Method and system for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning
By converting historical defect data into sequences and analyzing and planning defect sequences, building three-dimensional detection timing and dividing detection units, the problem of large amount of data in automatic detection of hyperbolic curtain wall units is solved, and efficient defect detection and category recognition are achieved.
Patent Information
- Application Number
- CN202411078616.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-07
AI Technical Summary
During the existing hyperbolic curtain wall unit automated detection process based on 3D scanning technology, the amount of data is large and the model defect type cannot be effectively identified, resulting in insufficiency of detection.
By collecting historical defect data and converting it into historical defect sequences, analyzing the planned defect sequence of the current hyperbolic curtain wall, building a three-dimensional detection timing, dividing the three-dimensional detection units, and performing model defect detection on these units, and identifying defect categories in combination with neural network models.
The data processing volume of defect detection is reduced, the accuracy and efficiency of defect category identification is improved, and the data processing volume is reduced.
Smart Images

Figure CN118604005B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of 3D scanning, and in particular to a method and system for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning. Background Art
[0002] 3D scanning technology refers to a high-tech technology that integrates optical, mechanical, electrical and computer technologies. It is mainly used to scan the spatial shape, structure and color of an object to obtain the spatial coordinates of the object's surface. A hyperbolic curtain wall refers to a curtain wall with curved surfaces on both sides. The purpose of unit automation detection is to detect defects in each curtain wall unit during the curtain wall manufacturing process, so as to repair the problematic curtain wall units in time. The defects of curtain wall units include scratches, dents, protrusions, etc. In addition, when assembling and splicing multiple curtain walls, it is also necessary to consider whether the joints between the spliced curtain walls are too large.
[0003] At present, the process of realizing unit automatic detection of hyperbolic curtain walls based on 3D scanning mainly uses 3D scanning technology to scan the outer surface of the manufactured hyperbolic curtain wall, thereby generating a three-dimensional point cloud model of the hyperbolic curtain wall, calculating the similarity between the three-dimensional point cloud model and the real template, and detecting whether the hyperbolic curtain wall has defects. In this process, if the hyperbolic curtain wall is large in size, the amount of data consumed in the defect detection process of the hyperbolic curtain wall will be relatively large, because it is necessary to compare the similarity between the coordinate points on the three-dimensional point cloud model and the coordinate points on the real template one by one, the workload of coordinate comparison is large, and the type of model defects cannot be detected. Image recognition technology is also needed to realize the defect category of the model surface. Therefore, a solution is urgently needed to reduce the amount of data processing for defect detection and realize the identification of defect categories. Summary of the invention
[0004] In order to solve the above problems, the present invention provides a method and system for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning, which can reduce the data processing amount of defect detection and realize the identification of defect categories.
[0005] In a first aspect, the present invention provides a method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning, comprising:
[0006] After emitting a detection laser to the current hyperbolic curtain wall, collecting the three-dimensional point cloud coordinates of the current hyperbolic curtain wall, and generating a three-dimensional point cloud model of the current hyperbolic curtain wall using the three-dimensional point cloud coordinates;
[0007] Collecting historical defect data of historical hyperbolic curtain walls, converting the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, and using the historical defect sequence to analyze a planned defect sequence corresponding to the current hyperbolic curtain wall;
[0008] querying a planning defect time of the planning defect sequence, constructing a three-dimensional detection time sequence of the three-dimensional point cloud model using the planning defect time, and dividing the three-dimensional detection units of the three-dimensional point cloud model using the planning defect sequence;
[0009] Based on the three-dimensional detection sequence, performing model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit to obtain a model defect detection result, and correcting the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall;
[0010] The modified hyperbolic curtain wall is assembled to obtain an assembled hyperbolic curtain wall, an assembly defect detection is performed on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and the model defect detection result and the assembly defect detection result are used as the unit automatic detection result of the current hyperbolic curtain wall.
[0011] In a possible implementation manner of the first aspect, collecting the three-dimensional point cloud coordinates of the current hyperbolic curtain wall includes:
[0012] The source-to-wall distance between the laser emission source of the detection laser and the current hyperbolic curtain wall is calculated using the following formula:
[0013]
[0014] in, represents the source-wall distance, It represents the time duration from the time when the laser emission source emits the detection laser to the time when the reflected laser of the current hyperbolic curtain wall is received. represents the emission speed of the detection laser;
[0015] Based on the source-wall distance, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are calculated using the following formula:
[0016]
[0017]
[0018]
[0019] in, Represents the coordinates of the 3D point cloud, represents the source-wall distance, represents the emission angle of the detection laser in the vertical direction, Indicates the emission angle of the detection laser on the Y axis.
[0020] In a possible implementation of the first aspect, collecting historical defect data of historical hyperbolic curtain walls includes:
[0021] respectively collecting the historical defect categories and historical defect locations of the historical hyperbolic curtain wall;
[0022] The defect penalty intensity of the historical hyperbolic curtain wall is calculated using the following formula:
[0023]
[0024] in, Indicates the historical defect category that occurred at the kth moment The penalty for defects caused by the responsible party of the hyperbolic curtain wall is Indicates the amount of fines imposed on the party responsible for the hyperbolic curtain wall. Indicates the duration of the fine imposed on the responsible party of the hyperbolic curtain wall. Indicates the job demotion caused to the party responsible for the hyperbolic curtain wall. , , express , , The weight of
[0025] The defect correction cost of the historic hyperbolic curtain wall was calculated using the following formula:
[0026]
[0027] in, Indicates the historical defect category that occurred at the kth moment The defect correction cost of the responsible party of the hyperbolic curtain wall is It represents the cost of the responsible party to correct the defects of the double-curved curtain wall. Indicates the time it takes for the responsible party to correct the defects of the hyperbolic curtain wall. Indicates the number of personnel dispatched by the responsible party to correct defects in the hyperbolic curtain wall. , , express , , The weight of
[0028] The scale of the defect consequences of the historical hyperbolic curtain wall is calculated using the following formula:
[0029]
[0030] in, Indicates the historical defect category that occurred at the kth moment The scale of the consequences of the defects caused by the hyperbolic curtain wall to the user, Indicates the amount of loss caused by the hyperbolic curtain wall to the user. Indicates the time loss caused by the hyperbolic curtain wall to the user. Indicates the loss of personnel caused by the hyperbolic curtain wall to the user. , , express , , The weight of
[0031] The historical defect category, the historical defect location, the defect penalty intensity, the defect correction cost and the defect consequence scale are used as the historical defect data of the historical hyperbolic curtain wall.
[0032] In a possible implementation manner of the first aspect, converting the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data includes:
[0033] Calculating a data standard deviation of the historical defect data based on the historical defect moments of the historical defect data;
[0034] constructing an absolute value threshold of the historical defect data using the data standard deviation;
[0035] Calculating an absolute value of a difference between the historical defect data and a mean value of the historical defect data;
[0036] When the absolute value of the difference is greater than the absolute value threshold, removing the target defect data corresponding to the absolute value of the difference in the historical defect data to obtain remaining defect data;
[0037] Detecting whether the remaining defect data has data missing;
[0038] When there are missing data in the remaining defective data, the following formula is used to interpolate the missing data in the remaining defective data to obtain interpolated defective data:
[0039]
[0040] in, represents interpolation defect data, express The remaining defect data at the time, express Remaining defect data at the moment;
[0041] Inserting the interpolation defect data into the remaining defect data to obtain the inserted defect data;
[0042] Based on the historical defect time of the historical defect data, the historical defect sequence of the inserted defect data is constructed using the following formula:
[0043]
[0044] in, express The historical defect sequence of row 5 and column 5, represents the sequence of historical defect categories in the inserted defect data, represents the total number of historical defect moments, The serial number of the historical defect moment, represents the sequence of historical defect positions in the inserted defect data, A sequence representing the magnitude of the consequence of defects in the inserted defect data.
[0045] In a possible implementation manner of the first aspect, analyzing the planned defect sequence corresponding to the current hyperbolic curtain wall by using the historical defect sequence includes:
[0046] Inputting the historical defect sequence into a preset defect analysis model, so as to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall through the defect analysis model;
[0047] Wherein, the defect analysis model includes:
[0048]
[0049]
[0050]
[0051]
[0052]
[0053]
[0054] in, represents the planning defect sequence, Indicates the current time The defect analysis model receives the historical defect sequence. Indicates the last moment The planning defect sequence under represents the output result of the input gate, represents the output of the forget gate, represents the output of the memory gate, Represents the output result of the output gate, Indicates the current time The unit status under Indicates the last moment The unit status under , , , represents the weight parameters and bias parameters of the input gate, , , , represents the weight parameter and bias parameter of the forget gate, , , , represents the weight parameters and bias parameters of the memory gate, , , , Represents the weight parameters and bias parameters of the output gate.
[0055] In a possible implementation of the first aspect, the three-dimensional detection timing of constructing the three-dimensional point cloud model by using the planned defect moment includes:
[0056] The following formula is used to calculate the number of moments added to the planning defect moment:
[0057]
[0058]
[0059] in, Indicates the number of moments increasing. represents the total number of moments in the jth planning defect position among the planning defect moments, express The total number of moments in the corresponding real scene, It represents the total number of moments after the total number of moments under the j-th planning defect position is increased;
[0060] Based on the time increase number, the defect time is increased at the target time position of the planned defect time to obtain the three-dimensional detection timing.
[0061] In a possible implementation manner of the first aspect, the three-dimensional detection unit that divides the three-dimensional point cloud model by using the planned defect sequence includes:
[0062] Based on the three-dimensional detection time sequence, adding defect positions to the planned defect sequence to obtain a three-dimensional defect sequence;
[0063] Determining whether there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence;
[0064] When there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence, the continuous defect positions are divided into the three-dimensional detection units based on the three-dimensional detection timing corresponding to the three-dimensional defect sequence.
[0065] In a possible implementation manner of the first aspect, performing model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection time sequence to obtain a model defect detection result includes:
[0066] Based on the three-dimensional detection timing, calculating the unit error between the coordinate point on the three-dimensional detection unit and the coordinate point on the preset standard unit template;
[0067] Calculating the error standard deviation of the unit error;
[0068] calculating an error-mean difference between the unit error and a mean of the unit errors;
[0069] Using the difference between the error standard deviation and the error-mean value to identify whether there is a model defect on the three-dimensional detection unit;
[0070] When there is a model defect on the three-dimensional detection unit, identifying a type of the model defect on the three-dimensional detection unit;
[0071] Using the model defect category as the model defect detection result;
[0072] When there is no model defect on the three-dimensional detection unit, the absence of the model defect on the three-dimensional detection unit is used as the model defect detection result.
[0073] In a possible implementation of the first aspect, performing assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result includes:
[0074] Obtaining a three-dimensional scanning model of the assembled hyperbolic curtain wall;
[0075] identifying a hyperbolic curtain wall boundary of the assembled hyperbolic curtain wall from the three-dimensional scanned model;
[0076] Querying the neighboring hyperbolic curtain wall boundaries of the hyperbolic curtain wall boundary;
[0077] Calculating a boundary error between the hyperbolic curtain wall boundary and the adjacent hyperbolic curtain wall boundary;
[0078] The boundary error is used to perform assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result.
[0079] In a second aspect, the present invention provides a unit automatic detection system for a hyperbolic curtain wall based on 3D scanning, the system comprising:
[0080] A model generation module, configured to collect three-dimensional point cloud coordinates of the current hyperbolic curtain wall after emitting a detection laser to the current hyperbolic curtain wall, and generate a three-dimensional point cloud model of the current hyperbolic curtain wall using the three-dimensional point cloud coordinates;
[0081] A sequence analysis module, used for collecting historical defect data of historical hyperbolic curtain walls, converting the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, and using the historical defect sequence to analyze a planned defect sequence corresponding to the current hyperbolic curtain wall;
[0082] a unit division module, used to query the planning defect time of the planning defect sequence, construct the three-dimensional detection time sequence of the three-dimensional point cloud model by using the planning defect time, and divide the three-dimensional detection unit of the three-dimensional point cloud model by using the planning defect sequence;
[0083] A curtain wall correction module, configured to perform model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection sequence to obtain a model defect detection result, and to correct the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall;
[0084] An assembly detection module is used to assemble the modified hyperbolic curtain wall to obtain an assembled hyperbolic curtain wall, perform assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and use the model defect detection result and the assembly defect detection result as a unit automatic detection result of the current hyperbolic curtain wall.
[0085] Compared with the prior art, the technical principle and beneficial effects of this solution are:
[0086] The embodiment of the present invention collects historical defect data of historical hyperbolic curtain walls to use the historical defect data as an influencing factor for the occurrence of future defects. Furthermore, the embodiment of the present invention converts the historical defect data into a historical defect sequence based on the historical defect moment of the historical defect data to convert the historical defect data into a time series sequence, so as to facilitate the subsequent neural network model to identify the features of the historical defect data. Furthermore, the embodiment of the present invention uses the historical defect sequence to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall to analyze at which positions defects of the hyperbolic curtain wall will occur in the future period. The embodiment of the present invention uses the planned defect moment to construct a three-dimensional point cloud model. The detection sequence is used to calculate the probability of the model prediction result occurring in the actual scene, and the probability is used to reconstruct the planning defect moment, thereby increasing the accuracy of the predicted planning moment. Furthermore, the embodiment of the present invention divides the three-dimensional detection unit of the three-dimensional point cloud model by using the planning defect sequence, so as to divide the continuous coordinates in the predicted planning defect sequence into the same area, and in the subsequent defect detection process, only the three-dimensional detection unit on the surface of the hyperbolic curtain wall is detected, and other positions on the hyperbolic curtain wall are not detected, which greatly reduces the amount of data processing. The embodiment of the present invention performs model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection sequence to detect the defect category. Therefore, the embodiment of the present invention proposes a method and system for realizing automatic detection of units of hyperbolic curtain walls based on 3D scanning, which can reduce the amount of data processing for defect detection and realize the identification of defect categories. BRIEF DESCRIPTION OF THE DRAWINGS
[0087] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0088] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0089] Figure 1 A schematic flow chart of a method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning provided by one embodiment of the present invention;
[0090] Figure 2 A schematic diagram of a module of a unit automatic detection system for a hyperbolic curtain wall based on 3D scanning provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0091] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.
[0092] The embodiment of the present invention provides a method for realizing unit automation detection of hyperbolic curtain walls based on 3D scanning, and the execution subject of the method for realizing unit automation detection of hyperbolic curtain walls based on 3D scanning includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present invention. In other words, the method for realizing unit automation detection of hyperbolic curtain walls based on 3D scanning can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.
[0093] See also Figure 1 FIG. 1 is a flow chart of a method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning according to an embodiment of the present invention. Figure 1 The method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning described in the specification includes:
[0094] S1. After emitting a detection laser to a current hyperbolic curtain wall, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are collected, and a three-dimensional point cloud model of the current hyperbolic curtain wall is generated using the three-dimensional point cloud coordinates.
[0095] In one embodiment of the present invention, the collecting of the three-dimensional point cloud coordinates of the current hyperbolic curtain wall includes: calculating the source-wall distance between the laser emission source of the detection laser and the current hyperbolic curtain wall using the following formula:
[0096]
[0097] in, represents the source-wall distance, It represents the time duration from the time when the laser emission source emits the detection laser to the time when the reflected laser of the current hyperbolic curtain wall is received. represents the emission speed of the detection laser;
[0098] Based on the source-wall distance, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are calculated using the following formula:
[0099]
[0100]
[0101]
[0102] in, Represents the coordinates of the 3D point cloud, represents the source-wall distance, represents the emission angle of the detection laser in the vertical direction, Indicates the emission angle of the detection laser on the Y axis.
[0103] S2. Collect historical defect data of historical hyperbolic curtain walls, convert the historical defect data into a historical defect sequence based on historical defect moments of the historical defect data, and use the historical defect sequence to analyze a planned defect sequence corresponding to the current hyperbolic curtain wall.
[0104] The embodiment of the present invention collects historical defect data of historical hyperbolic curtain walls, so as to use the historical defect data as an influencing factor for the occurrence of future defects.
[0105] The historical hyperbolic curtain wall refers to a product of the same type as the current hyperbolic curtain wall constructed in a historical period.
[0106] In one embodiment of the present invention, the collecting of historical defect data of the historical hyperbolic curtain wall includes: collecting the historical defect categories and historical defect locations of the historical hyperbolic curtain wall respectively; and calculating the defect penalty intensity of the historical hyperbolic curtain wall using the following formula:
[0107]
[0108] in, Indicates the historical defect category that occurred at the kth moment The penalty for defects caused by the responsible party of the hyperbolic curtain wall is Indicates the amount of fines imposed on the party responsible for the hyperbolic curtain wall. Indicates the duration of the fine imposed on the responsible party of the hyperbolic curtain wall. Indicates the job demotion caused to the party responsible for the hyperbolic curtain wall. , , express , , The weight of
[0109] The defect correction cost of the historic hyperbolic curtain wall was calculated using the following formula:
[0110]
[0111] in, Indicates the historical defect category that occurred at the kth moment The defect correction cost of the responsible party of the hyperbolic curtain wall is It represents the cost of the responsible party to correct the defects of the double-curved curtain wall. Indicates the time it takes for the responsible party to correct the defects of the hyperbolic curtain wall. Indicates the number of personnel dispatched by the responsible party to correct defects in the hyperbolic curtain wall. , , express , , The weight of
[0112] The scale of the defect consequences of the historical hyperbolic curtain wall is calculated using the following formula:
[0113]
[0114] in, Indicates the historical defect category that occurred at the kth moment The scale of the consequences of the defects caused by the hyperbolic curtain wall to the user, Indicates the amount of loss caused by the hyperbolic curtain wall to the user. Indicates the time loss caused by the hyperbolic curtain wall to the user. Indicates the loss of personnel caused by the hyperbolic curtain wall to the user. , , express , , The weight of
[0115] The historical defect category, the historical defect location, the defect penalty intensity, the defect correction cost and the defect consequence scale are used as the historical defect data of the historical hyperbolic curtain wall.
[0116] Among them, the historical defect categories include scratches, dents, bulges, etc. It should be noted that the historical defect categories are numerical values obtained by one-hot encoding text categories such as scratches, dents, bulges, etc. The historical defect location refers to the coordinates where the defect appears on the hyperbolic curtain wall. The defect penalty intensity refers to the degree of punishment for the responsible party when a defect occurs. The defect correction cost refers to the cost of correcting the defect. The defect consequence scale refers to the consequences and losses caused by the defect to the customer.
[0117] It should be noted that each e represents The corresponding number, rather than the data at the kth moment.
[0118] Furthermore, an embodiment of the present invention converts the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, so as to convert the historical defect data into a time series sequence, thereby facilitating feature recognition of the historical defect data by a subsequent neural network model.
[0119] The historical defect sequence refers to historical defect data carrying time information from small to large.
[0120] In one embodiment of the present invention, the converting the historical defect data into a historical defect sequence based on the historical defect moment of the historical defect data includes: calculating the data standard deviation of the historical defect data based on the historical defect moment of the historical defect data; constructing the absolute value threshold of the historical defect data using the data standard deviation; calculating the absolute value of the difference between the historical defect data and the mean of the historical defect data; when the absolute value of the difference is greater than the absolute value threshold, removing the target defect data corresponding to the absolute value of the difference in the historical defect data to obtain remaining defect data; detecting whether there is data missing in the remaining defect data; when there is data missing in the remaining defect data, interpolating the data missing part in the remaining defect data using the following formula to obtain interpolated defect data:
[0121]
[0122] in, represents interpolation defect data, express The remaining defect data at the time, express Remaining defect data at the moment;
[0123] The interpolated defect data is inserted into the remaining defect data to obtain the inserted defect data; based on the historical defect time of the historical defect data, the historical defect sequence of the inserted defect data is constructed using the following formula:
[0124]
[0125] in, express The historical defect sequence of row 5 and column 5, represents the sequence of historical defect categories in the inserted defect data, represents the total number of historical defect moments, The serial number of the historical defect moment, represents the sequence of historical defect positions in the inserted defect data, A sequence representing the magnitude of the consequence of defects in the inserted defect data.
[0126] The data standard deviation refers to the standard deviation of the data at different times, and the absolute value threshold refers to the product of a preset constant and the data standard deviation. The preset constant can be a decimal or an integer, and is intended to reduce or amplify the data standard deviation.
[0127] Optionally, the process of detecting whether there is data missing in the remaining defect data refers to a process of observing whether a time sequence of the remaining defect data is arranged at a fixed time interval.
[0128] Furthermore, the embodiment of the present invention uses the historical defect sequence to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall, so as to analyze at which positions defects of the hyperbolic curtain wall will occur in the future period.
[0129] The planned defect sequence refers to the sequence of locations where defects occur on the hyperbolic curtain wall in the future period.
[0130] In one embodiment of the present invention, the step of analyzing the planned defect sequence corresponding to the current hyperbolic curtain wall using the historical defect sequence comprises: inputting the historical defect sequence into a preset defect analysis model, so as to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall using the defect analysis model; wherein the defect analysis model comprises:
[0131]
[0132]
[0133]
[0134]
[0135]
[0136]
[0137] in, represents the planning defect sequence, Indicates the current time The defect analysis model receives the historical defect sequence. Indicates the last moment The planning defect sequence under represents the output result of the input gate, represents the output of the forget gate, represents the output of the memory gate, Represents the output result of the output gate, Indicates the current time The unit status under Indicates the last moment The unit status under , , , represents the weight parameters and bias parameters of the input gate, , , , represents the weight parameter and bias parameter of the forget gate, , , , represents the weight parameters and bias parameters of the memory gate, , , , Represents the weight parameters and bias parameters of the output gate.
[0138] S3. Query the planning defect time of the planning defect sequence, use the planning defect time to construct the three-dimensional detection sequence of the three-dimensional point cloud model, and use the planning defect sequence to divide the three-dimensional detection units of the three-dimensional point cloud model.
[0139] The embodiment of the present invention utilizes the planning defect moment to construct the three-dimensional detection time sequence of the three-dimensional point cloud model for calculating the probability of the model prediction result occurring in the actual scenario, and utilizes the probability to reconstruct the planning defect moment, thereby increasing the accuracy of the predicted planning moment.
[0140] In one embodiment of the present invention, the three-dimensional detection timing of constructing the three-dimensional point cloud model using the planning defect moment includes: calculating the time increase number of the planning defect moment using the following formula:
[0141]
[0142]
[0143] in, Indicates the number of moments increasing. represents the total number of moments in the jth planning defect position among the planning defect moments, express The total number of moments in the corresponding real scene, It represents the total number of moments after the total number of moments under the j-th planning defect position is increased;
[0144] Based on the time increase number, the defect time is increased at the target time position of the planned defect time to obtain the three-dimensional detection timing.
[0145] Among them, the target time position refers to the position before the time of the j-th planned defect position, that is, the time of the moment increase number is inserted before the time of the j-th planned defect position. For example, the time of the j-th planned defect position is 5, and the time increase number is 2, then 3 and 4 are inserted before 5.
[0146] Furthermore, the embodiment of the present invention divides the three-dimensional detection units of the three-dimensional point cloud model by utilizing the planned defect sequence to divide the continuous coordinates in the predicted planned defect sequence into the same area, and in the subsequent defect detection process, only the three-dimensional detection units on the surface of the hyperbolic curtain wall are detected, and other positions on the hyperbolic curtain wall are not detected, thereby greatly reducing the amount of data processing.
[0147] In one embodiment of the present invention, the three-dimensional detection unit that divides the three-dimensional point cloud model using the planned defect sequence includes: based on the three-dimensional detection timing, adding defect positions to the planned defect sequence to obtain a three-dimensional defect sequence; judging whether there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence; when there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence, dividing the continuous defect positions into the three-dimensional detection units based on the three-dimensional detection timing corresponding to the three-dimensional defect sequence.
[0148] Exemplarily, the process of adding defect positions to the planned defect sequence based on the three-dimensional detection timing to obtain a three-dimensional defect sequence refers to: in the aforementioned, the moment of the j-th planned defect position is 5, and the number of moments increased is 2, then 3 and 4 are inserted before 5, and the j-th planned defect position corresponding to 5 is also used as the j-th planned defect position corresponding to 3 and 4.
[0149] Optionally, the process of determining whether there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence can be performed by determining whether the distance between the coordinate positions at the continuous defect moments is less than a preset distance. If it is less than the preset distance, it can be determined as a continuous defect position.
[0150] S4. Based on the three-dimensional detection sequence, perform model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit to obtain a model defect detection result, and correct the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall.
[0151] The embodiment of the present invention performs model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection sequence to detect defect categories.
[0152] In one embodiment of the present invention, based on the three-dimensional detection timing, model defect detection is performed on the three-dimensional point cloud model on the three-dimensional detection unit to obtain a model defect detection result, including: calculating the unit error between the coordinate point on the three-dimensional detection unit and the coordinate point on a preset standard unit template based on the three-dimensional detection timing; calculating the error standard deviation of the unit error; calculating the error-mean difference between the unit error and the mean of the unit errors; using the error standard deviation and the error-mean difference to identify whether there is a model defect on the three-dimensional detection unit; when there is a model defect on the three-dimensional detection unit, identifying the model defect category on the three-dimensional detection unit; using the model defect category as the model defect detection result; when there is no model defect on the three-dimensional detection unit, using the absence of model defects on the three-dimensional detection unit as the model defect detection result.
[0153] The unit error refers to the coordinate point error.
[0154] Optionally, the process of using the difference between the error standard deviation and the error-mean to identify whether there is a model defect on the three-dimensional detection unit is similar to the above-mentioned process of using the data standard deviation to construct an absolute value threshold of the historical defect data; calculating the absolute value of the difference between the historical defect data and the mean of the historical defect data; when the absolute value of the difference is greater than the absolute value threshold, removing the target defect data corresponding to the absolute value of the difference in the historical defect data to obtain the remaining defect data. Further details are not given here. The process of identifying the model defect category on the three-dimensional detection unit means that after identifying the unit error between the coordinate point on the three-dimensional detection unit and the coordinate point on the preset standard unit template, the defect category can be identified based on the relative position between the coordinate point and the template coordinate point. For example, when the defect category is concave, the coordinate point on the three-dimensional detection unit will be lower than the template coordinate point.
[0155] S5. Assembling the modified hyperbolic curtain wall to obtain an assembled hyperbolic curtain wall, performing assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and using the model defect detection result and the assembly defect detection result as the unit automatic detection result of the current hyperbolic curtain wall.
[0156] In the embodiment of the present invention, the assembly defect detection result refers to the result of detecting whether there is an abnormal splicing seam.
[0157] In one embodiment of the present invention, the performing assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result includes: acquiring a three-dimensional scanning model of the assembled hyperbolic curtain wall; identifying a hyperbolic curtain wall boundary of the assembled hyperbolic curtain wall from the three-dimensional scanning model; querying a neighboring hyperbolic curtain wall boundary of the hyperbolic curtain wall boundary; calculating a boundary error between the hyperbolic curtain wall boundary and the neighboring hyperbolic curtain wall boundary; and performing assembly defect detection on the assembled hyperbolic curtain wall using the boundary error to obtain an assembly defect detection result.
[0158] The neighboring hyperbolic curtain wall boundary refers to a boundary on another hyperbolic curtain wall whose coordinates are closest to the hyperbolic curtain wall boundary, and the other hyperbolic curtain wall refers to a hyperbolic curtain wall spliced with the modified hyperbolic curtain wall.
[0159] Optionally, the principle of using the boundary error to perform assembly defect detection on the assembled hyperbolic curtain wall to obtain the assembly defect detection result is similar to the above-mentioned principles of calculating the unit error between the coordinate point on the three-dimensional detection unit and the coordinate point on the preset standard unit template; calculating the error standard deviation of the unit error; calculating the error-mean difference between the unit error and the mean of the unit errors; using the error standard deviation and the error-mean difference to identify whether there is a model defect on the three-dimensional detection unit; when there is a model defect on the three-dimensional detection unit, identifying the model defect category on the three-dimensional detection unit; using the model defect category as the model defect detection result; when there is no model defect on the three-dimensional detection unit, using the absence of model defects on the three-dimensional detection unit as the model defect detection result, and no further details are given here.
[0160] It can be seen that the embodiment of the present invention collects historical defect data of historical hyperbolic curtain walls to use the historical defect data as an influencing factor for the occurrence of future defects. Furthermore, the embodiment of the present invention converts the historical defect data into a historical defect sequence based on the historical defect moment of the historical defect data to convert the historical defect data into a time series sequence, which is convenient for the subsequent neural network model to identify the features of the historical defect data. Furthermore, the embodiment of the present invention analyzes the planned defect sequence corresponding to the current hyperbolic curtain wall by using the historical defect sequence to analyze at which positions the hyperbolic curtain wall will have defects in the future period. The embodiment of the present invention constructs the three-dimensional point cloud model by using the planned defect moment. The three-dimensional detection sequence is used to calculate the probability of the model prediction result occurring in the actual scene, and the probability is used to reconstruct the planning defect moment, thereby increasing the accuracy of the predicted planning moment. Furthermore, the embodiment of the present invention divides the three-dimensional detection unit of the three-dimensional point cloud model by using the planning defect sequence to divide the continuous coordinates in the predicted planning defect sequence into the same area, and in the subsequent defect detection process, only the three-dimensional detection unit on the surface of the hyperbolic curtain wall is detected, and other positions on the hyperbolic curtain wall are not detected, which greatly reduces the amount of data processing. The embodiment of the present invention performs model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection sequence to detect the defect category. Therefore, the embodiment of the present invention proposes a method and system for realizing automatic detection of units of hyperbolic curtain walls based on 3D scanning, which can reduce the amount of data processing for defect detection and realize the identification of defect categories.
[0161] like Figure 2 As shown, it is a functional module diagram of the unit automatic detection system of the hyperbolic curtain wall based on 3D scanning of the present invention.
[0162] The unit automatic detection system 200 for realizing hyperbolic curtain wall based on 3D scanning of the present invention can be installed in an electronic device. According to the functions implemented, the unit automatic detection system for realizing hyperbolic curtain wall based on 3D scanning can include a model generation module 201, a sequence analysis module 202, a unit division module 203, a curtain wall correction module 204 and an assembly detection module 205. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, which are stored in the memory of the electronic device.
[0163] In the embodiment of the present invention, the functions of each module / unit are as follows:
[0164] The model generation module 201 is used to collect the three-dimensional point cloud coordinates of the current hyperbolic curtain wall after emitting a detection laser to the current hyperbolic curtain wall, and generate a three-dimensional point cloud model of the current hyperbolic curtain wall using the three-dimensional point cloud coordinates;
[0165] The sequence analysis module 202 is used to collect historical defect data of historical hyperbolic curtain walls, convert the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, and use the historical defect sequence to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall;
[0166] The unit division module 203 is used to query the planning defect time of the planning defect sequence, construct the three-dimensional detection sequence of the three-dimensional point cloud model by using the planning defect time, and divide the three-dimensional detection units of the three-dimensional point cloud model by using the planning defect sequence;
[0167] The curtain wall correction module 204 is used to perform model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection time sequence to obtain a model defect detection result, and correct the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall;
[0168] The assembly detection module 205 is used to assemble the modified hyperbolic curtain wall to obtain an assembled hyperbolic curtain wall, perform assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and use the model defect detection result and the assembly defect detection result as the unit automatic detection result of the current hyperbolic curtain wall.
[0169] In detail, the modules in the unit automatic detection system 200 for realizing hyperbolic curtain wall based on 3D scanning in the embodiment of the present invention are used in the same manner as above. Figure 1 The same technical means as the method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning described in the specification and the like can produce the same technical effects, which will not be repeated here.
[0170] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.
[0171] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0172] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.
[0173] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0174] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.
Claims
1. A method for realizing automatic detection of hyperbolic curtain wall units based on 3D scanning, characterized in that: The method comprises: After emitting a detection laser to the current hyperbolic curtain wall, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are collected, and a three-dimensional point cloud model of the current hyperbolic curtain wall is generated using the three-dimensional point cloud coordinates, wherein the collecting of the three-dimensional point cloud coordinates of the current hyperbolic curtain wall includes: The source-to-wall distance between the laser emission source of the detection laser and the current hyperbolic curtain wall is calculated using the following formula: in, represents the source-wall distance, It represents the time duration from the time when the laser emission source emits the detection laser to the time when the reflected laser of the current hyperbolic curtain wall is received. represents the emission speed of the detection laser; Based on the source-wall distance, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are calculated using the following formula: in, Represents the coordinates of the 3D point cloud, represents the source-wall distance, represents the emission angle of the detection laser in the vertical direction, Indicates the emission angle of the detection laser on the Y axis; Collecting historical defect data of historical hyperbolic curtain walls, converting the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, and using the historical defect sequence to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall, wherein collecting the historical defect data of the historical hyperbolic curtain wall includes: respectively collecting the historical defect categories and historical defect locations of the historical hyperbolic curtain wall; The defect penalty intensity of the historical hyperbolic curtain wall is calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The penalty for defects caused by the responsible party of the hyperbolic curtain wall is Indicates the amount of fines imposed on the party responsible for the hyperbolic curtain wall. Indicates the duration of the fine imposed on the responsible party of the hyperbolic curtain wall. Indicates the job demotion caused to the party responsible for the hyperbolic curtain wall. , , express , , The weight of The defect correction cost of the historic hyperbolic curtain wall was calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The defect correction cost of the responsible party of the hyperbolic curtain wall is It represents the cost of the responsible party to correct the defects of the double-curved curtain wall. Indicates the time it takes for the responsible party to correct the defects of the hyperbolic curtain wall. Indicates the number of personnel dispatched by the responsible party to correct defects in the hyperbolic curtain wall. , , express , , The weight of The scale of the defect consequences of the historical hyperbolic curtain wall is calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The scale of the consequences of the defects caused by the hyperbolic curtain wall to the user, Indicates the amount of loss caused by the hyperbolic curtain wall to the user. Indicates the time loss caused by the hyperbolic curtain wall to the user. Indicates the loss of personnel caused by the hyperbolic curtain wall to the user. , , express , , The weight of The historical defect category, the historical defect location, the defect penalty intensity, the defect correction cost and the defect consequence scale are used as the historical defect data of the historical hyperbolic curtain wall; The using the historical defect sequence to analyze the planning defect sequence corresponding to the current hyperbolic curtain wall includes: Inputting the historical defect sequence into a preset defect analysis model, so as to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall through the defect analysis model; Wherein, the defect analysis model includes: in, represents the planning defect sequence, Indicates the current time The defect analysis model receives the historical defect sequence. Indicates the last moment The planning defect sequence under represents the output result of the input gate, represents the output of the forget gate, represents the output of the memory gate, Represents the output result of the output gate, Indicates the current time The unit status under Indicates the last moment The unit status under , , , represents the weight parameters and bias parameters of the input gate, , , , represents the weight parameter and bias parameter of the forget gate, , , , represents the weight parameters and bias parameters of the memory gate, , , , Represents the weight parameters and bias parameters of the output gate; querying a planning defect time of the planning defect sequence, constructing a three-dimensional detection time sequence of the three-dimensional point cloud model using the planning defect time, and dividing the three-dimensional detection units of the three-dimensional point cloud model using the planning defect sequence; Based on the three-dimensional detection sequence, performing model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit to obtain a model defect detection result, and correcting the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall; The modified hyperbolic curtain wall is assembled to obtain an assembled hyperbolic curtain wall, an assembly defect detection is performed on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and the model defect detection result and the assembly defect detection result are used as the unit automatic detection result of the current hyperbolic curtain wall.
2. The method according to claim 1, characterized in that The converting the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data comprises: Calculating a data standard deviation of the historical defect data based on the historical defect moments of the historical defect data; constructing an absolute value threshold of the historical defect data using the data standard deviation; Calculating an absolute value of a difference between the historical defect data and a mean value of the historical defect data; When the absolute value of the difference is greater than the absolute value threshold, removing the target defect data corresponding to the absolute value of the difference in the historical defect data to obtain remaining defect data; Detecting whether the remaining defect data has data missing; When there are missing data in the remaining defective data, the following formula is used to interpolate the missing data in the remaining defective data to obtain interpolated defective data: in, represents interpolation defect data, express The remaining defect data at the time, express Remaining defect data at the moment; Inserting the interpolation defect data into the remaining defect data to obtain the inserted defect data; Based on the historical defect time of the historical defect data, the historical defect sequence of the inserted defect data is constructed using the following formula: in, express The historical defect sequence of row 5 and column 5, represents the sequence of historical defect categories in the inserted defect data, represents the total number of historical defect moments, The serial number representing the moment of historical defect, represents the sequence of historical defect positions in the inserted defect data, A sequence representing the magnitude of the defect consequence in the inserted defect data.
3. The method according to claim 1, characterized in that: The three-dimensional detection sequence of constructing the three-dimensional point cloud model by using the planned defect moment includes: The following formula is used to calculate the number of times the planning defect time increases: in, Indicates the number of moments increasing. represents the total number of moments in the jth planning defect position among the planning defect moments, express The total number of moments in the corresponding real scene, It represents the total number of moments after the total number of moments under the j-th planning defect position is increased; Based on the time increase number, the defect time is increased at the target time position of the planned defect time to obtain the three-dimensional detection timing.
4. The method according to claim 1, characterized in that: The three-dimensional detection unit that divides the three-dimensional point cloud model using the planned defect sequence includes: Based on the three-dimensional detection time sequence, adding defect positions to the planned defect sequence to obtain a three-dimensional defect sequence; Determining whether there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence; When there are continuous defect positions at continuous defect moments in the three-dimensional defect sequence, the continuous defect positions are divided into the three-dimensional detection units based on the three-dimensional detection timing corresponding to the three-dimensional defect sequence.
5. The method according to claim 1, characterized in that The performing model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection time sequence to obtain a model defect detection result includes: Based on the three-dimensional detection timing, calculating the unit error between the coordinate point on the three-dimensional detection unit and the coordinate point on the preset standard unit template; Calculating the error standard deviation of the unit error; calculating an error-mean difference between the unit error and a mean of the unit errors; Using the difference between the error standard deviation and the error-mean value to identify whether there is a model defect on the three-dimensional detection unit; When there is a model defect on the three-dimensional detection unit, identifying a type of the model defect on the three-dimensional detection unit; Using the model defect category as the model defect detection result; When there is no model defect on the three-dimensional detection unit, the absence of the model defect on the three-dimensional detection unit is used as the model defect detection result.
6. The method according to claim 1, characterized in that The step of performing assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result includes: Obtaining a three-dimensional scanning model of the assembled hyperbolic curtain wall; identifying a hyperbolic curtain wall boundary of the assembled hyperbolic curtain wall from the three-dimensional scanned model; Querying the neighboring hyperbolic curtain wall boundaries of the hyperbolic curtain wall boundary; Calculating a boundary error between the hyperbolic curtain wall boundary and the adjacent hyperbolic curtain wall boundary; The boundary error is used to perform assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result.
7. A unit automatic detection system for hyperbolic curtain wall based on 3D scanning, characterized in that: The system comprises: The model generation module is used to collect the three-dimensional point cloud coordinates of the current hyperbolic curtain wall after emitting a detection laser to the current hyperbolic curtain wall, and generate a three-dimensional point cloud model of the current hyperbolic curtain wall using the three-dimensional point cloud coordinates, wherein the collecting of the three-dimensional point cloud coordinates of the current hyperbolic curtain wall includes: The source-to-wall distance between the laser emission source of the detection laser and the current hyperbolic curtain wall is calculated using the following formula: in, represents the source-wall distance, It represents the time duration from the time when the laser emission source emits the detection laser to the time when the reflected laser of the current hyperbolic curtain wall is received. represents the emission speed of the detection laser; Based on the source-wall distance, the three-dimensional point cloud coordinates of the current hyperbolic curtain wall are calculated using the following formula: in, Represents the coordinates of the 3D point cloud, represents the source-wall distance, represents the emission angle of the detection laser in the vertical direction, Indicates the emission angle of the detection laser on the Y axis; A sequence analysis module is used to collect historical defect data of historical hyperbolic curtain walls, convert the historical defect data into a historical defect sequence based on the historical defect moments of the historical defect data, and use the historical defect sequence to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall, wherein the collecting of historical defect data of historical hyperbolic curtain walls includes: respectively collecting the historical defect categories and historical defect locations of the historical hyperbolic curtain wall; The defect penalty intensity of the historical hyperbolic curtain wall is calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The penalty for defects caused by the responsible party of the hyperbolic curtain wall is Indicates the amount of fines imposed on the party responsible for the hyperbolic curtain wall. Indicates the duration of the fine imposed on the responsible party of the hyperbolic curtain wall. Indicates the job demotion caused to the party responsible for the hyperbolic curtain wall. , , express , , The weight of The defect correction cost of the historic hyperbolic curtain wall was calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The defect correction cost of the responsible party of the hyperbolic curtain wall is It represents the cost of the responsible party to correct the defects of the double-curved curtain wall. Indicates the time it takes for the responsible party to correct the defects of the hyperbolic curtain wall. Indicates the number of personnel dispatched by the responsible party to correct defects in the hyperbolic curtain wall. , , express , , The weight of The scale of the defect consequences of the historical hyperbolic curtain wall is calculated using the following formula: in, Indicates the historical defect category that occurred at the kth moment The scale of the consequences of the defects caused by the hyperbolic curtain wall to the user, Indicates the amount of loss caused by the hyperbolic curtain wall to the user. Indicates the time loss caused by the hyperbolic curtain wall to the user. Indicates the loss of personnel caused by the hyperbolic curtain wall to the user. , , express , , The weight of The historical defect category, the historical defect location, the defect penalty intensity, the defect correction cost and the defect consequence scale are used as the historical defect data of the historical hyperbolic curtain wall; The using the historical defect sequence to analyze the planning defect sequence corresponding to the current hyperbolic curtain wall includes: Inputting the historical defect sequence into a preset defect analysis model, so as to analyze the planned defect sequence corresponding to the current hyperbolic curtain wall through the defect analysis model; Wherein, the defect analysis model includes: in, represents the planning defect sequence, Indicates the current time The defect analysis model receives the historical defect sequence. Indicates the last moment The planning defect sequence under represents the output result of the input gate, represents the output of the forget gate, represents the output of the memory gate, Represents the output result of the output gate, Indicates the current time The unit status under Indicates the last moment The unit status under , , , represents the weight parameters and bias parameters of the input gate, , , , represents the weight parameter and bias parameter of the forget gate, , , , represents the weight parameters and bias parameters of the memory gate, , , , Represents the weight parameters and bias parameters of the output gate; a unit division module, used to query the planning defect time of the planning defect sequence, construct the three-dimensional detection time sequence of the three-dimensional point cloud model by using the planning defect time, and divide the three-dimensional detection unit of the three-dimensional point cloud model by using the planning defect sequence; A curtain wall correction module, configured to perform model defect detection on the three-dimensional point cloud model on the three-dimensional detection unit based on the three-dimensional detection sequence to obtain a model defect detection result, and to correct the current hyperbolic curtain wall based on the model defect detection result to obtain a corrected hyperbolic curtain wall; An assembly detection module is used to assemble the modified hyperbolic curtain wall to obtain an assembled hyperbolic curtain wall, perform assembly defect detection on the assembled hyperbolic curtain wall to obtain an assembly defect detection result, and use the model defect detection result and the assembly defect detection result as a unit automatic detection result of the current hyperbolic curtain wall.
