence image are obtained. The feedback signal is determined based on the fluorescence intensity and distribution uniformity of the fluorescence image.

5. The high-definition fluorescence endoscopic imaging system as described in claim 4, characterized in that, The coarse matching includes: Calculate the distance between structural feature points and the edge of the target tissue, and configure an error threshold to remove mismatched points; The weights of matching feature points are dynamically adjusted based on the uniformity of distribution. The number of iterations and error threshold for coarse matching are optimized based on the type of the target organization.

6. The high-definition fluorescence endoscopic imaging system as described in claim 5, characterized in that, The generation of the overlay image includes: The registered white light image and fluorescence image are fused based on the boundary of the target tissue. Perform color adjustments and optimizations on the merged image; The color-adjusted and optimized image is output as the final overlay image.

7. The high-definition fluorescence endoscopic imaging system as described in claim 6, characterized in that, The triggering of the early warning process includes: Real-time monitoring of excitation light power and target tissue temperature; The power of the excitation light and the temperature of the target tissue are compared with the configured power threshold and temperature threshold, respectively, to determine whether they exceed the limits. If the limit is exceeded, an early warning will be triggered. The warning level will be determined and a response will be initiated based on the degree and duration of the exceedance.

Images