Engineering progress three-dimensional display system for constructional engineering management

Abstract

The invention discloses an engineering progress three-dimensional display system for building engineering management, relates to the technical field of engineering model display, solves the problem that the image fusion processing process is not comprehensive enough, and greatly reduces the data calculation dimension and simplifies the feature processing flow through image graying processing. While the calculation complexity is reduced, the data processing speed is effectively improved, and an efficient foundation is laid for subsequent three-dimensional modeling; in the feature point calibration link, a unique gray value difference judgment and feature polygon comparison mode is adopted, the influence caused by the difference of the shooting angles of the unmanned aerial vehicle is fully considered, the same feature points are accurately recognized through scaling processing, and compared with a traditional method, the accuracy of feature point calibration is remarkably improved; and the precision of subsequent image fusion and modeling is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering model display, and specifically to a three-dimensional display system for construction project progress management in the construction project management field. Background Art

[0002] In the field of construction project management, accurately and intuitively displaying the project progress is crucial for project planning, monitoring, and decision-making. With the rapid development of unmanned aerial vehicle (UAV) aerial photography technology and three-dimensional modeling technology, using UAVs to capture images for three-dimensional modeling of project progress has become an important means.

[0003] However, there are many problems in the existing technology in this process: First, in terms of image processing, the influence of UAV shooting angles and focal length changes on image features is not fully considered. Traditional feature point extraction and matching algorithms are difficult to accurately identify the same feature points in overlapping images, resulting in errors in subsequent image fusion and affecting the accuracy of the three-dimensional model. Second, in the image fusion stage, there is a lack of an effective alignment mechanism, and the point position deviation caused by shooting differences cannot be eliminated, making the fused images have geometric deformations and reducing the reliability of three-dimensional modeling. Third, the processing of pixel points in the overlapping area is not delicate enough, and the situation of pixel value superposition distortion is likely to occur, further affecting the quality and visualization effect of the three-dimensional model. These problems limit the accuracy and practicality of three-dimensional display of project progress and are difficult to meet the requirements of efficient and accurate progress display in construction project management. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the present invention provides a three-dimensional display system for construction project progress management, which solves the problem that the image fusion processing process is not comprehensive enough.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A three-dimensional display system for construction project progress management includes: An image processing terminal that grayscales the images, confirms the grayscale images, and then sorts a number of associated grayscale images according to the shooting time sequence to generate a grayscale image set. The specific method is as follows: Grayscale the captured images, confirm the grayscale values associated with different points based on the different RGB values associated with different points, and convert the corresponding images into grayscale images based on the confirmed grayscale values; Sort a number of confirmed grayscale images according to the time sequence based on the shooting time associated with different images to confirm the grayscale image set; A same feature point calibration terminal that locks the feature points of the confirmed grayscale image set and confirms the same feature points associated between adjacent grayscale images based on the feature points locked in each grayscale image. The specific method is as follows: Confirm the gray value associated with each pixel in the grayscale image and label it as H i , where i represents different pixels, and perform feature confirmation on each group of pixels: Identify the four groups of pixels associated with the periphery of each group of pixels. The four groups of pixels are located at the four positions above, below, left, and right of this pixel respectively, and label the gray values associated with the four groups of peripheral pixels in sequence as: Z1 i , Z2 i , Z3 i and Z4 i , and use: |Z1 i -H i |, |Z2 i -H i |, |Z3 i -H i |, |Z4 i -H i | to confirm the four groups of feature differences. If all four groups of feature differences are greater than Y1, where Y1 is a preset value, then label this pixel as a feature point; otherwise, do not perform any labeling; Identify the same feature points associated between adjacent grayscale images from the grayscale image set: According to the identified feature points and the four groups of associated feature differences, use this feature point as the center point and the value associated with the feature difference as the unit length to confirm the scalar line associated with the corresponding unit length. The starting point of the scalar line is this center point. Based on the scalar lines associated with the four directions above, below, left, and right of the center point, connect the endpoints of adjacent scalar lines to generate a quadrilateral, and use this quadrilateral as the feature polygon of this feature point; Compare and verify the feature polygons associated with different feature points in adjacent grayscale images, and make the center points of the two groups of feature polygons coincide: If the two feature polygons do not completely coincide, lock the corresponding coincident scalar lines, and from the locked coincident scalar lines, determine the coincident scalar line with the smallest difference in unit length. Based on the current coincident scalar line, extend the scalar line with the smaller unit length until the lengths of the two groups of coincident scalar lines are the same. During the extension process, the other three groups of scalar lines of the corresponding feature polygon are extended synchronously. After the extension is completed, identify the coincident area of the two feature polygons and determine the area ratio of the coincident area. If the area ratio ≥ 98%, then label the two groups of feature points associated with the two feature polygons as the same feature points; otherwise, do not perform any labeling; If the two feature polygons completely coincide, then label the two groups of feature points as the same feature points; The image fusion terminal, based on the same feature points determined between adjacent grayscale images in the grayscale image set, combines the determined same feature points, places the same feature points at the same image position, completes the image fusion of adjacent grayscale images, and generates a preliminary fused image belonging to this grayscale image set. The specific processing method is as follows: Based on the same feature points calibrated in adjacent grayscale images, randomly select a group of the same feature points as the initial points, then select a group of the same feature points as the end points, move from the initial points to the end points to confirm a group of movement vectors, and record the angles and lengths of these movement vectors; It is assumed that there are corresponding movement vectors determined in the previous group of images within adjacent grayscale images. In the subsequent group of images, confirm the same initial points and end points, and determine whether the feature vectors associated with the initial points and end points are consistent with the movement vectors: If they are not consistent, rotate or scale the subsequent group of images until the processed feature vectors are consistent with the movement vectors, then overlap the same feature points associated with the two processed adjacent images, and then sequentially complete the preliminary fusion process of the subsequent two groups of images to obtain a preliminary fused image belonging to this grayscale image set; If they are consistent, overlap the same feature points associated with the two images, and sequentially complete the preliminary fusion process of the two groups of images to obtain a preliminary fused image belonging to this grayscale image set; The point position optimization processing terminal, based on the determined preliminary fused image, optimizes the different grayscale values associated with the same point positions in the overlapping area of the preliminary fused image, confirms the final grayscale value associated with the corresponding point positions, and performs point position optimization processing on the entire preliminary fused image to obtain an optimized fused image. The specific method is as follows: For the overlapping pixel points at the outermost periphery of the overlapping area, perform mean processing on the multiple groups of grayscales associated with the overlapping pixel points to confirm the final grayscale value associated with the corresponding overlapping pixel points; For the overlapping pixel points that do not belong to the outermost periphery of the overlapping area, confirm the grayscale characteristics of the overlapping pixel points in the grayscale image before non - overlapping. Denote the pixel points associated with the corresponding overlapping pixel points in the grayscale image as single - position characteristics. Then, confirm four groups of pixel points in the four directions of up, down, left, and right of the single - position characteristics in the grayscale image, and confirm the pixel differences between the four groups of pixel points and the pixel points associated with this single - position characteristic. If the pixel difference > 0, then sum up the four groups of pixel differences to confirm the characteristic value belonging to the single - position feature; Sequentially confirm the single - position characteristics of a single group of overlapping pixel points located in different grayscale images, and simultaneously confirm the characteristic values of the corresponding single - position features. Then, sequentially perform ratio processing on several groups of characteristic values to confirm the ratio sequence belonging to the multiple groups of characteristic values; According to the confirmed ratio sequence, confirm the weight factor associated with the single-position feature. Denote the ratio associated with the single-position feature as the single-position ratio, and then sum all the ratios in the ratio sequence to confirm the total ratio. The weight factor = single-position ratio ÷ total ratio, and sequentially confirm the weight factors associated with different single-position features; Confirm the gray value associated with each single-position feature, and then combine the weight factors associated with different single-position features. Multiply its gray value by the corresponding weight factor, and then sum the products confirmed by multiple different single-position features to confirm the final gray value associated with this overlapping pixel point; Based on the final gray values associated with different overlapping pixel points in the overlapping area of the preliminary fusion image, perform gray value verification and adjustment on them to obtain an optimized fusion image after optimization processing.

[0006] Preferably, the image modeling end performs three-dimensional modeling based on the optimized fusion image after processing, and displays the generated three-dimensional solid model.

[0007] The present invention provides a three-dimensional display system for construction project management. Compared with the prior art, it has the following beneficial effects: By graying the image, the present invention greatly reduces the data calculation dimension, simplifies the feature processing flow, effectively improves the data processing speed while reducing the calculation complexity, and lays an efficient foundation for subsequent three-dimensional modeling; in the feature point calibration link, a unique gray value difference judgment and feature polygon comparison method is adopted, which not only fully considers the influence brought by the difference in the shooting angle of the drone, but also accurately identifies the same feature points through scaling processing. Compared with the traditional method, the accuracy of feature point calibration is significantly improved, thus ensuring the accuracy of subsequent image fusion and modeling; Based on the matching mechanism of the movement vector and the feature vector of the same feature point, it can adaptively adjust the rotation and scaling of adjacent images to ensure the accurate coincidence of the same feature points, effectively avoiding the position error caused by the changes in the shooting angle and focal length, making the fused image more accurate in geometric position, and providing a high-quality data source for three-dimensional modeling; The whole system is designed from image acquisition, processing to three-dimensional modeling, closely around the needs of construction project management, forming a complete and efficient solution. Through automated and precise processing methods, it effectively reduces manual intervention and errors, improves the efficiency and accuracy of three-dimensional display of project progress, and can provide intuitive and reliable decision-making basis for project managers, having broad application prospects and significant practical value in the field of construction project management. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0009] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0010] Please refer to Figure 1 , this application provides a three-dimensional display system for construction project management, including an image processing terminal, a same feature point calibration terminal, an image fusion terminal, a point position optimization processing terminal, and an image modeling terminal. Among them, the image processing terminal, the same feature point calibration terminal, the image fusion terminal, the point position optimization processing terminal, and the image modeling terminal are electrically connected in sequence from the output node to the input node; For the three-dimensional model of the project progress associated with the corresponding construction project, it is necessary to take images of the relevant construction project by a drone, and based on the specific image features associated with the corresponding images, confirm the image feature points. During the image processing process, the images taken at intervals before and after need to meet an 80% overlap rate. The drone travels according to a preset flight route. After the corresponding flight speed is confirmed, the interval shooting time can be set, so as to ensure that the images taken before and after meet the corresponding overlap rate. During the verification process of the spatial point positions, identify the same feature points, complete the overlapping process between multiple adjacent images, and obtain the images after overlapping verification; Among them, the image processing terminal grayscales the images taken by the drone to confirm the grayscale images associated with the corresponding images. Then, according to the shooting time sequence, sort the associated groups of grayscale images to generate a grayscale image set and transmit it to the same feature point calibration terminal. Specifically, in order to facilitate subsequent actual calculation processes, after grayscaling the corresponding images, the corresponding grayscale values can be quickly confirmed for feature processing. After completing the corresponding three-dimensional modeling process, based on the preset mapping processing relationship, convert the corresponding grayscale images into original images, perform three-dimensional modeling, and perform three-dimensional display. The specific method for detailed processing is as follows: Grayscale the taken images. Based on the different RGB values associated with different points, confirm the grayscale values associated with different points. Based on the confirmed grayscale values, convert the corresponding images into grayscale images. When converting, assign different weights to the RGB values. Based on the weight factor, confirm the grayscale values associated with the corresponding points. The grayscale value = 0.299×R + 0.587×G + 0.114×B; Sort the confirmed groups of grayscale images according to the time sequence of different images to confirm the grayscale image set (there is an overlap area of more than 80% between two adjacent grayscale images).

[0011] Among them, at the same feature point calibration end, the feature points of the confirmed gray image set are locked. Based on the feature points locked in each gray image, the same feature points associated between adjacent gray images are confirmed and marked with the same category. Specifically, because there are many overlapping areas between the two, there are a large number of same feature points in the corresponding overlapping area. Subsequently, based on the corresponding same feature points, image fusion can be performed, which is convenient for subsequent 3D modeling; Among them, the specific method for locking the feature points in the gray image is as follows: Confirm the gray value associated with each pixel point in the gray image and calibrate it as H i , where i represents different pixel points, and feature confirmation is performed on each group of pixel points: Identify the four groups of pixel points associated with the periphery of each group of pixel points (that is, the adjacent four groups of pixel points). The four groups of pixel points are located at the upper, lower, left, and right positions of this pixel point respectively, and the gray values associated with the four groups of pixel points on the periphery are calibrated in sequence as: Z1 i , Z2 i , Z3 i and Z4 i , and use: |Z1 i -H i |, |Z2 i -H i |, |Z3 i -H i |, |Z4 i -H i | to confirm the four groups of feature differences. If all four groups of feature differences are greater than Y1, where Y1 is a preset value, which is determined in advance by the operator according to experience and generally takes a value between 50 and 80, then this pixel point is calibrated as a feature point; otherwise, no calibration is performed; Confirm the same feature points associated between adjacent gray images from the gray image set: According to the confirmed feature points and the four groups of associated feature differences, take this feature point as the center point, and take the value associated with the feature difference as the unit length (if the feature difference is 50, then the unit length is 50), and confirm the scalar line associated with the corresponding unit length (the length unit of the scalar line is determined by the operator. Here, the corresponding unit does not need to be considered, so no calibration is required, and any unit can be determined as its unit length). The starting point of the scalar line is this center point. Based on the scalar lines associated with the four directions of the center point (up, down, left, and right), connect the endpoints of the adjacent scalar lines to generate a quadrilateral, and take this quadrilateral as the feature polygon of this feature point; Compare and verify the feature polygons associated with different feature points in adjacent gray images, and overlap the centers of the two groups of feature polygons: If two characteristic polygons completely coincide, the two sets of characteristic points are designated as the same characteristic points; If two characteristic polygons do not completely coincide, lock the corresponding coincident scalar lines, and from the locked coincident scalar lines, determine the coincident scalar line with the smallest difference in unit length. Based on the current coincident scalar line, extend the scalar line with a smaller unit length until the lengths of the two sets of coincident scalar lines are the same. During the extension process, the other three sets of scalar lines of the corresponding characteristic polygons are extended synchronously. After the extension is completed, identify the coincident area of the two characteristic polygons and determine the area ratio of the coincident area (the area ratio = the area of the coincident area ÷ the total area covered by the two characteristic polygons). If the area ratio ≥ 98%, designate the two sets of characteristic points associated with the two characteristic polygons as the same characteristic points; otherwise, do not perform any designation. Specifically, the reason why there are cases where the characteristics of the same point position are different in the two images is that there are angular differences during the UAV shooting process, which cause characteristic differences. Such characteristic differences will result in different generated characteristic polygons. To identify the misjudgment situations caused by such differences, the corresponding characteristics are scaled to confirm the comprehensive characteristics. Although the pixel points and their surrounding pixel points are affected by the shooting angle, the associated pixel change characteristics should remain unchanged. Therefore, through the process of scaling, the same characteristic points with similar characteristics can be further locked to achieve a more accurate calibration effect.

[0012] Among them, in the image fusion end, based on the same characteristic points determined between adjacent grayscale images in the grayscale image set, combine the determined same characteristic points so that the same characteristic points are placed at the same image position, complete the image fusion of adjacent grayscale images, and generate a preliminary fusion image belonging to this grayscale image set. Specifically, there are the same same characteristic points between each adjacent image, and the corresponding same characteristic points belong to the same point position. During the calibration and scaling process, the images can be scaled in real time to achieve the specific fusion process; Among them, the specific processing method for image fusion is as follows: Based on the same characteristic points calibrated in adjacent grayscale images, randomly select a set of same characteristic points as the starting point, and then select a set of same characteristic points as the ending point. Move from the starting point to the ending point to confirm a set of movement vectors, and record the angle and length of this movement vector; It is proposed that there is a corresponding motion vector in the previous set of images within adjacent grayscale images. The same initial point and end point are confirmed in the subsequent set of images, and it is determined whether the feature vectors associated with the initial point and the end point are consistent with the motion vector. If they are consistent, the corresponding feature points associated in the two images are superimposed in position, and the preliminary fusion process of the two sets of images is completed in sequence to obtain the preliminary fusion image belonging to this grayscale image set. If they are not consistent, the subsequent set of images is rotated or scaled until the processed feature vector is consistent with the motion vector, and then the corresponding feature points associated with the two processed adjacent images are superimposed in position, and the subsequent preliminary fusion process of the two sets of images is completed in sequence to obtain the preliminary fusion image belonging to this grayscale image set. Specifically, there are the same corresponding feature points in both images. Then, when performing front and rear image recognition, it is necessary to perform position verification on the feature vectors between each feature point based on the position differences between the corresponding same feature points, so that the corresponding feature points can be effectively and fully fused, thereby ensuring that the corresponding positions in the two corresponding images can be fully aligned, that is, it will not cause shooting errors due to images related to shooting angles or focal lengths, etc., and can effectively ensure the alignment and fusion effect of the positions of the two images, so as to obtain the corresponding preliminary fusion image. For such images, subsequent corresponding optimization processing is performed to avoid excessive superposition of pixel values between different pixels.

[0013] Among them, for the point position optimization processing end, based on the determined preliminary fusion image, the different grayscale values of the same point positions associated with the overlapping area in the preliminary fusion image are optimized, the final grayscale value associated with the corresponding point positions is confirmed, and the point position optimization processing is performed on the entire preliminary fusion image to obtain the optimized fusion image. The specific method for performing the point position optimization processing is as follows: For the overlapping pixel points at the outermost periphery of the overlapping area, the multiple sets of grayscales associated with the overlapping pixel points are averaged to confirm the final grayscale value associated with the corresponding overlapping pixel points; For the overlapping pixel points that do not belong to the outermost periphery of the overlapping area, the grayscale characteristics of the overlapping pixel points in the grayscale image before non-overlapping are confirmed (that is, the grayscale characteristics of the point positions corresponding to the original grayscale image). The pixel points associated with the corresponding overlapping pixel points in the grayscale image are denoted as single-position characteristics, and then four sets of pixel points in the four directions of up, down, left, and right of the single-position characteristics are confirmed in the grayscale image, and the pixel differences between the four sets of pixel points and the pixel points associated with this single-position characteristic are confirmed. The pixel difference > 0, and then the four sets of pixel differences are summed to confirm the characteristic value belonging to the single-position characteristic; The single-position characteristics of a single set of overlapping pixel points located in different grayscale images are confirmed in sequence, and the characteristic values corresponding to the single-position characteristics are confirmed synchronously. Then, the characteristic values of several groups are processed by ratio in sequence to confirm the ratio sequence belonging to multiple groups of characteristic values. For example, an overlapping pixel point is associated with three single-position characteristics, namely A, B, and C, and the associated characteristic values are 20, 30, and 40 respectively. Then the generated ratio sequence is: 2:3:4; According to the confirmed ratio sequence, the weight factors associated with the corresponding single-position characteristics are confirmed. The ratio associated with the single-position characteristic is denoted as the single-position ratio, and then the sum of all ratios in the ratio sequence is calculated to confirm the total ratio. The weight factor = single-position ratio ÷ total ratio, and the weight factors associated with different single-position characteristics are confirmed in sequence. For example, for point A above, the associated weight factor is (2÷9), that is, two-ninths; The grayscale value associated with each single-position characteristic is confirmed, and then combined with the weight factors associated with different single-position characteristics, the grayscale value is multiplied by the corresponding weight factor, and then the sums of the products confirmed by multiple different single-position characteristics are calculated to confirm the final grayscale value associated with this overlapping pixel point; Based on the final grayscale values associated with different overlapping pixel points in the overlapping area of the preliminary fusion image, the grayscale values are verified and adjusted to obtain the optimized fusion image after optimization processing.

[0014] Among them, on the image modeling side, based on the optimized fusion image after processing, the corresponding RGB values are confirmed by inverse mapping based on the grayscale values after optimized processing at the corresponding positions to obtain the corresponding RGB image; The sparse three-dimensional point cloud is calculated using the principle of triangulation; first, the relative pose relationship between two images is determined through the epipolar geometry relationship, and then the coordinates of the three-dimensional points are calculated according to the coordinates of the feature points in the image and the internal and external parameters of the camera; for multiple images, the initial three-dimensional point cloud structure can be constructed by gradually adding images and performing joint optimization; Based on the sparse three-dimensional point cloud, algorithms such as Poisson reconstruction and moving least squares are used for surface reconstruction to generate a triangular mesh model of the target scene; The generated triangular mesh model is optimized, including operations such as removing redundant vertices and triangles, smoothing the mesh surface, and adjusting the topological structure of the mesh to improve the quality and visualization effect of the model; algorithms such as Laplacian smoothing and Taubin smoothing can be used to reduce the noise and irregularities on the mesh surface; Map the texture information in the original image to the surface of the 3D model to give the model a realistic appearance. By calculating the texture coordinates, map the pixel values in the image to the triangular patches of the 3D model to achieve texture fitting. During the texture mapping process, issues such as texture resolution, stretching, and distortion need to be considered to ensure the quality and authenticity of the texture.

[0015] Since obtaining the specific method of 3D modeling through the corresponding image is relatively common in the prior art, it will not be elaborated here. Specific introductions can be found in related applications such as patent applications (CN109544686B, CN115661360A, CN115661360A), so it will not be elaborated here.

[0016] Some of the data in the above formula are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0017] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A three-dimensional display system for project progress in construction project management, characterized in that, Including: An image processing terminal that grayscales an image, confirms the grayscale image, and then sorts several groups of associated grayscale images according to the shooting time sequence to generate a grayscale image set. An identical feature point calibration terminal that locks the feature points of the confirmed grayscale image set and confirms the identical feature points associated between adjacent grayscale images based on the locked feature points in each grayscale image. An image fusion terminal that combines the identified identical feature points based on the identical feature points determined between adjacent grayscale images in the grayscale image set, places the identical feature points at the same image position, completes the image fusion of adjacent grayscale images, and generates a preliminary fusion image belonging to this grayscale image set. A point position optimization processing terminal that optimizes the different grayscale values of the same point positions associated with the overlapping area in the preliminary fusion image based on the determined preliminary fusion image, confirms the final grayscale value associated with the corresponding point positions, and performs point position optimization processing on the entire preliminary fusion image to obtain an optimized fusion image.

2. The three-dimensional display system for construction project progress management according to claim 1, characterized in that, The specific method for the image processing terminal to generate the grayscale image set is as follows: Grayscale the captured image, confirm the grayscale value associated with different points based on the different RGB values associated with different points, and convert the corresponding image into a grayscale image based on the confirmed grayscale value. Sort the confirmed several groups of grayscale images according to the time sequence based on the shooting time associated with different images to confirm the grayscale image set.

3. A three-dimensional display system for construction project progress management according to claim 1, characterized in that, The specific method for the identical feature point calibration terminal to confirm the identical feature points associated between adjacent grayscale images is as follows: Confirm the gray value associated with each pixel in the grayscale image and label it as H i , where i represents different pixels, and perform feature confirmation on each group of pixels: Identify the four groups of pixel points associated with the periphery of each group of pixel points. The four groups of pixel points are located at the upper, lower, left, and right positions of this pixel point respectively, and sequentially label the gray values associated with the four groups of pixel points on the periphery as: Z1 i , Z2 i , Z3 i and Z4 i , and adopt: |Z1 i -H i |, |Z2 i -H i |, |Z3 i -H i |, |Z4 i -H i | to confirm the four groups of feature differences. If all four groups of feature differences are greater than Y1, where Y1 is a preset value, then label this pixel point as a feature point; otherwise, do not perform any labeling; Identify the identical feature points associated between adjacent grayscale images from the grayscale image set: Taking the identified feature point as the center point and the value associated with the feature difference as the unit length, confirm the scalar line associated with the corresponding unit length. The starting point of the scalar line is this center point. Connect the endpoints of adjacent scalar lines according to the scalar lines associated with the four directions of up, down, left, and right of the center point to generate a quadrilateral, and use this quadrilateral as the feature polygon of this feature point. Compare and verify the feature polygons associated with different feature points in adjacent grayscale images, and overlap the center points of the two groups of feature polygons. If the two feature polygons do not completely overlap, lock the corresponding overlapping scalar lines, and determine the overlapping scalar line with the smallest difference in unit length from the locked overlapping scalar lines. Based on the current overlapping scalar line, extend the scalar line with a smaller unit length until the lengths of the two groups of overlapping scalar lines are the same. During the extension process, the other three groups of scalar lines of the corresponding feature polygons are extended synchronously. After the extension is completed, identify the overlapping area of the two feature polygons and determine the area ratio of the overlapping area. If the area ratio ≥ 98%, calibrate the two groups of feature points associated with the two feature polygons as identical feature points; otherwise, do not perform any calibration.

4. A three-dimensional display system for construction project progress management according to claim 1, characterized in that, If the two feature polygons completely overlap, calibrate the two groups of feature points as identical feature points.

5. A three-dimensional display system for construction project progress management according to claim 1, characterized in that, The specific processing method for the image fusion terminal to perform image fusion is as follows: Based on the same feature points calibrated in adjacent grayscale images, a group of the same feature points is randomly selected as the initial points, and another group of the same feature points is selected as the end points. Starting from the initial points, move towards the end points to confirm a group of motion vectors, and record the angles and lengths of these motion vectors. It is assumed that there are corresponding motion vectors determined in the previous group of images within adjacent grayscale images. In the subsequent group of images, confirm the same initial points and end points, and determine whether the feature vectors associated with the initial points and end points are consistent with the motion vectors: If they are not consistent, rotate or scale the subsequent group of images until the processed feature vectors are consistent with the motion vectors, then stop. Overlap the positions of the same feature points associated with the two processed adjacent images, and then sequentially complete the preliminary fusion process of the subsequent two groups of images to obtain the preliminary fusion image belonging to this grayscale image set.

6. The three-dimensional display system for construction project progress management according to claim 5, characterized in that, If they are consistent, overlap the positions of the same feature points associated with the two images, and sequentially complete the preliminary fusion process of the two groups of images to obtain the preliminary fusion image belonging to this grayscale image set.

7. An engineering progress three-dimensional display system for construction project management according to claim 6, characterized in that, The specific method for performing point position optimization processing at the point position optimization processing end is as follows: For the overlapping pixel points at the outermost periphery of the overlapping area, perform average processing on the multiple groups of grayscales associated with the overlapping pixel points to confirm the final grayscale value associated with the corresponding overlapping pixel points; For the overlapping pixel points that do not belong to the outermost periphery of the overlapping area, confirm the grayscale characteristics of the grayscale image where the corresponding overlapping pixel points are located before non - overlapping. Denote the pixel points associated with the corresponding overlapping pixel points in the grayscale image as single - position characteristics. Then, confirm four groups of pixel points in the four directions of up, down, left, and right of the single - position characteristics in the grayscale image, and confirm the pixel differences between the four groups of pixel points and the pixel points associated with this single - position characteristic. If the pixel difference > 0, then sum up the four groups of pixel differences to confirm the characteristic value belonging to the single - position feature; Sequentially confirm the single - position characteristics of a single group of overlapping pixel points located in different grayscale images, and simultaneously confirm the characteristic values of the corresponding single - position features. Then, sequentially perform ratio processing on several groups of characteristic values to confirm the ratio sequence belonging to the multiple groups of characteristic values; Based on the confirmed ratio sequence, confirm the weight factors associated with the corresponding single - position features. Denote the ratio associated with the single - position feature as the single - position ratio, and then sum up all the ratios in the ratio sequence to confirm the total ratio. The weight factor = single - position ratio ÷ total ratio, and sequentially confirm the weight factors associated with different single - position features; Confirm the grayscale value associated with each single - position characteristic, and then combine the weight factors associated with different single - position characteristics. Multiply its grayscale value by the corresponding weight factor, and then sum up the products confirmed by multiple different single - position characteristics to confirm the final grayscale value associated with this overlapping pixel point; Based on the final grayscale values associated with different overlapping pixel points in the overlapping area of the preliminary fusion image, perform grayscale value verification and adjustment to obtain the optimized fusion image after optimization processing.

8. The three-dimensional display system for construction project progress management according to claim 7, characterized in that The image modeling end performs three - dimensional modeling based on the optimized fusion image after processing, and displays the generated three - dimensional solid model.

Images