Prefabrication yard layout planning modeling method and system based on digitalized auxiliary design

Abstract

The present application relates to the technical field of field layout modeling processing, and particularly relates to a prefabricated field layout planning modeling method and system based on digital auxiliary design. The present application screens multiple construction feature points of each three-dimensional point cloud model; then obtains multiple model regions at a real-time time; according to the number of construction feature points, the position distribution of different points and the initial voxel size in each model region at the real-time time, obtains the adjusted voxel size of each model region at the real-time time, and obtains the three-dimensional adjusted point cloud model at the real-time time; obtains the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model, and obtains the unregistered point region model of the initial three-dimensional point cloud model; splices the three-dimensional adjusted point cloud model at the real-time time and the unregistered point region model, and obtains the prefabricated field layout model at the next time. The present application obtains the accurate offset trend in the construction process, and improves the accuracy of planning modeling.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of field layout modeling processing, and particularly relates to a prefabrication field layout planning modeling method and system based on digital auxiliary design. BACKGROUND

[0002] The prefabrication field layout planning modeling is to reasonably arrange production lines and optimize logistics paths in a limited field through modeling technology, so as to realize efficient, safe and low-cost prefabricated component production. In the prior art, when the digital technology is used for planning modeling, a corresponding three-dimensional model is usually constructed based on BIM technology, so as to guide the specific construction process. However, in the actual construction process, due to the influence of the surrounding environment, the installation coordinates in the prefabrication stage and the model deviation are enlarged due to the difference in segmented settlement of the super-long linear structure in the construction, and there is a difference between the initial modeling model and the construction sequence, and the model precision requirements of different construction stages are different, so that the accuracy of the existing construction planning modeling is poor. SUMMARY

[0003] In order to solve the technical problem that the accuracy of the existing construction planning modeling is poor due to the influence of the environment and the different model precision requirements of different construction stages, the purpose of the present application is to provide a prefabrication field layout planning modeling method and system based on digital auxiliary design, and the technical solution adopted is as follows:

[0004] The present application provides a prefabrication field layout planning modeling method based on digital auxiliary design, which comprises the following steps:

[0005] Obtaining binocular depth images containing a construction area at each time for different monitoring positions in a prefabrication field, and an initial three-dimensional point cloud model;

[0006] Obtaining three-dimensional point cloud models corresponding to the binocular depth images at each time for all monitoring positions; screening a plurality of construction feature points according to the position distribution characteristics of different points in each three-dimensional point cloud model; registering different three-dimensional point cloud models based on the construction feature points, and obtaining a plurality of model regions at a real-time time according to the position change characteristics of the registered points in the three-dimensional point cloud models between different adjacent times;

[0007] According to the number of construction feature points, the position distribution of different points and the initial voxel size in each model region at the real-time time, obtaining the adjusted voxel size of each model region at the real-time time, and obtaining the three-dimensional adjusted point cloud model at the real-time time; according to the position distribution characteristics of different points between the three-dimensional adjusted point cloud model at the real-time time and the initial three-dimensional point cloud model, and the number of construction feature points in all model regions, obtaining the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model, and obtaining the unregistered point region model of the initial three-dimensional point cloud model;

[0008] The three-dimensional adjustment point cloud model and the unregistered point area model at the real-time moment are spliced to obtain a prefabrication yard layout model at the next moment.

[0009] Further, the construction feature point acquisition method comprises:

[0010] According to the position distribution characteristics of different points in each three-dimensional point cloud model, the construction feature contribution degree of each point is obtained.

[0011] If the construction feature contribution degree of a point is greater than a preset contribution threshold, the corresponding point is taken as a construction feature point.

[0012] Further, the construction feature contribution degree acquisition method comprises:

[0013] For each three-dimensional point cloud model, the depth value of each point is obtained.

[0014] The local surface fitting of all points in the neighborhood range of each point in the three-dimensional point cloud model is performed to obtain the maximum principal curvature and normal vector of each point.

[0015] According to the similarity of the normal vector between each point and all other points in the neighborhood range, the difference in the depth value, and the maximum principal curvature of each point, the construction feature contribution degree of each point is obtained, the difference in the depth value and the maximum principal curvature are positively correlated with the construction feature contribution degree, and the similarity of the normal vector is negatively correlated with the construction feature contribution degree.

[0016] Further, the model area acquisition method comprises:

[0017] The relative distance between the registration points in the three-dimensional point cloud model between each moment and the next moment is obtained, and the average of the relative distances of the corresponding registration points between all moments is taken as the matching difference degree of the registration points.

[0018] The ratio of the matching difference degree between each point and other points in the neighborhood range of each point is obtained, if the ratio is greater than a preset matching threshold, the corresponding point is taken as a same-area point, and all same-area points form a model area.

[0019] Further, the adjusted voxel size acquisition method specifically comprises:

[0020] According to the number of construction feature points in each model area at the real-time moment and the position distribution of different points, the high-precision demand degree of each model area is obtained.

[0021] The high-precision demand degree is negatively correlated mapped, and the initial voxel size is gain-adjusted according to the negatively correlated mapping result to obtain the adjusted voxel size of each model area.

[0022] Further, the high-precision demand degree acquisition method comprises:

[0023] The number of construction feature points in each model region at the real-time moment is normalized, and the product of the normalized result and the average of the matching difference degrees in all model regions is calculated as the high-precision demand degree of each model region.

[0024] Further, the method for obtaining the adjusted voxel size comprises:

[0025] The high-precision demand degree is negatively correlated, and the sum of the positive integer 1 and the negatively correlated mapping result is obtained as the weight; the product of the initial voxel size and the weight is calculated as the adjusted voxel size.

[0026] Further, the method for obtaining the construction offset direction vector and the construction offset amount comprises:

[0027] According to the position distribution characteristics of the different points between the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model at the real-time moment, and the number of construction feature points in each model region, the single adjustment amplitude coefficient of each unregistered point in the initial three-dimensional point cloud model is obtained.

[0028] The sum of the coordinate offset vectors of all registered points between the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model at the real-time moment is obtained as the construction offset direction vector.

[0029] The difference between the positive integer 1 and the single adjustment amplitude coefficient is obtained, and the product between the difference and the modulus value of the construction offset direction vector is calculated as the construction offset amount.

[0030] Further, the method for obtaining the single adjustment amplitude coefficient comprises:

[0031] The position coordinate difference between each pair of registered points in the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model is obtained to form a coordinate offset vector.

[0032] For the unregistered points in the initial three-dimensional point cloud model, according to the relative distance between the center of the construction area where each unregistered point is located and the real-time construction position, the high-precision demand degree of all model regions, and the modulus value of the coordinate offset vector of all registered points in the real-time three-dimensional adjustment point cloud model, the single adjustment amplitude coefficient of each unregistered point is obtained, the relative distance and the high-precision demand degree are positively correlated with the single adjustment amplitude coefficient, and the modulus value of the coordinate offset vector is negatively correlated with the single adjustment amplitude coefficient.

[0033] The application further provides a prefabrication yard layout planning modeling system based on digital auxiliary design, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the steps of any one of the prefabrication yard layout planning modeling methods based on digital auxiliary design are implemented.

[0034] The application has the following advantages:

[0035] The application obtains three-dimensional point cloud models corresponding to binocular depth image pairs at each moment at all monitoring positions, eliminating single-view blind areas; according to the position distribution characteristics of different points in each three-dimensional point cloud model, a plurality of construction feature points are screened out to provide stable and significant anchor points for registration and change detection, improving registration accuracy and algorithm efficiency; different three-dimensional point cloud models are registered based on the construction feature points, a plurality of model regions at a real-time moment are obtained according to the position change characteristics of the registered points in the three-dimensional point cloud models between different adjacent moments; the adjustment voxel size of each model region at the real-time moment is obtained according to the number of construction feature points, the position distribution of different points and the initial voxel size in each model region at the real-time moment, and the three-dimensional adjustment point cloud model at the real-time moment is obtained, balancing accuracy and speed, and retaining regions with large shape changes and complex construction structures; the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model are obtained according to the position distribution characteristics of different points between the three-dimensional adjustment point cloud model at the real-time moment and the initial three-dimensional point cloud model, and the number of construction feature points in all model regions, and the unregistered point region model of the initial three-dimensional point cloud model is obtained; the three-dimensional adjustment point cloud model at the real-time moment and the unregistered point region model are spliced to obtain the prefabrication yard layout model at the next moment. The application improves the accuracy of planning modeling by obtaining the accurate offset trend in the construction process. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0037] Figure 1 A flowchart of a prefabrication yard layout planning modeling method based on digital auxiliary design provided by an embodiment of the present application;

[0038] Figure 2 A flowchart of a method for obtaining construction offset direction vectors and construction offset amounts provided by an embodiment of the present application. DETAILED DESCRIPTION

[0039] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the specific implementation, structure, features and effects of a prefabrication yard layout planning modeling method and system based on digital aided design according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0041] The specific scheme of the prefabrication yard layout planning modeling method and system based on digital aided design provided by the present application is described in detail below in combination with the drawings.

[0042] Please refer to Figure 1 which shows a method flowchart of a prefabrication yard layout planning modeling method based on digital aided design according to an embodiment of the present application, which specifically includes:

[0043] Step S1: Obtain binocular depth images containing construction areas at each time point for different monitoring positions in the prefabrication yard, and an initial three-dimensional point cloud model.

[0044] In the embodiment of the present application, considering that different stages of construction have different accuracy requirements for the model, the prefabrication yard model at different stages needs to be analyzed, and the model accuracy of different areas needs to be adjusted. First, monitoring positions are uniformly selected in the construction site, binocular cameras are installed to ensure coverage of the entire construction site, Zhang's calibration method is used to calibrate all binocular cameras, binocular images at each time point for different monitoring positions in the prefabrication yard are obtained, which is beneficial for semantic recognition to identify the construction area in the monocular image, and the binocular images of the construction area corresponding to the monitoring position are registered by ORB to obtain binocular depth images containing depth information. In order to better understand the accuracy of real-time construction, an initial three-dimensional point cloud model of the construction project needs to be obtained according to the initial construction plan CAD or BIM, which is beneficial for comparison and adjustment. The specific semantic recognition and ORB registration are well-known technical means to those skilled in the art, and will not be described here.

[0045] It should be noted that in the embodiment of the present application, the construction stage is monitored every 24 hours as a monitoring period.

[0046] Step S2: obtaining a three-dimensional point cloud model corresponding to a binocular depth image pair at each moment for all monitoring positions; screening a plurality of construction feature points according to position distribution characteristics of different points in each three-dimensional point cloud model; registering different three-dimensional point cloud models based on the construction feature points, and obtaining a plurality of model regions at a real-time moment according to position change characteristics of the registered points in the three-dimensional point cloud models between different adjacent moments.

[0047] In order to realize high-precision, full-time-domain, and full-space-dimension construction state digitalization management and control, a plurality of monitoring positions are analyzed to eliminate a single-view blind area, and a three-dimensional point cloud model corresponding to a binocular depth image at each moment for all monitoring positions is obtained.

[0048] It should be noted that in the embodiments of the present application, the binocular depth images at each moment for all monitoring positions are spliced to obtain a three-dimensional point cloud model at each moment, and specific means are technical means familiar to those skilled in the art, which are not described here.

[0049] In a precast yard construction scene, a precast component is a narrow structure, and the depth of internal points changes obviously, and the geometric characteristics are more obvious, therefore, a plurality of construction feature points are screened according to the position distribution characteristics of different points in each three-dimensional point cloud model.

[0050] Preferably, in an embodiment of the present application, the construction feature point acquisition method comprises:

[0051] According to the position distribution characteristics of different points in each three-dimensional point cloud model, the construction feature contribution degree of each point is obtained.

[0052] Preferably, in an embodiment of the present application, the construction feature contribution degree acquisition method comprises:

[0053] For each three-dimensional point cloud model, the depth value of each point is obtained.

[0054] It should be noted that the depth value is the height distance of a point in the point cloud model from the ground.

[0055] The local surface fitting of all points in the neighborhood range of each point in the three-dimensional point cloud model is performed to obtain the maximum principal curvature and the normal vector of each point.

[0056] According to the similarity of the normal vector between each point and all other points in the neighborhood range, the difference of the depth value, and the maximum principal curvature of each point, the construction feature contribution degree of each point is obtained, the difference of the depth value and the maximum principal curvature are positively correlated with the construction feature contribution degree, and the similarity of the normal vector is negatively correlated with the construction feature contribution degree.

[0057] It should be noted that in an embodiment of the present application, the similarity of the normal vector is reflected by calculating the cosine value of the included angle between the normal vectors, the greater the cosine value of the included angle, the greater the similarity, the closer the direction of the normal vector, and the smaller the construction feature contribution degree; the specific means is a means familiar to those skilled in the art, which will not be repeated here.

[0058] In an embodiment of the present application, the average of the cosine values of the included angles between the normal vectors of each point and all other points in the neighborhood range is obtained as the similarity of the normal vector; the average of the depths of all points in the neighborhood range of each point is obtained as the average depth; the difference between the depth value of each point and the average depth is obtained as the depth difference; the ratio of the maximum principal curvature of each point to the similarity of the normal vector is obtained, the product of the ratio result and the depth difference is calculated, and the product is normalized as the construction feature contribution degree of each point; therefore, the correlation between the depth value difference, the maximum principal curvature, the similarity of the normal vector and the construction feature contribution degree is constructed based on the above mathematical operation, that is, the greater the depth value difference, the greater the maximum principal curvature, the smaller the similarity of the normal vector, the more obvious the change of the geometric feature, and the greater the construction feature contribution degree.

[0059] If the construction feature contribution degree of a certain point is greater than the preset contribution threshold, the corresponding point is taken as a construction feature point.

[0060] It should be noted that in an embodiment of the present application, the size of the preset contribution threshold is 0.5; in other embodiments of the present application, the size of the preset contribution threshold can be set according to specific circumstances, which will not be limited or repeated here.

[0061] In order to align different three-dimensional point cloud models to the same coordinate system for comparison, the different three-dimensional point cloud models are registered based on the construction feature points, and it should be noted that in an embodiment of the present application, the three-dimensional point cloud models are all sampled under the same voxel size, and the construction feature points of the three-dimensional point cloud models are input into ICP for matching, and the specific means is a means familiar to those skilled in the art, which will not be repeated here.

[0062] Considering that the state of the construction stage is different, the contribution degree is inconsistent, and it is necessary to analyze it in different regions, according to the position change characteristics of the registration points in the three-dimensional point cloud models between different adjacent time points, a plurality of model regions at real time are obtained.

[0063] Preferably, in an embodiment of the present application, the method for obtaining the model region comprises:

[0064] The relative distance of the registration points in the three-dimensional point cloud model between each time point and the next time point is obtained, and the average of the relative distances of the corresponding registration points between all time points is obtained as the matching difference degree of the registration points;

[0065] Obtaining the ratio of other points in the neighborhood range of each point and the matching difference between each point, if the ratio is greater than the preset matching threshold, the corresponding point is taken as a same region point; all the same region points form a model region.

[0066] It should be noted that in the embodiments of the present application, the relative distance is calculated by using the Euclidean distance or Manhattan distance, and the specific means is a technology familiar to those skilled in the art, which will not be repeated here.

[0067] It should be noted that in one embodiment of the present application, the size of the preset matching threshold is 0.6; in other embodiments of the present application, the size of the preset matching threshold can be set according to specific circumstances, which will not be limited and repeated here.

[0068] Step S3: According to the number of construction feature points in each model region at the real-time moment, the position distribution of different points and the initial voxel size, the adjusted voxel size of each model region at the real-time moment is obtained, and the three-dimensional adjusted point cloud model at the real-time moment is obtained; according to the position distribution characteristics of different points between the three-dimensional adjusted point cloud model at the real-time moment and the initial three-dimensional point cloud model, and the number of construction feature points in all model regions, the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model are obtained, and the unregistered point region model of the initial three-dimensional point cloud model is obtained.

[0069] During the construction process, the regions that have been constructed and have simple structures have a smaller contribution to the real-time construction, and retaining high precision for the corresponding three-dimensional model will cause data redundancy, while the regions that are unstable in the current construction process and have complex structures should be retained with higher precision, so it is necessary to adjust the voxel size; the points with greater depth and geometric feature changes have more construction feature points, the points with greater depth and geometric feature changes are more, the structure is more complex, the contribution to the construction is greater, and more accurate modeling is needed, the voxel size adjustment is smaller, and the size is closer to the size when the precision is greater. According to the number of construction feature points in each model region at the real-time moment, the position distribution of different points and the initial voxel size, the adjusted voxel size of each model region at the real-time moment is obtained, and the three-dimensional adjusted point cloud model at the real-time moment is obtained.

[0070] Preferably, in one embodiment of the present application, the method for obtaining the adjusted voxel size comprises:

[0071] According to the number of construction feature points in each model region at the real-time moment, and the position distribution of different points, the degree of high precision demand of each model region is obtained.

[0072] Preferably, in one embodiment of the present application, the method for obtaining the degree of high precision demand comprises:

[0073] The number of construction feature points in each model region at the real-time moment is normalized, and the product of the normalized result and the average of the matching difference degree in all model regions is calculated as the high-precision demand degree of each model region.

[0074] It should be noted that the ratio of the number of construction feature points in each model region at the real-time moment to the total number of points is calculated, that is, the number of construction feature points in each model region at the real-time moment is normalized. The more the number of construction feature points, the greater the change in geometric features, and the more the need for high-precision modeling. The closer the distance between points in the region, the smaller the matching difference degree, the smaller the contribution of the stable region to the real-time construction, and the smaller the high-precision demand degree. In other embodiments of the present application, normalization can also be performed by linear normalization or a normalization function. The specific means are well-known to those skilled in the art, and will not be repeated here.

[0075] The high-precision demand degree is negatively correlated, and the initial voxel size is adjusted according to the negative correlation mapping result to obtain the adjusted voxel size of each model region.

[0076] It should be noted that the greater the high-precision demand degree, the greater the number of construction feature points, the more obvious the change in geometric features and depth of the model region, the greater the contribution to the construction, and the greater the need to maintain the modeling precision of the model region. The smaller the degree of increase in the adjusted voxel size, the closer to the initial voxel size.

[0077] It should be noted that in one embodiment of the present application, the reciprocal of the high-precision demand degree is negatively correlated, and in other embodiments of the present application, the reciprocal of the high-precision demand degree can be negatively correlated by an exponential function with a natural constant as the base. The specific means are well-known to those skilled in the art, and will not be limited or repeated here.

[0078] In one embodiment of the present application, the sum of the positive integer 1 and the negative correlation mapping result is obtained as the weight, and the product of the initial voxel size and the weight is calculated as the adjusted voxel size.

[0079] It should be noted that in the embodiments of the present application, the initial voxel size is 0.03mm.

[0080] It should be noted that in the embodiments of the present application, the relative distance can be obtained by existing distance calculation methods such as Euclidean distance or Manhattan distance. The specific means are well-known to those skilled in the art, and will not be repeated here.

[0081] Based on the adjusted voxel size of each model region, the three-dimensional point cloud model at the real-time moment is sampled to obtain a three-dimensional adjusted point cloud model at the real-time moment.

[0082] In actual construction process, due to the influence of external factors, such as uneven settlement or construction vibration interference, there will be certain error between the actual construction and the preset initial three-dimensional model, and the next time the prefabrication yard needs to be adjusted, so the unregistered area is adjusted based on the offset direction and offset degree of the constructed part. According to the position distribution characteristics of the different points between the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model, and the number of construction feature points in all model areas, the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model are obtained, and the unregistered point area model of the initial three-dimensional point cloud model is obtained.

[0083] Preferably, in an embodiment of the present application, the method for obtaining the construction offset direction vector and the construction offset amount is as follows: Figure 2 , which shows a flow chart of a method for obtaining a construction offset direction vector and a construction offset amount, comprising:

[0084] Step S201: According to the position distribution characteristics of the different points between the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model, and the number of construction feature points in each model area, the single adjustment amplitude coefficient of each unregistered point in the initial three-dimensional point cloud model is obtained.

[0085] Preferably, in an embodiment of the present application, the method for obtaining the single adjustment amplitude coefficient comprises:

[0086] The position coordinate difference of each pair of registered points between the real-time three-dimensional adjustment point cloud model and the initial three-dimensional point cloud model is obtained, and a coordinate offset vector is formed;

[0087] For the unregistered points in the initial three-dimensional point cloud model, according to the relative distance between the center of the construction area where each unregistered point is located and the real-time construction position, the high-precision demand degree of all model areas, and the coordinate offset vector modulus value of all registered points in the three-dimensional adjustment point cloud model at the real-time moment, the single adjustment amplitude coefficient of each unregistered point is obtained. The relative distance and the high-precision demand degree are positively correlated with the single adjustment amplitude coefficient, and the coordinate offset vector modulus value is negatively correlated with the single adjustment amplitude coefficient.

[0088] It should be noted that the greater the relative distance, the less susceptible to unstable settlement interference, the greater the degree of adjustment, the greater the high-precision demand degree, the more unstable the construction area, the greater the need for adjustment, and the greater the single adjustment amplitude coefficient; the smaller the offset degree, the closer to the initial model, and the smaller the adjustment amplitude.

[0089] In an embodiment of the present application, the relative distance between the center of the construction area where each point is located and the real-time construction position is obtained; the mean value of the high-precision requirement degree of all model areas is obtained; the cumulative value of the coordinate offset vector modulus of all registration points in the three-dimensional adjusted point cloud model at the real-time time is obtained; the product between the relative distance and the mean value of the high-precision requirement degree is obtained, the ratio of the product result and the cumulative value of the coordinate offset vector modulus is calculated, and normalization is performed, which is taken as the single adjustment amplitude coefficient. Therefore, the correlation between the relative distance, the high-precision requirement degree, the cumulative value and the single amplitude coefficient is constructed based on the above basic mathematical operations, that is, the larger the relative distance, the larger the high-precision requirement degree, the smaller the cumulative value, and the larger the single amplitude coefficient.

[0090] Step S202: obtaining the sum of the coordinate offset vectors of all registration points between the three-dimensional adjusted point cloud model and the initial three-dimensional point cloud model at the real-time time as the construction offset direction vector.

[0091] By taking the mean value, the overall trend of the coordinate offset vectors of all registration points can be quantified.

[0092] Step S203: obtaining the difference between the positive integer 1 and the single adjustment amplitude coefficient, calculating the product between the difference and the modulus value of the construction offset direction vector as the construction offset amount.

[0093] Based on this, by obtaining the construction offset direction vector and the construction offset amount of each unregistered point in the initial three-dimensional point cloud model, starting from the initial coordinates of each unregistered point in the initial three-dimensional point cloud model, moving the construction offset amount in the direction of the construction offset direction vector, the new coordinates of each unregistered point are obtained, and the unregistered point area model of the initial three-dimensional point cloud model is constructed.

[0094] Step S4: splicing the three-dimensional adjusted point cloud model at the real-time time and the unregistered point area model to obtain the precast yard layout model at the next time.

[0095] By fusing the real-time construction point cloud and the unregistered area data, the precast yard layout model is dynamically updated to reflect the real-time state of component installation, site occupation, etc., optimize component production, stacking and transportation path planning, and improve the accuracy of the yard layout planning modeling.

[0096] To sum up, the application screens a plurality of construction feature points of each three-dimensional point cloud model, then obtains a plurality of model regions at a real-time moment, obtains an adjusted voxel size of each model region at the real-time moment according to the number of construction feature points in each model region at the real-time moment, the position distribution of different points and the initial voxel size, obtains a three-dimensional adjusted point cloud model at the real-time moment, obtains a construction offset direction vector and a construction offset amount of each unregistered point in the initial three-dimensional point cloud model, and obtains an unregistered point region model of the initial three-dimensional point cloud model, splices the three-dimensional adjusted point cloud model at the real-time moment and the unregistered point region model to obtain a precast yard layout model at the next moment. The application obtains an accurate offset trend in the construction process, and improves the accuracy of planning modeling.

[0097] The application further provides a precast yard layout planning modeling system based on digital auxiliary design, which comprises a memory, a processor, a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of any one of the precast yard layout planning modeling methods based on digital auxiliary design when executing the computer program.

[0098] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0099] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the differences from other embodiments.

Claims

1. A prefabrication yard layout planning and modeling method based on digital-aided design, characterized in that, The method includes: Acquire binocular depth images of the construction area at different monitoring locations in the prefabrication yard at each time point, as well as an initial 3D point cloud model; Obtain 3D point cloud models corresponding to the binocular depth images of all monitoring locations at each time step; select multiple construction feature points based on the positional distribution characteristics of different points in each 3D point cloud model; register different 3D point cloud models based on the construction feature points; and obtain multiple model regions at the real time step based on the positional change characteristics of the registration points in the 3D point cloud models between different adjacent time steps. Based on the number of construction feature points in each model region at real time, the positional distribution of different points, and the initial voxel size, the adjusted voxel size of each model region at real time is obtained, and the 3D adjusted point cloud model at real time is obtained; based on the positional distribution characteristics of different points between the 3D adjusted point cloud model at real time and the initial 3D point cloud model, and the number of construction feature points in all model regions, the construction offset direction vector and construction offset of each unregistered point in the initial 3D point cloud model are obtained, and the unregistered point region model of the initial 3D point cloud model is obtained. The three-dimensional adjusted point cloud model and the unregistered point area model at real time are stitched together to obtain the prefabrication field layout model at the next time step. The method for obtaining the model region includes: The relative distance between the registration points in the 3D point cloud model between each time step and the next time step is obtained. The average relative distance between the corresponding registration points between all time steps is obtained as the degree of matching difference of the registration points. Obtain the ratio of the degree of matching difference between each point and other points within the neighborhood of each point. If the ratio is greater than the preset matching threshold, the corresponding point is regarded as a point in the same region; all points in the same region are combined into a model region. The method for obtaining the adjusted voxel size specifically includes: Based on the number of construction feature points in each model area at real time and the location distribution of different points, the high precision requirement of each model area is obtained; A negative correlation mapping is performed on the degree of high precision requirement, and the initial voxel size is adjusted according to the negative correlation mapping result to obtain the adjusted voxel size of each model region. The method for obtaining the required level of high precision includes: The number of construction feature points in each model region at real time is normalized, and the product of the normalization result and the mean of the degree of matching difference in all model regions is calculated as the high precision requirement of each model region.

2. The method for prefabrication yard layout planning and modeling based on digital-aided design according to claim 1, characterized in that, The method for obtaining the construction feature points includes: Based on the positional distribution characteristics of different points in each 3D point cloud model, the construction feature contribution of each point is obtained; If the contribution of a construction feature at a certain point is greater than the preset contribution threshold, the corresponding point will be used as a construction feature point.

3. The method for prefabrication yard layout planning and modeling based on digital-aided design according to claim 2, characterized in that, The method for obtaining the contribution of construction features includes: For each 3D point cloud model, obtain the depth value of each point; Local surface fitting is performed on all points within the neighborhood of each point in the 3D point cloud model to obtain the maximum principal curvature and normal vector of each point. The construction feature contribution of each point is obtained based on the similarity of the normal vectors between each point and all other points in the neighborhood, the difference in depth values, and the maximum principal curvature of each point. The difference in depth values ​​and the maximum principal curvature are positively correlated with the construction feature contribution, while the similarity of the normal vectors is negatively correlated with the construction feature contribution.

4. The method for prefabrication yard layout planning and modeling based on digital-aided design according to claim 1, characterized in that, The method for obtaining the adjusted voxel size includes: A negative correlation mapping is performed on the degree of high precision requirement, and the sum of the positive integer 1 and the negative correlation mapping result is obtained as the weight; the product of the initial voxel size and the weight is calculated as the adjusted voxel size.

5. The method for prefabrication yard layout planning and modeling based on digital-aided design according to claim 1, characterized in that, The methods for obtaining the construction offset direction vector and the construction offset amount include: Based on the positional distribution characteristics of different points between the real-time adjusted 3D point cloud model and the initial 3D point cloud model, and the number of construction feature points in each model region, the single adjustment amplitude coefficient of each unregistered point in the initial 3D point cloud model is obtained. The sum of the coordinate offset vectors of all registration points between the real-time adjusted 3D point cloud model and the initial 3D point cloud model is obtained and used as the construction offset direction vector. Obtain the difference between the positive integer 1 and the single adjustment amplitude coefficient, calculate the product between the difference and the magnitude of the construction offset direction vector, and use it as the construction offset amount.

6. The method for prefabrication yard layout planning and modeling based on digital-aided design according to claim 5, characterized in that, The method for obtaining the single adjustment amplitude coefficient includes: The position coordinate difference of each pair of registration points between the real-time adjusted 3D point cloud model and the initial 3D point cloud model is obtained to form a coordinate offset vector. For unregistered points in the initial 3D point cloud model, the single adjustment amplitude coefficient of each unregistered point is obtained based on the relative distance between the center of the construction area where each unregistered point is located and the real-time construction position, the high precision requirement of all model areas, and the coordinate offset vector magnitude of all registered points in the 3D adjustment point cloud model at the real time. The relative distance and the high precision requirement are positively correlated with the single adjustment amplitude coefficient, while the coordinate offset vector magnitude is negatively correlated with the single adjustment amplitude coefficient.

7. A prefabrication yard layout planning and modeling system based on digital-aided design, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the prefabrication yard layout planning and modeling method as described in any one of claims 1 to 6.

Images