A laser scanning path planning method and system applied to laser scar removal

Abstract

The application provides a laser scanning path planning method and system applied to laser scar removal, fusing a structured light scanning mode and a speckle imaging mode to obtain surface shape data of a target scar, using an optical coherence tomography mode to obtain deep shape data of the target scar, and constructing a three-dimensional structure model of the target scar; a multi-modal sensing array is used to detect physiological parameters of the target scar; a convolutional neural network outputs a type of the target scar, and selects a laser source suitable for a laser spot distribution mode of the target scar; a spatial boundary constraint condition is set, a fast random tree algorithm is used to generate a simulated laser scanning path suitable for a depth feature of the target scar in the three-dimensional structure model of the target scar; in a simulation software, a laser source is enabled to simulate irradiation on the target scar according to the simulated laser scanning path, a laser irradiation simulation graph of the target scar is generated, the simulated laser scanning path is verified and secondarily adjusted, and an actual laser scanning path is generated.

Description

Technical Field

[0001] The present invention belongs to the technical field of laser treatment, and in particular relates to a laser scanning path planning method and system for laser scar removal. Background Art

[0002] Laser scar removal is a treatment method that uses laser technology to improve or remove scars. This technique uses laser light of a specific wavelength and intensity to target scar tissue, promoting skin regeneration, reducing scar volume, and improving scar color and smoothness. Laser scar removal relies on several mechanisms, including stimulating collagen reorganization and regeneration, selective photothermolysis, improving blood circulation, and reducing inflammation.

[0003] However, under existing technologies, when manually operating the laser source or controlling the movement of the laser source by a conventional drive mechanism, the extensive path planning causes the laser focus to easily shift, and the laser cannot ensure full coverage, which not only easily damages normal tissues but also causes scars to be in a sub-therapeutic state. At the same time, existing treatment methods lack efficient and accurate detection mechanisms in key dimensions such as scar morphology description, depth identification, dermal temperature monitoring, and tissue elasticity assessment, resulting in a mismatch between the energy output of the laser source and the dynamic treatment needs, which can easily lead to clinical complications such as excessive thermal damage or insufficient treatment energy. Summary of the Invention

[0004] The present invention aims to provide a laser scanning path planning method and system for laser scar removal, so as to solve the technical problem that the laser scanning path planning of conventional laser scar removal operations cannot be fully adapted to the morphology and physiological parameters of the scar under the existing technology.

[0005] To solve the above problems, the technical solution of the present invention is: a laser scanning path planning method for laser scar removal, comprising the following steps:

[0006] S1: Surface morphological data of the target scar are obtained by fusing structured light scanning and speckle imaging, and deep morphological data of the target scar are obtained by optical coherence tomography. Based on these surface and deep morphological data, a three-dimensional structural model of the target scar is constructed.

[0007] S2: Detecting physiological parameters of target scar using a multimodal sensing array;

[0008] S3: Inputting the three-dimensional structural model and physiological parameters of the target scar into the pre-trained convolutional neural network, the convolutional neural network outputs the type of the target scar and selects a laser spot distribution pattern of the laser source suitable for the target scar;

[0009] S4: Setting spatial boundary constraints, using a fast random tree algorithm to generate a simulated laser scanning path suitable for the depth characteristics of the target scar in the 3D structural model of the target scar;

[0010] S5: In the simulation software, enable the laser source to simulate irradiation of the target scar according to the simulated laser scanning path, generate a simulation graph of the target scar along the laser irradiation, verify and re-adjust the simulated laser scanning path based on the laser irradiation simulation graph, and generate an actual laser scanning path.

[0011] Preferably, in S1, obtaining surface morphological data of the target scar using a structured light scanning method includes the following steps:

[0012] S11: Using a structured light scanner, a dynamic grating is used to generate multiple sets of light beams at different angles and projected onto different areas of the target scar. For the concave areas of the target scar, the operating frequency of the dynamic grating is manually reduced, while for the convex areas of the target scar, the operating frequency of the dynamic grating is manually increased.

[0013] Using a camera to capture multiple sets of deformation image data of grating patterns on the target scar;

[0014] S12: Based on the deformation image data of the target scar, on the premise of obtaining the horizontal position information of each pixel point on the surface layer of the target scar, the vertical position information of each pixel point on the surface layer of the target scar is calculated by using a phase shift algorithm, a phase unwrapping algorithm and a triangulation algorithm;

[0015] S13: generating first three-dimensional point cloud data of the target scar surface based on the horizontal position information and the vertical position information of each pixel point in the target scar, wherein the first three-dimensional point cloud data is used to indicate three-dimensional geometric contour information of the target scar surface.

[0016] Preferably, obtaining surface morphological data of the target scar using speckle imaging in S1 includes the following steps:

[0017] S14: using a coherent laser generator to generate multiple groups of coherent lasers at several different angles to irradiate different areas of the target scar;

[0018] Using a camera to capture multiple sets of speckle image data reflected from the target scar surface;

[0019] S15: For the speckle image data of the target scar, the contrast of the speckle image is calculated using the spatial speckle contrast algorithm and the temporal speckle contrast algorithm to evaluate the speckle image quality;

[0020] Calculate the average size of the speckle image;

[0021] The speckle image is processed using a complete empirical mode decomposition algorithm, and the texture profile of the surface layer of different regions in the target scar is extracted by combining the contrast and size information of the speckle image, to generate second three-dimensional point cloud data of the surface layer of the target scar, which is used to indicate the roughness information of the surface layer of the target scar.

[0022] In the same coordinate system, the first three-dimensional point cloud data and the second three-dimensional point cloud data are superimposed to obtain the surface morphology data of the target scar.

[0023] Preferably, the deep layer morphology data of the target scar is obtained in S1 using an optical coherence tomography method, including the following steps:

[0024] S16: Using an optical coherence tomography system, a broadband near-infrared laser is irradiated onto different regions of the target scar in a direction perpendicular to the surface of the target scar, and the corresponding reflected light of the different regions of the target scar is collected, wherein the reflected light generated by the normal skin tissue under the target scar is removed by screening.

[0025] S17: For the reflected light of the target scar in different regions, the reflected light frequency domain signal of the target scar is converted into depth information of the target scar by Fourier transform, and in the same coordinate system of the first three-dimensional point cloud data and the second three-dimensional point cloud data, third three-dimensional point cloud data of the deep layer of the target scar is generated, which is used to indicate the depth information of the deep layer of the target scar.

[0026] Preferably, after constructing the three-dimensional structure model of the target scar in S1, the following steps are further included:

[0027] S18: The three-dimensional structure model image of the target scar is preprocessed, which includes:

[0028] The three-dimensional structure model image of the target scar is subjected to noise suppression processing by a Gaussian filtering algorithm;

[0029] The boundaries of the concave region, the convex region, and the surface undulating region in the target scar, and the boundary between the target scar and the surrounding normal skin tissue are identified using a Canny edge detection algorithm;

[0030] The image of the target scar is subjected to edge enhancement by a gradient calculation method of a Sobel operator, which is used to highlight the boundary line between the target scar and the surrounding normal skin tissue;

[0031] The surface curvature distribution information of the target scar in the three-dimensional coordinates is calculated by a Gaussian curvature algorithm or a least squares surface fitting algorithm.

[0032] Preferably, the physiological parameters of the target scar are detected in S2 using a multi-modal sensing array, including the following steps:

[0033] S21: In the coordinate system of the target scar three-dimensional structural model, infrared thermal imaging data of the target scar region is obtained using an infrared thermal imager. After performing thermal field reconstruction processing on the infrared thermal imaging data, a two-dimensional temperature distribution map of the target scar region in the horizontal direction is calculated according to a temperature measurement formula. The two-dimensional temperature distribution map is used to indicate the thermal conductivity of the skin tissue in the target scar region.

[0034] S22: In the coordinate system of the target scar three-dimensional structural model, using a humidity sensor to measure and obtain surface moisture content data within the target scar region, and performing Kalman filtering and wavelet noise reduction on the moisture content data. The surface moisture content data is used to indicate the metabolic intensity and hardness of the skin tissue within the target scar region;

[0035] S23: In the coordinate system of the target scar three-dimensional structural model, a blood oxygen saturation detection device is used to measure and obtain blood oxygen saturation data in the target scar area, and Kalman filtering and wavelet noise reduction processing are performed on the blood oxygen saturation data. The blood oxygen saturation data is used to indicate the blood circulation status and skin tissue metabolic intensity in the target scar area.

[0036] Preferably, the convolutional neural network is configured to perform iterative training based on the three-dimensional structural model and physiological parameters of an existing scar. After training, the convolutional neural network is used to identify the type of target scar and select the laser spot distribution pattern based on the three-dimensional structural model and physiological parameters of the target scar.

[0037] Target scar types include mature flat scars, hypertrophic scars, keloid scars, atrophic scars, contracture scars, and depressed scars;

[0038] Laser spot distribution modes include annular spot, grid spot and multi-focal spot. Different laser spot distribution modes are suitable for scar treatment needs of different types, three-dimensional structural models and physiological parameters.

[0039] Preferably, in S4, a spatial boundary constraint condition is set, and in the three-dimensional structural model of the target scar, a fast random tree algorithm is used to generate a simulated laser scanning path suitable for the depth characteristics of the target scar, comprising the following steps:

[0040] S41: Using the boundary line between the target scar and the surrounding normal skin tissue as the spatial boundary constraint for the laser irradiation operation, within the two-dimensional closed boundary area formed by the target scar and the surrounding normal skin tissue, based on the deep morphological data of the target scar and using the depth characteristics of the target scar as a judgment criterion, a number of grid points are generated with different distribution densities from deep to shallow;

[0041] Initialize the laser scanning path of the laser source. The laser source is configured to face perpendicularly to the target scar surface. Randomly sample a preceding grid point x_rand_1 in the surface plane of the 3D structural model of the target scar. Align the center point of the laser irradiation position of the laser source with the preceding grid point x_rand_1.

[0042] S42: Calculate the position of any subsequent grid point x_nearest_1 that is closest to the previous grid point x_rand_1;

[0043] According to the laser spot distribution pattern, the coverage area of ​​a single laser irradiation of the laser source is determined. The diameter or maximum length of the coverage area of ​​a single laser irradiation of the laser source is used as the laser source movement step length. Between the previous grid point x_rand_1 and the next grid point x_nearest_1, several extension nodes x_new are generated with the laser source movement step length as the interval.

[0044] S43: Based on the three-dimensional structural model of the target scar, determine whether the extended node x_new is in a legal position and whether there are obstacles from adjacent grid points to the extended node, or between adjacent extended nodes. If legal and there are no obstacles, enable the laser source to move along the distribution path of the extended node x_new and irradiate until it reaches the subsequent grid point x_nearest_1;

[0045] S44: After the laser source reaches the subsequent grid point x_nearest_1, a new scanning path planning is performed between adjacent grid points. The current subsequent grid point x_nearest_1 is used as the new preceding grid point x_rand_2. In the scanning path planning between adjacent grid points, the position of any subsequent grid point x_nearest_2 closest to the preceding grid point x_rand_2 is calculated, excluding the preceding grid point x_rand_1. Steps S42-S43 are repeated.

[0046] S45: When the laser source irradiation position covers all grid points in the target scar area, a sequence {x_0, x_1, ..., x_n} simulating the laser scanning path is backtracked from the three-dimensional structure model of the target scar.

[0047] Preferably, S4 further includes the following steps:

[0048] S46: Keeping the spatial boundary constraint conditions and the grid point distribution conditions unchanged, repeating steps S42-S45 to generate several groups of simulated laser scanning paths, and selecting the group of simulated laser scanning paths with the least total number of laser source movement steps as the optimal simulated laser scanning path;

[0049] In addition, the laser source has a maximum number of laser irradiation times for any grid point. When the number of laser irradiation times for any grid point reaches the upper limit, the grid point will no longer be selected as the subsequent grid point x_nearest_n in the subsequent scanning path planning between adjacent grid points.

[0050] Preferably, in S5, a laser irradiation simulation graphic of the target scar is generated in the simulation software, and the simulated laser scanning path is verified and adjusted again based on the laser irradiation simulation graphic, including the following steps:

[0051] S51: establishing and running a simulation program in MATLAB or Simulink simulation software to simulate the laser irradiation of the target scar area by the laser source along the simulated laser scanning path;

[0052] The laser irradiation simulation graphics include a visual laser irradiation movement path in the target scar area, and color markings formed by the total laser heat received by different areas of the target scar;

[0053] S52: Manually verify whether the laser irradiation area of ​​the laser source completely covers the target scar, and whether the total laser heat received by different areas of the target scar matches the depth characteristics of the target scar.

[0054] Based on the same concept, the present invention also provides a laser scanning path planning system for laser scar removal, which is used to execute any one of the laser scanning path planning methods for laser scar removal described above, comprising:

[0055] Scar morphology acquisition module, used to collect surface morphology data and deep morphology data of target scars;

[0056] A model building module is used to establish a three-dimensional structural model of the target scar;

[0057] A physiological parameter acquisition module, used to acquire physiological parameters of the target scar;

[0058] An identification module, used to identify the type of target scar and select a laser spot distribution pattern suitable for the target scar;

[0059] A laser scanning path generation module is used to generate a simulated laser scanning path and an actual laser scanning path;

[0060] The simulation module is used to simulate the laser irradiation of the target scar area by the laser source along the simulated laser scanning path, and generate a laser irradiation simulation graphic of the target scar.

[0061] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art:

[0062] The present invention provides a laser scanning path planning method and system for laser scar removal. First, a structured light scanning method, a speckle imaging method, and an optical coherence tomography method are integrated to obtain the surface morphological data and deep morphological data of the target scar, and to construct a three-dimensional structural model of the target scar. The three-dimensional structural model of the target scar can fully express its three-dimensional geometric contour, surface roughness, and depth information. At the same time, a multimodal sensor array is used to detect the various physiological parameters of the target scar, so that the laser treatment method and the physiological parameters of the target scar are fully adapted. Based on the three-dimensional structural model and physiological parameters of the target scar, a laser spot distribution pattern suitable for the target scar is selected to achieve personalized laser treatment. Finally, in the three-dimensional structural model of the target scar, a fast random tree algorithm is used to generate a simulated laser scanning path suitable for the depth characteristics of the target scar. The laser scanning path planning can effectively combine the depth characteristics of the target scar to achieve accurate laser scanning path planning and efficient laser treatment of scars of different types and morphologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A flow chart of a laser scanning path planning method for laser scar removal provided by the present invention;

[0064] Figure 2 A schematic structural diagram of a laser scanning path planning system for laser scar removal provided by the present invention. DETAILED DESCRIPTION

[0065] The following is a detailed description of a laser scanning path planning method and system for laser scar removal proposed by the present invention in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description and claims.

[0066] First embodiment

[0067] See Figure 1-Figure 2 This embodiment provides a laser scanning path planning method for laser scar removal, comprising the following steps:

[0068] S1: Structured light scanning and speckle imaging are used to obtain surface morphological data of the target scar, and optical coherence tomography is used to obtain deep morphological data of the target scar. Based on the surface and deep morphological data of the target scar, a three-dimensional structural model of the target scar is constructed.

[0069] The three-dimensional structural model of the target scar is used to accurately display and evaluate the location, size, depth, geometric contour shape of the target scar and its impact on surrounding tissues. The three-dimensional structural model of the target scar provides key basic data for subsequent laser scanning path planning, thereby achieving precise and personalized implementation of scar treatment surgery.

[0070] S2: Use a multimodal sensor array to detect the physiological parameters of the target scar. The physiological parameters of the target scar can be used to further evaluate the scar characteristics and realize the design of personalized laser treatment plan.

[0071] S3: The three-dimensional structural model and physiological parameters of the target scar are input into the pre-trained convolutional neural network. The convolutional neural network outputs the type of the target scar and selects a laser spot distribution pattern suitable for the target scar.

[0072] The laser spot distribution pattern determines the distribution of laser irradiation on the target scar surface. Different laser spot distribution patterns are suitable for scars of different types and structures to achieve the best laser treatment effect.

[0073] S4: Set spatial boundary constraints and use a fast random tree algorithm to generate a simulated laser scanning path suitable for the depth characteristics of the target scar in the three-dimensional structure model of the target scar.

[0074] S5: In the simulation software, enable the laser source to simulate irradiation of the target scar according to the simulated laser scanning path, generate a laser irradiation simulation graphic of the target scar, verify and re-adjust the simulated laser scanning path based on the laser irradiation simulation graphic, and generate an actual laser scanning path.

[0075] By simulating and verifying the simulated laser scanning path, the accuracy of the actual laser scanning path generated is improved.

[0076] In summary, this embodiment provides a laser scanning path planning method for laser scar removal. First, a three-dimensional structural model of the target scar is established, and physiological parameters of the target scar are obtained. Based on this, a convolutional neural network adaptively selects a laser spot distribution pattern suitable for the target scar. Finally, a simulated laser scanning path is generated within the three-dimensional structural model of the target scar. This simulated laser scanning path is manually verified to generate the actual laser scanning path. This embodiment effectively achieves full adaptation of the laser treatment method to the target scar's morphology, type, and physiological parameters, improves the accuracy and stability of the laser scanning path, and ensures the effectiveness of laser scar removal treatment.

[0077] The following further details the specific steps and functions of the laser scanning path planning method for laser scar removal provided in this embodiment:

[0078] Preferably, in one embodiment, obtaining surface morphological data of the target scar using structured light scanning in S1 includes the following steps:

[0079] S11: Using a structured light scanner, a dynamic grating is used to generate multiple sets of light beams at different angles and projected onto different areas of the target scar. For the concave areas of the target scar, the operating frequency of the dynamic grating is manually reduced, while for the convex areas of the target scar, the operating frequency of the dynamic grating is manually increased.

[0080] A camera is used to capture multiple sets of deformation image data of grating patterns on the target scar.

[0081] For scars, which have complex geometric shapes, properly adjusting the grating frequency can significantly improve scanning performance. Specifically, in the concave areas of a target scar, where the surface shape changes significantly, reducing the grating frequency increases the width of the grating stripes, thereby obtaining more sampling points within the same field of view and improving the resolution of the entire concave area of ​​the target scar.

[0082] On the contrary, in the raised part of the target scar, the shape of its surface changes more smoothly. By increasing the grating frequency, the width of the grating stripes can be reduced, thereby obtaining denser sampling points in the same field of view and reducing sampling errors.

[0083] S12: For the deformation image data of the target scar, on the premise of obtaining the horizontal position information of each pixel point on the surface of the target scar, the vertical position information of each pixel point on the surface of the target scar is calculated through a phase shift algorithm, a phase unwrapping algorithm and a triangulation algorithm.

[0084] Specifically, for the deformation image data of the target scar, the horizontal position information of each pixel point on the surface of the target scar is directly obtained based on the two-dimensional plane image of the target scar. Subsequently, the phase information of each pixel is calculated. The phase shift algorithm refers to calculating the phase by analyzing the images taken under four different phase shifts. In this embodiment, the four phase shift amounts are preset to be 0, 、 、 , the corresponding grayscale values ​​of the target scar deformation image are I0(x,y), I1(x,y), I2(x,y) and I3(x,y).

[0085] The expression of the phase shift algorithm is:

[0086]

[0087] in, is the wrapped phase, that is, the calculated phase value is arrive between.

[0088] Since the phase calculation result is a wrapped phase, there is The periodic jumps in the image need to be unwrapped into a continuous phase. In this embodiment, a quality-guided phase unwrapping algorithm is used to select appropriate pixels as the unwrapping starting point based on the quality information of the target scar deformation image. Phase unwrapping is then performed along a high-quality path, effectively avoiding the propagation of noise and errors.

[0089] Finally, the depth information corresponding to each pixel is calculated through a triangulation algorithm. In this embodiment, the focal length of the preset camera is f, the baseline length (the distance between the camera and the laser source) is b, and the pixel offset (the horizontal offset of the pixel obtained by phase calculation) is Δx.

[0090] The expression of the triangulation algorithm is:

[0091]

[0092] Where d is the calculated surface depth of the target scar. The triangulation algorithm exploits the similarity of triangles. In structured light scanning, the change in phase is proportional to the change in depth of the object's surface. Therefore, the horizontal offset of the pixel is calculated using the phase difference, and the surface depth of the target scar can be calculated using trigonometric relationships.

[0093] S13: generating first three-dimensional point cloud data of the target scar surface based on the horizontal position information and the vertical position information of each pixel point in the target scar, wherein the first three-dimensional point cloud data is used to indicate three-dimensional geometric contour information of the target scar surface.

[0094] Furthermore, in one embodiment, obtaining surface morphological data of a target scar using speckle imaging in S1 includes the following steps:

[0095] S14: using a coherent laser generator to generate multiple groups of coherent lasers at several different angles to irradiate different areas of the target scar;

[0096] A camera is used to capture multiple sets of speckle image data reflected from the target scar surface.

[0097] S15: For the speckle image data of the target scar, the contrast of the speckle image is calculated using a spatial speckle contrast algorithm and a temporal speckle contrast algorithm to evaluate the quality of the speckle image.

[0098] Specifically, speckle contrast is an important indicator for measuring speckle image quality. It is calculated by first calculating the standard deviation of the speckle intensity (indicating the magnitude of intensity variation) and then dividing it by the mean speckle intensity. This yields a dimensionless contrast value. A larger value indicates better speckle pattern quality.

[0099] The expression of the spatial speckle contrast algorithm is:

[0100]

[0101] in, is the standard deviation of the speckle intensity, is the average value of the speckle intensity, calculated in a spatial window.

[0102] The expression of the temporal speckle contrast algorithm is:

[0103]

[0104] in, is the standard deviation of the speckle intensity over time, is the average value of the speckle intensity over time.

[0105] Then, the average size of the speckle image is calculated.

[0106] Specifically, the calculation expression of the average size of the speckle image is:

[0107]

[0108] Among them, r y represents the average size of the speckle image, λ is the wavelength, M is the system magnification, and F is the F number of the imaging system.

[0109] Subsequently, the speckle image was processed using the complete empirical mode decomposition algorithm. The contrast and size information of the speckle image were combined to extract the texture contours of the surface of different areas in the target scar, and the second three-dimensional point cloud data of the target scar surface was generated. The second three-dimensional point cloud data was used to indicate the roughness information of the target scar surface.

[0110] Specifically, define the operator , which aims to add noise to the original signal to facilitate decomposition.

[0111] Operator The expression of the formula is:

[0112]

[0113] in, is the original signal, are different realizations of zero-mean Gaussian white noise with unit variance, is the noise standard deviation.

[0114] The signal is decomposed using the complete empirical mode decomposition algorithm, which can decompose complex signals into a series of simple intrinsic mode functions (IMFs).

[0115] The expression of the complete empirical mode decomposition algorithm is:

[0116]

[0117] Where I represents the number of iterations, E1 represents the complete empirical mode decomposition algorithm of the first iteration, is the noise standard deviation.

[0118] Finally, the IMF layers of different frequencies are merged, and the resulting intrinsic mode functions are combined according to frequency to reconstruct a clearer and more accurate speckle image. Combining speckle contrast and size information, the texture contours of different surface areas within the target scar are extracted, generating a second 3D point cloud of the target scar surface.

[0119] In the same coordinate system, the first three-dimensional point cloud data and the second three-dimensional point cloud data are superimposed to obtain the surface morphological data of the target scar. That is, the surface morphological data of the target scar can not only represent the overall three-dimensional contour shape of the target scar surface, but also accurately represent the roughness detail data of the target scar surface, thereby improving the resolution effect of the target scar surface morphological data.

[0120] Furthermore, in one embodiment, in S1, obtaining deep morphological data of the target scar using optical coherence tomography includes the following steps:

[0121] S16: Using an optical coherence tomography system to generate a broadband near-infrared laser in a direction perpendicular to the surface of the target scar, the laser is irradiated onto different areas of the target scar, and the reflected light corresponding to the different areas of the target scar is collected. Since the collagen fibers of scar tissue are significantly different from those of normal skin, there are also significant differences in the reflected light generated by the laser between scar tissue and normal skin tissue. In this embodiment, the reflected light generated by the normal skin tissue below the target scar needs to be screened and removed first.

[0122] S17: For the reflected light of the target scar in different areas, the frequency domain signal of the reflected light of the target scar is converted into the depth information of the target scar through Fourier transform, and in the same coordinate system of the first three-dimensional point cloud data and the second three-dimensional point cloud data, a third three-dimensional point cloud data of the deep layer of the target scar is generated. The third three-dimensional point cloud data is used to indicate the depth information of the deep layer of the target scar, that is, the thickness parameters of different areas of the target scar.

[0123] Specifically, the expression of Fourier transform is:

[0124]

[0125] Among them, I OCT (z) is the signal intensity at depth z, α(k) is the spectrum signal at wave number k, k is the wave number, e ikz is the phase change of the light wave at depth z.

[0126] Furthermore, in one embodiment, after constructing the three-dimensional structural model of the target scar in S1, the following steps are further included:

[0127] S18: Preprocessing the three-dimensional structural model image of the target scar, including:

[0128] The noise of the three-dimensional structural model image of the target scar is suppressed by using the Gaussian filtering algorithm to reduce the random noise in the image.

[0129] The Canny edge detection algorithm was used to identify the boundaries of the concave area, convex area, and surface undulating area in the target scar, as well as the boundary between the target scar and the surrounding normal skin tissue.

[0130] The edge of the target scar image was enhanced by the gradient calculation method of the Sobel operator to highlight the boundary between the target scar and the surrounding normal skin tissue.

[0131] The surface curvature distribution information of the target scar in three-dimensional coordinates is calculated using the Gaussian curvature algorithm or the least squares surface fitting algorithm.

[0132] Preferably, in one embodiment, a multimodal sensor array is used in S2 to detect physiological parameters of the target scar, the multimodal sensor array including an integrated body of an infrared thermal imager, a humidity sensor, and a blood oxygen saturation detection device, including the following steps:

[0133] S21: In the coordinate system of the three-dimensional structural model of the target scar, use an infrared thermal imager to measure and obtain infrared thermal imaging data within the target scar area. After performing thermal field reconstruction processing on the infrared thermal imaging data, calculate and obtain a two-dimensional temperature distribution map within the target scar area in the horizontal direction according to the temperature measurement formula. The two-dimensional temperature distribution map is used to indicate the thermal conductivity performance of the skin tissue in the target scar area.

[0134] Specifically, an infrared thermal imager is used to capture infrared radiation from the target scar area, generating an initial thermal image. The grayscale value of the image reflects the surface temperature of the object, with higher grayscale values ​​indicating higher temperatures. Thermal field reconstruction is performed through interpolation and image fusion to improve the accuracy of infrared thermal imaging data. In the reconstructed thermal field image, the grayscale values ​​are converted to temperature values ​​using the infrared thermal imager's temperature measurement formula, resulting in a two-dimensional temperature distribution map.

[0135] The expression of the temperature measurement formula is:

[0136] T=a×I+b

[0137] Where T is temperature, I is grayscale value, and a and b are calibration coefficients.

[0138] S22: In the coordinate system of the target scar three-dimensional structure model, a humidity sensor is used to measure and obtain surface moisture content data in the target scar area, and Kalman filtering and wavelet noise reduction processing are performed on the moisture content data. The surface moisture content data is used to indicate the metabolic intensity and hardness of the skin tissue in the target scar area.

[0139] S23: In the coordinate system of the target scar three-dimensional structural model, a blood oxygen saturation detection device is used to measure and obtain blood oxygen saturation data in the target scar area, and Kalman filtering and wavelet noise reduction processing are performed on the blood oxygen saturation data. The blood oxygen saturation data is used to indicate the blood circulation status and skin tissue metabolic intensity in the target scar area.

[0140] Preferably, in one embodiment, the convolutional neural network is configured to perform iterative training based on the three-dimensional structural model and physiological parameters of an existing scar. The trained convolutional neural network is used to identify the type of target scar and select the laser spot distribution pattern based on the three-dimensional structural model and physiological parameters of the target scar.

[0141] Target scar types include mature flat scars, hypertrophic scars, keloids, atrophic scars, contracture scars, and depressed scars.

[0142] Laser spot distribution modes include annular spot, grid spot and multi-focal spot. Different laser spot distribution modes are suitable for scar treatment needs of different types, three-dimensional structural models and physiological parameters.

[0143] For example, for mature flat scars, due to their hard texture, the use of a grid spot can evenly cover the entire scar area, ensuring even distribution of laser energy. For hypertrophic scars, due to their large size and uneven surface, a multifocal spot can cover multiple focal points simultaneously, improving treatment efficiency and better adapting to the irregular shape of the scar. For keloids, which usually have clear boundaries but a hard texture and may be invasive, the use of a ring spot can be used to treat along the scar's border, effectively inhibiting scar growth and reducing damage to the surrounding normal skin.

[0144] Preferably, in one embodiment, a spatial boundary constraint condition is set in S4, and a fast random tree algorithm is used to generate a simulated laser scanning path suitable for the depth characteristics of the target scar in the three-dimensional structural model of the target scar, comprising the following steps:

[0145] S41: The boundary line between the target scar and the surrounding normal skin tissue is used as the spatial boundary constraint condition for the laser irradiation operation of the laser source. Within the two-dimensional closed boundary area formed by the target scar and the surrounding normal skin tissue, according to the deep morphological data of the target scar and the depth characteristics of the target scar as the judgment standard, a number of grid points with different distribution densities are generated from deep to shallow. That is, based on the three-dimensional structural model of the target scar, more densely distributed grid points are generated in the thicker area of ​​the target scar, and sparsely distributed grid points are generated in the thinner area of ​​the target scar.

[0146] Initialize the laser scanning path of the laser source. The laser source is configured to face perpendicular to the target scar surface. Randomly sample a preceding grid point x_rand_1 in the surface plane of the target scar's three-dimensional structural model. Align the center point of the laser irradiation position with the preceding grid point x_rand_1, that is, the preceding grid point x_rand_1 serves as the starting operating point of the laser source.

[0147] S42: Calculate the position of any subsequent grid point x_nearest_1 that is closest to the preceding grid point x_rand_1.

[0148] Based on the laser spot distribution pattern, the laser source's single laser irradiation coverage area is determined. The diameter or maximum length of the laser source's single laser irradiation coverage area is used as the laser source movement step size. Between the preceding grid point x_rand_1 and the succeeding grid point x_nearest_1, a number of extended nodes x_new are generated, spaced at intervals equal to the laser source movement step size. Specifically, in this embodiment, in planning a single scan path between adjacent grid points, two adjacent grid points are used as the starting and ending positions, and several extended nodes are generated along the straight line between the adjacent grid points. The laser source then moves and irradiates along the direction in which these extended nodes extend.

[0149] S43: Based on the three-dimensional structural model of the target scar, determine whether the extended node x_new is in a legal position and whether there are obstacles from adjacent grid points to the extended node, or between adjacent extended nodes. If it is legal and there are no obstacles, enable the laser source to move along the distribution path of the extended node x_new and irradiate until it reaches the subsequent grid point x_nearest_1.

[0150] Because scars have irregular shapes, normal skin tissue may exist between adjacent grid points. If the planned expansion node is located on normal skin, this means that the expansion node x_new is in an illegal position. Furthermore, scars may contain scabs, which can be defined as obstacles.

[0151] S44: After the laser source reaches the posterior grid point x_nearest_1, a new adjacent grid point scanning path is planned, taking the current posterior grid point x_nearest_1 as the new anterior grid point x_rand_2. In the adjacent grid point scanning path, the position of any posterior grid point x_nearest_2 closest to the anterior grid point x_rand_2 is calculated, and the steps S42-S43 are repeated.

[0152] It is worth noting that in this embodiment, in the adjacent grid point scanning path, the anterior grid point in the previous scanning path is no longer selected as the termination position in the current scanning path, avoiding multiple laser irradiation in the region between two grid points, causing overheating damage.

[0153] According to the planning rules of the laser scanning path, more repeated laser irradiation will be automatically formed in the region with a deeper target scar depth, i.e., the target scar absorbs more laser heat, while in the position with a shallower target scar depth, the number of repeated laser irradiation is reduced to prevent the scar from overheating, thereby realizing the adaptation of the laser scanning path to the characteristics of the target scar depth.

[0154] S45: When the laser source irradiation position covers all grid points in the target scar region, i.e., any grid point region has at least received one laser irradiation, the sequence {x_0, x_1,..., x_n} of the simulated laser scanning path can be generated from the three-dimensional structure model of the target scar.

[0155] Further, in one embodiment, the following steps are further included in S4:

[0156] S46: Keep the spatial boundary constraint condition and the grid point distribution condition unchanged, repeat the steps S42-S45 to generate several groups of simulated laser scanning paths, and select the one with the least total number of laser source movement steps as the optimal simulated laser scanning path.

[0157] Moreover, the laser source has a maximum number of laser irradiation for any grid point, and when the number of laser irradiation for any grid point reaches the upper limit, the grid point will not be selected as the posterior grid point x_nearest_n in the subsequent adjacent grid point scanning path planning, avoiding multiple laser irradiation in a single grid point, causing skin overheating damage.

[0158] Preferably, in one embodiment, in S5, a laser irradiation simulation graph of the target scar is generated in the simulation software, and the simulated laser scanning path is verified and adjusted based on the laser irradiation simulation graph, including the following steps:

[0159] S51: Establish and run a simulation program in MATLAB or Simulink simulation software to simulate the laser irradiation performed by the laser source on the target scar area along the simulated laser scanning path.

[0160] The laser irradiation simulation graphics include a visualized laser irradiation movement path in the target scar area, and color markings formed by the total laser heat received by different areas of the target scar.

[0161] Color marking specifically means that when a certain area of ​​the target scar receives a higher total laser heat, the color mark formed in that area will be darker, which is used to achieve visual analysis of the total laser heat received by different areas of the target scar.

[0162] S52: Manually verify whether the laser irradiation area of ​​the laser source completely covers the target scar, and whether the total laser heat received by different areas of the target scar matches the depth characteristics of the target scar. If it is not completely covered or not completely matched, the movement path of the laser source can be manually modified, or a second laser supplementary treatment can be manually performed on the local area after the current laser treatment.

[0163] Second embodiment

[0164] Based on the same concept, the present invention further provides a laser scanning path planning system for laser scar removal, which is used to execute the laser scanning path planning method for laser scar removal as described in any one of the first embodiments, including:

[0165] The scar morphology acquisition module is used to collect surface morphology data and deep morphology data of the target scar.

[0166] The model building module is used to establish a three-dimensional structural model of the target scar.

[0167] The physiological parameter acquisition module is used to acquire the physiological parameters of the target scar.

[0168] The identification module is used to identify the type of the target scar and select a laser spot distribution pattern of the laser source suitable for the target scar.

[0169] The laser scanning path generation module is used to generate simulated laser scanning paths and actual laser scanning paths.

[0170] The simulation module is used to simulate the laser irradiation of the target scar area by the laser source along the simulated laser scanning path, and generate a laser irradiation simulation graphic of the target scar.

[0171] The functional implementation of each module or unit in the above-mentioned laser scanning path planning system for laser scar removal corresponds to the steps in the above-mentioned laser scanning path planning method embodiment for laser scar removal, and their functions and implementation processes will not be repeated here one by one.

[0172] The embodiments of the present application are explained in detail above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments. Even if various changes are made to the present application, if the changes fall within the scope of the present claims and their equivalents, they are still within the protective scope of the present application.

Claims

1. A laser scanning path planning method for laser scar removal, characterized in that: The steps include: S1: Surface morphological data of the target scar are obtained by fusing structured light scanning and speckle imaging, and deep morphological data of the target scar are obtained by optical coherence tomography. Based on these surface and deep morphological data, a three-dimensional structural model of the target scar is constructed. S2: Detecting physiological parameters of target scar using a multimodal sensing array; S3: Inputting the three-dimensional structural model and physiological parameters of the target scar into the pre-trained convolutional neural network, the convolutional neural network outputs the type of the target scar and selects a laser spot distribution pattern of the laser source suitable for the target scar; S4: Setting spatial boundary constraints, using a fast random tree algorithm to generate a simulated laser scanning path suitable for the depth characteristics of the target scar in the 3D structural model of the target scar; S5: In the simulation software, enable the laser source to simulate irradiation of the target scar according to the simulated laser scanning path, generate a laser irradiation simulation graphic of the target scar, verify and re-adjust the simulated laser scanning path based on the laser irradiation simulation graphic, and generate an actual laser scanning path; In S1, the surface morphological data of the target scar is obtained using structured light scanning and speckle imaging, including the following steps: S11: Using a structured light scanner, a dynamic grating is used to generate multiple sets of light beams at different angles and projected onto different areas of the target scar. For the concave areas of the target scar, the operating frequency of the dynamic grating is manually reduced, while for the convex areas of the target scar, the operating frequency of the dynamic grating is manually increased. Using a camera to capture multiple sets of deformation image data of grating patterns on the target scar; S12: Based on the deformation image data of the target scar, on the premise of obtaining the horizontal position information of each pixel point on the surface layer of the target scar, the vertical position information of each pixel point on the surface layer of the target scar is calculated by using a phase shift algorithm, a phase unwrapping algorithm and a triangulation algorithm; S13: generating first three-dimensional point cloud data of the target scar surface based on the horizontal position information and the vertical position information of each pixel point in the target scar, wherein the first three-dimensional point cloud data is used to indicate three-dimensional geometric contour information of the target scar surface; S14: using a coherent laser generator to generate multiple groups of coherent lasers at several different angles to irradiate different areas of the target scar; Using a camera to capture multiple sets of speckle image data reflected from the target scar surface; S15: For the speckle image data of the target scar, the contrast of the speckle image is calculated using the spatial speckle contrast algorithm and the temporal speckle contrast algorithm to evaluate the speckle image quality; Calculate the average size of the speckle image; The speckle image is processed using a complete empirical mode decomposition algorithm, and the texture contours of the surface layers of different regions in the target scar are extracted by combining the contrast and size information of the speckle image to generate second three-dimensional point cloud data of the target scar surface layer. The second three-dimensional point cloud data is used to indicate the roughness information of the target scar surface layer; In the same coordinate system, the first three-dimensional point cloud data and the second three-dimensional point cloud data are superimposed to obtain surface morphological data of the target scar.

2. The laser scanning path planning method for laser scar removal according to claim 1, characterized in that: In S1, optical coherence tomography is used to obtain deep morphological data of the target scar, including the following steps: S16: Using an optical coherence tomography system to generate a broadband near-infrared laser and irradiate different areas of the target scar in a direction perpendicular to the surface of the target scar, and collecting reflected light corresponding to different areas of the target scar, wherein the reflected light generated by the normal skin tissue below the target scar is screened and removed; S17: For the reflected light of the target scar in different areas, the frequency domain signal of the reflected light of the target scar is converted into the depth information of the target scar through Fourier transform, and in the same coordinate system of the first three-dimensional point cloud data and the second three-dimensional point cloud data, a third three-dimensional point cloud data of the deep layer of the target scar is generated, and the third three-dimensional point cloud data is used to indicate the depth information of the deep layer of the target scar.

3. The laser scanning path planning method for laser scar removal according to claim 2, wherein: After constructing the three-dimensional structural model of the target scar in S1, the following steps are further included: S18: Preprocessing the three-dimensional structural model image of the target scar, including: The noise of the three-dimensional structure model image of the target scar is suppressed by using Gaussian filtering algorithm; Use the Canny edge detection algorithm to identify the boundaries of the target scar's concave areas, convex areas, and surface undulations, as well as the boundary between the target scar and the surrounding normal skin tissue; The edge of the target scar image is enhanced by the gradient calculation method of the Sobel operator to highlight the boundary between the target scar and the surrounding normal skin tissue; The surface curvature distribution information of the target scar in three-dimensional coordinates is calculated using the Gaussian curvature algorithm or the least squares surface fitting algorithm.

4. The laser scanning path planning method for laser scar removal according to claim 1, wherein: In S2, a multimodal sensor array is used to detect physiological parameters of a target scar, including the following steps: S21: In the coordinate system of the target scar three-dimensional structural model, infrared thermal imaging data of the target scar region is obtained using an infrared thermal imager. After performing thermal field reconstruction processing on the infrared thermal imaging data, a two-dimensional temperature distribution map of the target scar region in the horizontal direction is calculated according to a temperature measurement formula. The two-dimensional temperature distribution map is used to indicate the thermal conductivity of the skin tissue in the target scar region. S22: In the coordinate system of the target scar three-dimensional structural model, using a humidity sensor to measure and obtain surface moisture content data within the target scar region, and performing Kalman filtering and wavelet noise reduction on the moisture content data. The surface moisture content data is used to indicate the metabolic intensity and hardness of the skin tissue within the target scar region; S23: In the coordinate system of the target scar three-dimensional structural model, a blood oxygen saturation detection device is used to measure and obtain blood oxygen saturation data in the target scar area, and Kalman filtering and wavelet noise reduction processing are performed on the blood oxygen saturation data. The blood oxygen saturation data is used to indicate the blood circulation status and skin tissue metabolic intensity in the target scar area.

5. The laser scanning path planning method for laser scar removal according to claim 1, wherein: The convolutional neural network is configured to perform iterative training based on the three-dimensional structural model and physiological parameters of an existing scar. After training, the convolutional neural network is used to identify the type of the target scar and select the laser spot distribution mode based on the three-dimensional structural model and physiological parameters of the target scar. Target scar types include mature flat scars, hypertrophic scars, keloid scars, atrophic scars, contracture scars, and depressed scars; Laser spot distribution modes include annular spot, grid spot and multi-focal spot. Different laser spot distribution modes are suitable for scar treatment needs of different types, three-dimensional structural models and physiological parameters.

6. The laser scanning path planning method for laser scar removal according to claim 1, characterized in that: The spatial boundary constraints are set in S4. In the three-dimensional structural model of the target scar, a fast random tree algorithm is used to generate a simulated laser scanning path suitable for the depth characteristics of the target scar, including the following steps: S41: Using the boundary line between the target scar and the surrounding normal skin tissue as the spatial boundary constraint for the laser irradiation operation, within the two-dimensional closed boundary area formed by the target scar and the surrounding normal skin tissue, based on the deep morphological data of the target scar and using the depth characteristics of the target scar as a judgment criterion, a number of grid points are generated with different distribution densities from deep to shallow; Initialize the laser scanning path of the laser source. The laser source is configured to face perpendicularly to the target scar surface. Randomly sample a preceding grid point x_rand_1 in the surface plane of the 3D structural model of the target scar. Align the center point of the laser irradiation position of the laser source with the preceding grid point x_rand_1. S42: Calculate the position of any subsequent grid point x_nearest_1 that is closest to the previous grid point x_rand_1; According to the laser spot distribution pattern, the coverage area of ​​a single laser irradiation of the laser source is determined. The diameter or maximum length of the coverage area of ​​a single laser irradiation of the laser source is used as the laser source movement step length. Between the previous grid point x_rand_1 and the next grid point x_nearest_1, several extension nodes x_new are generated with the laser source movement step length as the interval. S43: Based on the three-dimensional structural model of the target scar, determine whether the extended node x_new is in a legal position and whether there are obstacles from adjacent grid points to the extended node, or between adjacent extended nodes. If legal and there are no obstacles, enable the laser source to move along the distribution path of the extended node x_new and irradiate until it reaches the subsequent grid point x_nearest_1; S44: After the laser source reaches the subsequent grid point x_nearest_1, a new scanning path planning is performed between adjacent grid points. The current subsequent grid point x_nearest_1 is used as the new preceding grid point x_rand_2. In the scanning path planning between adjacent grid points, the position of any subsequent grid point x_nearest_2 closest to the preceding grid point x_rand_2 is calculated, excluding the preceding grid point x_rand_1. Steps S42-S43 are repeated. S45: When the laser source irradiation position covers all grid points in the target scar area, a sequence {x_0, x_1, ..., x_n} simulating the laser scanning path is backtracked from the three-dimensional structure model of the target scar.

7. The laser scanning path planning method for laser scar removal according to claim 6, characterized in that: S4 further includes the following steps: S46: Keeping the spatial boundary constraint conditions and the grid point distribution conditions unchanged, repeating steps S42-S45 to generate several groups of simulated laser scanning paths, and selecting the group of simulated laser scanning paths with the least total number of laser source movement steps as the optimal simulated laser scanning path; In addition, the laser source has a maximum number of laser irradiation times for any grid point. When the number of laser irradiation times for any grid point reaches the upper limit, the grid point will no longer be selected as the subsequent grid point x_nearest_n in the subsequent scanning path planning between adjacent grid points.

8. The laser scanning path planning method for laser scar removal according to claim 1, wherein: In S5, a laser irradiation simulation graphic of the target scar is generated in the simulation software. Based on the laser irradiation simulation graphic, the simulated laser scanning path is verified and adjusted again, including the following steps: S51: establishing and running a simulation program in MATLAB or Simulink simulation software to simulate the laser irradiation of the target scar area by the laser source along the simulated laser scanning path; The laser irradiation simulation graphics include a visual laser irradiation movement path in the target scar area, and color markings formed by the total laser heat received by different areas of the target scar; S52: Manually verify whether the laser irradiation area of ​​the laser source completely covers the target scar, and whether the total laser heat received by different areas of the target scar matches the depth characteristics of the target scar.

9. A laser scanning path planning system for laser scar removal, characterized in that: The method for executing the laser scanning path planning method for laser scar removal according to any one of claims 1 to 8 comprises: Scar morphology acquisition module, used to collect surface morphology data and deep morphology data of target scars; A model building module is used to establish a three-dimensional structural model of the target scar; A physiological parameter acquisition module, used to acquire physiological parameters of the target scar; An identification module is used to identify the type of target scar and select a laser spot distribution pattern suitable for the target scar; A laser scanning path generation module is used to generate a simulated laser scanning path and an actual laser scanning path; The simulation module is used to simulate the laser irradiation of the target scar area by the laser source along the simulated laser scanning path, and generate a laser irradiation simulation graphic of the target scar.

Images