Optimization processing method for three-dimensional image

Abstract

The invention discloses an optimization processing method for a three-dimensional image, and relates to the technical field of image optimization. The method comprises the following steps: acquiring modeling acquisition points, acquiring point cloud data and image data on the modeling acquisition points, constructing a color giving model of a three-dimensional object, acquiring acquisition angles, acquisition distances and distortion coefficients of the color giving model at different modeling acquisition points, constructing a corresponding distortion evaluation model, and setting actual acquisition points; acquiring an actual acquisition angle, an actual acquisition distance and an evaluation distortion coefficient of the three-dimensional object, acquiring actual image data on an actual acquisition point, acquiring toning image data in combination with the evaluation distortion coefficient, and performing image optimization on the three-dimensional object; through the technical scheme of the invention, the distortion degree of the image data under different acquisition angles and acquisition distances can be quantized into an accurate numerical value, and the distortion degree of the image data of the three-dimensional object can be reduced while the data processing amount is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of image optimization, and in particular to an optimization processing method for three-dimensional images. Background Art

[0002] With the development of information technology, more and more technical fields have appeared crossover phenomena, such as bionics and image design. By imitating the morphology, structure and function of organisms, three-dimensional image products with unique appearance can be designed, adding more vivid visual effects to three-dimensional image products. In order to combine bionics with image design, how to collect and model the shape data of organisms has become a problem that needs to be solved. In the existing technology, since most organisms are irregular three-dimensional objects, in order to accurately model them, it is necessary to collect data from different angles multiple times, which leads to excessive data processing. Moreover, when collecting image data, different collection angles and collection distances will cause the collected image data to be distorted, resulting in the constructed three-dimensional image being unable to accurately reflect the characteristics of the organism. How to solve these two problems at the same time? The prior art lacks effective means. In view of the shortcomings of the prior art, the present invention provides an optimized processing method for three-dimensional images. Summary of the invention

[0003] The object of the present invention is to provide an optimization processing method for three-dimensional images.

[0004] The object of the present invention can be achieved by the following technical solution: A method for optimizing a three-dimensional image comprises the following steps: Step S1: obtaining the flatness of the surface of the three-dimensional object, selecting a number of points to be collected according to the flatness, selecting a number of modeling collection points from the points to be collected, and obtaining point cloud data and image data of the three-dimensional object at the modeling collection points respectively; Step S2: constructing a three-dimensional model of the three-dimensional object in the digital twin space using the point cloud data, and performing color processing on the three-dimensional model using the image data to obtain a corresponding color model; Step S3: acquiring the acquisition angle and acquisition distance of the color model at different modeling acquisition points and their corresponding distortion coefficients in the digital twin space, and constructing a corresponding distortion evaluation model according to the different acquisition angles and acquisition distances and their distortion coefficients; Step S4: setting a number of actual acquisition points, obtaining the actual acquisition angle and actual acquisition distance of the three-dimensional object, and obtaining the corresponding evaluation distortion coefficient using the distortion evaluation model; Step S5: Acquire actual image data of the three-dimensional object at the actual acquisition point, acquire corresponding toned image data in combination with the estimated distortion coefficient, and optimize the image of the three-dimensional object using the toned image data.

[0005] Furthermore, the process of obtaining the flatness of the surface of the three-dimensional object and selecting a number of points to be collected according to the flatness includes: A laser interferometer is used to emit a laser beam to the surface of a three-dimensional object, and the reflected light formed is monitored. The reflected light is superimposed with the reference light to generate an interference pattern. The flatness of each point on the surface of the three-dimensional object relative to the ideal plane is obtained based on the position, shape and spacing of the fringes in the interference pattern. A flatness threshold is set, the flatness of each point on the surface of the three-dimensional object is compared with the flatness threshold, and each point whose flatness is greater than the flatness threshold is marked as a point to be collected.

[0006] Furthermore, a process of selecting a number of modeling acquisition points from the points to be acquired and respectively acquiring point cloud data and image data of the three-dimensional object at the modeling acquisition points includes: Take any point to be collected as the initial modeling collection point, continuously obtain the probability values ​​of other points to be collected being selected as the next modeling collection point, take the point to be collected corresponding to the maximum probability value as the next modeling collection point, and repeat this step until n modeling collection points are selected; Obtain the collection area of ​​each modeling collection point. In addition to its own modeling collection point, each collection area also contains several other modeling collection points. Use a laser scanner to obtain point cloud data of the corresponding collection area with each modeling collection point as the center. An image acquisition unit is provided, and the image data of the corresponding acquisition area is respectively acquired by the image acquisition unit with each modeling acquisition point as the center.

[0007] Furthermore, the process of constructing a three-dimensional model of the three-dimensional object in the digital twin space using the point cloud data and performing color processing on the three-dimensional model using the image data to obtain the corresponding color model includes: In the digital twin space, the point cloud data of each modeling collection point are integrated into a unified coordinate system, and the point cloud data of each modeling collection point are registered using the registration algorithm; Extract feature information from each point cloud data, fuse each point cloud data to obtain a point cloud data set, and use a mesh generation algorithm to convert the point cloud data set into a continuous three-dimensional model; The coloring process refers to mapping the image data of a single modeling acquisition point to the corresponding acquisition area in the three-dimensional model so that the acquisition area has a pixel value, using the three-dimensional model that has completed the coloring process as the coloring model, and obtaining the coloring models corresponding to each modeling acquisition point respectively.

[0008] Furthermore, the process of respectively obtaining the acquisition angle and acquisition distance of the color model at different modeling acquisition points and their corresponding distortion coefficients in the digital twin space includes: In the color model, the position where the image acquisition unit acquires its image data is taken as the image acquisition point, and the line connecting the image acquisition point and its vertical point on the color model is taken as the acquisition height line; Each other modeling acquisition point in the color model is used as a distortion acquisition point, the line connecting the image acquisition point and each distortion acquisition point is used as an acquisition length line, the angle between the acquisition length line of the distortion acquisition point and the acquisition height line is used as its acquisition angle, and the length of its acquisition length line is used as its acquisition distance; The pixel value of each distorted acquisition point in its own color model is used as the standard pixel value, and the pixel value of each distorted acquisition point in the current color model is used as the distorted pixel value, and the distortion coefficient S of each distorted acquisition point in the current color model is obtained; ; Among them, R1, G1, and B1 are standard pixel values ​​of distorted pixels, and R2, G2, and B2 are distorted pixel values ​​of distorted pixels. The acquisition angle, acquisition distance, and distortion coefficient of all distorted pixels in each color assignment model are obtained respectively.

[0009] Furthermore, the process of constructing a corresponding distortion evaluation model according to different acquisition angles and acquisition distances and their distortion coefficients includes: Generate a distortion evaluation set according to different acquisition angles and acquisition distances and their corresponding distortion coefficients, and divide the obtained distortion evaluation set into a training set and a test set; Construct a convolutional neural network, use different acquisition angles and acquisition distances in the training set as input data of the convolutional neural network, use the corresponding distortion coefficients in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The test set is used to verify the model of the initial convolutional neural network, and the output of the initial convolutional neural network that is less than or equal to the preset test error threshold is used as the distortion assessment model.

[0010] Furthermore, a number of actual acquisition points are set, and the actual acquisition angle and actual acquisition distance of the three-dimensional object are obtained. The process of obtaining the corresponding evaluation distortion coefficient using the distortion evaluation model includes: In actual application scenarios, a number of actual acquisition points are set for a three-dimensional object, wherein the actual acquisition points are not located on the surface of the three-dimensional object, and the acquisition area thereof can cover the entire surface of the three-dimensional object; Taking each actual acquisition point as the center, the actual acquisition angle and actual acquisition distance of each point in its acquisition area are obtained respectively, and the actual acquisition angle and actual acquisition distance are input into the distortion assessment model, and the distortion assessment model is used to output the corresponding evaluation distortion coefficient.

[0011] Furthermore, actual image data of the three-dimensional object is obtained at the actual acquisition point, and corresponding toned image data is obtained in combination with the estimated distortion coefficient. The process of optimizing the image of the three-dimensional object using the toned image data includes: Taking a single actual acquisition point as the center, the image acquisition unit obtains the actual image data in the corresponding acquisition area and the acquisition pixel values ​​of each point inside it, combined with the evaluation distortion coefficient S of each point p , collect pixel values ​​R for each point c , G c , B c Optimize and obtain the optimized pixel value R y , G y , B y ; ; ; ; in, To optimize the amplitude, the optimized pixel value of each point is used as the color-adjusting image data in the acquisition area; Taking each actual acquisition point as the center, the color image data in the corresponding acquisition area is obtained respectively. For a point with only one optimized pixel value, its acquisition pixel value is adjusted to the corresponding optimized pixel value. For a point with multiple optimized pixel values, its acquisition pixel value is adjusted to the optimized pixel value corresponding to the smallest optimization amplitude.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention can use the uneven position of the surface of a three-dimensional object as the object of data collection by acquiring the points to be collected according to the flatness and then acquiring the modeling collection points, which is beneficial to improving the applicability and pertinence of model construction. By respectively constructing a corresponding coloring model for each modeling collection point, acquiring the collection angle and collection distance of each distorted collection point in the coloring model, and acquiring the corresponding distortion coefficient, the distortion degree of the image data under different collection angles and collection distances can be quantified into an accurate value, which is beneficial to obtaining the influence of different collection angles and collection distances on image collection; 2. The present invention constructs a distortion assessment model and sets actual acquisition points in actual application scenarios to obtain the actual acquisition angle and actual acquisition distance of the three-dimensional object. The distortion assessment model is used to obtain the evaluation distortion coefficient, and the actual image data is optimized according to the evaluation distortion coefficient. This is beneficial to reducing the distortion degree of the image data of the three-dimensional object while reducing the amount of data processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0014] like Figure 1 As shown, the optimization processing method for three-dimensional images includes the following steps: Step S1: obtaining the flatness of the surface of the three-dimensional object, selecting a number of points to be collected according to the flatness, selecting a number of modeling collection points from the points to be collected, and obtaining point cloud data and image data of the three-dimensional object at the modeling collection points respectively; Step S2: constructing a three-dimensional model of the three-dimensional object in the digital twin space using the point cloud data, and performing color processing on the three-dimensional model using the image data to obtain a corresponding color model; Step S3: acquiring the acquisition angle and acquisition distance of the color model at different modeling acquisition points and their corresponding distortion coefficients in the digital twin space, and constructing a corresponding distortion evaluation model according to the different acquisition angles and acquisition distances and their distortion coefficients; Step S4: setting a number of actual acquisition points, obtaining the actual acquisition angle and actual acquisition distance of the three-dimensional object, and obtaining the corresponding evaluation distortion coefficient using the distortion evaluation model; Step S5: Acquire actual image data of the three-dimensional object at the actual acquisition point, acquire corresponding toned image data in combination with the estimated distortion coefficient, and optimize the image of the three-dimensional object using the toned image data.

[0015] It should be further explained that, in the specific implementation process, the process of obtaining the flatness of the surface of the three-dimensional object and selecting a number of points to be collected according to the flatness includes: The three-dimensional object is an irregular object with an uneven surface. A laser interferometer is used to emit a laser beam to the surface of the three-dimensional object, and the reflected light formed on the surface is monitored. The reflected light is superimposed with a reference light (a beam of light waves with known flatness) to generate a corresponding interference pattern. In the interference pattern, the bright lines (or dark lines) corresponding to the concave parts will appear first, while the lines corresponding to the convex parts will appear later. According to the parameters such as the position, shape and spacing of the lines in the interference pattern, the degree of deviation of each point on the surface of the three-dimensional object relative to the ideal plane, that is, the flatness, can be obtained; Set a flatness threshold, compare the flatness of each point on the surface of the three-dimensional object with the flatness threshold, mark all points whose flatness is greater than the flatness threshold as points to be collected, and do not perform any other operations on points whose flatness is less than or equal to the flatness threshold.

[0016] It should be further explained that, in the specific implementation process, a process of selecting a number of modeling acquisition points from the points to be acquired and respectively acquiring point cloud data and image data of the three-dimensional object at the modeling acquisition points includes: On the surface of a three-dimensional object, any point to be collected is used as the initial modeling collection point, the shortest distance between other points to be collected and the current modeling collection point is obtained, recorded as d, and the probability value of other points to be collected being selected as the next modeling collection point is obtained, recorded as P; ; The point to be collected corresponding to the maximum probability value is used as the next modeling collection point, and this step is repeated until n modeling collection points are selected to obtain the collection area of ​​each modeling collection point, where the collection area refers to the local area that can be collected on the surface of the three-dimensional object when the modeling collection point is taken as the center; The number of the modeling collection points is less than the number of the points to be collected, and each collection area needs to include at least 3 other modeling collection points in addition to its own modeling collection points, and the laser scanner is used to obtain the point cloud data of the corresponding collection area with each modeling collection point as the center; An image acquisition unit is set up, and the image acquisition unit is used to obtain image data of corresponding acquisition areas respectively with each modeling acquisition point as the center, and the point cloud data and image data of each acquisition area are bound to its corresponding modeling acquisition point.

[0017] It should be further explained that, in the specific implementation process, the process of constructing a three-dimensional model of a three-dimensional object in the digital twin space using point cloud data and coloring the three-dimensional model using image data to obtain a corresponding color model includes: The digital twin space, also known as a digital twin or digital mapping, is used to digitally represent a physical entity. In the digital twin space, the point cloud data of each modeling acquisition point is integrated into a unified coordinate system, and the point cloud data of each modeling acquisition point is registered using a registration algorithm; Extract feature information of three-dimensional objects from various point cloud data, including corner points, edges, surface normals, etc., fuse the various point cloud data after the above processing to obtain a point cloud data set of the three-dimensional object, and use a mesh generation algorithm to convert the point cloud data set into a continuous mesh model, i.e., a three-dimensional model; Using the image data of each modeling acquisition point to perform coloring processing on the corresponding acquisition area on the three-dimensional model, the coloring processing refers to mapping the image data of a single modeling acquisition point to the corresponding acquisition area in the three-dimensional model so that the acquisition area has a corresponding pixel value; The three-dimensional model that has completed the coloring process is used as a coloring model, and the coloring model only contains image data of a single modeling acquisition point. The coloring models corresponding to each modeling acquisition point are obtained respectively, and the coloring models of different modeling acquisition points are also different.

[0018] It should be further explained that, in the specific implementation process, the process of obtaining the acquisition angle and acquisition distance of the color model at different modeling acquisition points and their corresponding distortion coefficients in the digital twin space includes: Taking any color model as an example, the position where the image acquisition unit acquires the image data contained therein is marked as an image acquisition point, and the line between the image acquisition point and its vertical point on the color model is used as an acquisition height line; Since a single acquisition area contains at least 3 other modeling acquisition points, each other modeling acquisition point is marked as a distortion acquisition point, and the connection line between the image acquisition point and each distortion acquisition point is used as the acquisition length line; Taking any distorted acquisition point as an example, the angle (acute angle) between the corresponding acquisition length line and the acquisition height line is marked as the acquisition angle of the distorted acquisition point, and the length of the corresponding acquisition length line is marked as the acquisition distance of the distorted acquisition point; Since each distorted acquisition point has a corresponding color model, the pixel value of each distorted acquisition point in its color model is marked as a standard pixel value, and the pixel value of each distorted acquisition point in the color model is marked as a distorted pixel value, and the distortion coefficient of the distorted acquisition point is obtained, which is recorded as S; ; Among them, R1, G1, and B1 are standard pixel values ​​of the distorted pixels, and R2, G2, and B2 are distorted pixel values ​​of the distorted pixels. This method is used to obtain the acquisition angle and acquisition distance of each distorted pixel point in the color model and its corresponding distortion coefficient, as well as the acquisition angle and acquisition distance of all distorted pixel points in other color models and their corresponding distortion coefficients.

[0019] It should be further explained that, in the specific implementation process, the process of constructing a corresponding distortion evaluation model according to different acquisition angles and acquisition distances and their distortion coefficients includes: Generate a distortion evaluation set according to different acquisition angles and acquisition distances and their corresponding distortion coefficients, and divide the obtained distortion evaluation set into a training set and a test set; Construct a convolutional neural network, use different acquisition angles and acquisition distances in the training set as input data of the convolutional neural network, use the corresponding distortion coefficients in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The test set is used to verify the model of the initial convolutional neural network, and the output of the initial convolutional neural network that is less than or equal to the preset test error threshold is used as the corresponding distortion assessment model.

[0020] It should be further explained that, in the specific implementation process, a number of actual acquisition points are set, and the actual acquisition angle and actual acquisition distance of the three-dimensional object are obtained. The process of obtaining the corresponding evaluation distortion coefficient using the distortion evaluation model includes: In an actual application scenario, a number of actual acquisition points are set for a three-dimensional object, the number of the actual acquisition points is less than the number of modeling acquisition points, the positions of the actual acquisition points are not on the surface of the three-dimensional object, and their acquisition areas can cover the entire surface of the three-dimensional object; Taking each actual collection point as the center, the actual collection angle and actual collection distance of each point in its collection area are respectively obtained, and the actual collection angle and actual collection distance are obtained in the same way as the collection angle and collection distance in the color model; The actual acquisition angle and the actual acquisition distance are input into the distortion evaluation model, and the distortion evaluation model is used to obtain the evaluation distortion coefficient corresponding to the actual acquisition angle and the actual acquisition distance.

[0021] It should be further explained that, in the specific implementation process, the actual image data of the three-dimensional object is obtained at the actual acquisition point, the corresponding color-adjusted image data is obtained in combination with the evaluation distortion coefficient, and the process of optimizing the image of the three-dimensional object using the color-adjusted image data includes: Taking a single actual acquisition point as the center, the image acquisition unit obtains the actual image data in the corresponding acquisition area, and obtains the acquisition pixel value of each point in the corresponding acquisition area according to the actual image data, and combines the evaluation distortion coefficient of each point (denoted as S p ), collect pixel values ​​of each point (denoted as R c , G c , B c ) is optimized and the corresponding optimized pixel value (denoted as R y , G y , B y ); ; ; ; in, is the optimization amplitude, the smaller its value is, the more effective the optimization is, and the optimized pixel value of each point obtained is used as the color adjustment image data in the acquisition area; This method is adopted to take each actual acquisition point as the center and obtain the color image data in the corresponding acquisition area. Since the acquisition areas of different actual acquisition points overlap, a single point may exist in one or more acquisition areas at the same time and have one or more optimized pixel values. For points with only one optimized pixel value, image optimization is achieved by adjusting its collected pixel value to the corresponding optimized pixel value. For points with multiple optimized pixel values, image optimization is achieved by adjusting its collected pixel value to the optimized pixel value corresponding to the smallest optimized amplitude.

[0022] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An optimization processing method for a three-dimensional image, characterized in that: The following steps are involved: Step S1: obtaining the flatness of the surface of the three-dimensional object, selecting a number of points to be collected according to the flatness, selecting a number of modeling collection points from the points to be collected, and obtaining point cloud data and image data of the three-dimensional object at the modeling collection points respectively; Step S2: constructing a three-dimensional model of the three-dimensional object in the digital twin space using the point cloud data, and performing color processing on the three-dimensional model using the image data to obtain a corresponding color model; Step S3: acquiring the acquisition angle and acquisition distance of the color model at different modeling acquisition points and their corresponding distortion coefficients in the digital twin space, and constructing a corresponding distortion evaluation model according to the different acquisition angles and acquisition distances and their distortion coefficients; Step S4: setting a number of actual acquisition points, obtaining the actual acquisition angle and actual acquisition distance of the three-dimensional object, and obtaining the corresponding evaluation distortion coefficient using the distortion evaluation model; Step S5: Acquire actual image data of the three-dimensional object at the actual acquisition point, acquire corresponding toned image data in combination with the estimated distortion coefficient, and optimize the image of the three-dimensional object using the toned image data.

2. The optimization processing method for three-dimensional images according to claim 1, characterized in that: The process of obtaining flatness and selecting points to be collected includes: A laser interferometer is used to emit a laser beam to the surface of a three-dimensional object, and the reflected light formed is monitored. The reflected light is superimposed with the reference light to generate an interference pattern. The flatness of each point on the surface of the three-dimensional object relative to the ideal plane is obtained based on the position, shape and spacing of the fringes in the interference pattern. A flatness threshold is set, the flatness of each point on the surface of the three-dimensional object is compared with the flatness threshold, and each point whose flatness is greater than the flatness threshold is marked as a point to be collected.

3. The optimization processing method for three-dimensional images according to claim 2, characterized in that: The process of selecting modeling collection points and obtaining point cloud data and image data includes: Take any point to be collected as the initial modeling collection point, continuously obtain the probability values ​​of other points to be collected being selected as the next modeling collection point, take the point to be collected corresponding to the maximum probability value as the next modeling collection point, and repeat this step until n modeling collection points are selected; Obtain the collection area of ​​each modeling collection point. In addition to its own modeling collection point, each collection area also contains several other modeling collection points. Use a laser scanner to obtain point cloud data of the corresponding collection area with each modeling collection point as the center. An image acquisition unit is provided, and the image data of the corresponding acquisition area is respectively acquired by the image acquisition unit with each modeling acquisition point as the center.

4. The optimization processing method for three-dimensional images according to claim 3, characterized in that: The process of constructing a three-dimensional model and coloring it to obtain a color-added model includes: In the digital twin space, the point cloud data of each modeling collection point is integrated into a unified coordinate system, and the point cloud data of each modeling collection point is registered using the registration algorithm; Extract feature information from each point cloud data, fuse each point cloud data to obtain a point cloud data set, and use a mesh generation algorithm to convert the point cloud data set into a continuous three-dimensional model; The coloring process refers to mapping the image data of a single modeling acquisition point to the corresponding acquisition area in the three-dimensional model so that the acquisition area has a pixel value, using the three-dimensional model that has completed the coloring process as the coloring model, and obtaining the coloring models corresponding to each modeling acquisition point respectively.

5. The optimization processing method for three-dimensional images according to claim 4, characterized in that: The process of obtaining the acquisition angle, acquisition distance and distortion coefficient at different modeling acquisition points includes: In the color model, the position where the image acquisition unit acquires its image data is taken as the image acquisition point, and the line connecting the image acquisition point and its vertical point on the color model is taken as the acquisition height line; Each other modeling acquisition point in the color model is used as a distortion acquisition point, the line connecting the image acquisition point and each distortion acquisition point is used as an acquisition length line, the angle between the acquisition length line of the distortion acquisition point and the acquisition height line is used as its acquisition angle, and the length of its acquisition length line is used as its acquisition distance; The pixel value of each distorted acquisition point in its own color model is used as the standard pixel value, and the pixel value of each distorted acquisition point in the current color model is used as the distorted pixel value, and the distortion coefficient S of each distorted acquisition point in the current color model is obtained; ； Among them, R1, G1, and B1 are standard pixel values ​​of distorted pixels, and R2, G2, and B2 are distorted pixel values ​​of distorted pixels. The acquisition angle, acquisition distance, and distortion coefficient of all distorted pixels in each color assignment model are obtained respectively.

6. The optimization processing method for three-dimensional images according to claim 5, characterized in that: The process of building a distortion assessment model includes: Generate a distortion evaluation set according to different acquisition angles and acquisition distances and their corresponding distortion coefficients, and divide the obtained distortion evaluation set into a training set and a test set; Construct a convolutional neural network, use different acquisition angles and acquisition distances in the training set as input data of the convolutional neural network, use the corresponding distortion coefficients in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The test set is used to verify the model of the initial convolutional neural network, and the output of the initial convolutional neural network that is less than or equal to the preset test error threshold is used as the distortion assessment model.

7. The optimization processing method for three-dimensional images according to claim 6, characterized in that: The process of obtaining the actual acquisition angle and the actual acquisition distance and evaluating the distortion coefficient includes: In actual application scenarios, a number of actual acquisition points are set for a three-dimensional object, wherein the actual acquisition points are not located on the surface of the three-dimensional object, and the acquisition area thereof can cover the entire surface of the three-dimensional object; Taking each actual acquisition point as the center, the actual acquisition angle and actual acquisition distance of each point in its acquisition area are obtained respectively, and the actual acquisition angle and actual acquisition distance are input into the distortion assessment model, and the distortion assessment model is used to output the corresponding evaluation distortion coefficient.

8. The optimization processing method for three-dimensional images according to claim 7, characterized in that: The process of obtaining actual image data and color-adjusted image data and optimizing the image of a three-dimensional object includes: Taking a single actual acquisition point as the center, the image acquisition unit obtains the actual image data in the corresponding acquisition area and the acquisition pixel values ​​of each point inside it, combined with the evaluation distortion coefficient S of each point p , collect pixel values ​​R for each point c , G c , B c Optimize and obtain the optimized pixel value R y , G y , B y ; ； ； ； in, To optimize the amplitude, the optimized pixel value of each point is used as the color-adjusting image data in the acquisition area; Taking each actual acquisition point as the center, the color image data in the corresponding acquisition area is obtained respectively. For a point with only one optimized pixel value, its acquisition pixel value is adjusted to the corresponding optimized pixel value. For a point with multiple optimized pixel values, its acquisition pixel value is adjusted to the optimized pixel value corresponding to the smallest optimization amplitude.

Images