A construction process management method and system based on AI

Abstract

The present application provides an AI-based construction process management method and system, the method comprising obtaining the first image data and the first point cloud data of the target tunnel in the first period of time through an image acquisition device and a laser scanning device respectively; inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, the excavation area recognition result including the corresponding position of each excavation area on the first image data; obtaining the second image data of the target tunnel in the second period of time through an image acquisition device; inputting the second image data and the excavation area recognition result into a pre-trained construction process recognition model to obtain a construction process recognition result output by the construction process recognition model. In this way, the construction process stage of the tunnel can be determined more accurately, providing accurate data support for subsequent construction progress control and construction safety control.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and specifically, to an AI-based construction process management method and system. Background Art

[0002] In tunnel construction scenarios, there are high requirements for construction safety and there is a problem of small construction space. Therefore, in order to ensure construction safety and improve construction efficiency, it is necessary to accurately and timely understand the current construction process or construction process. The existing method of manually recording and reporting construction processes or construction processes is prone to misreporting or untimely information updates. Summary of the invention

[0003] In order to overcome the above-mentioned deficiencies in the prior art, the purpose of this application is to provide an AI-based construction process management method, the method comprising:

[0004] Acquire first image data and first point cloud data of the target tunnel in the first period of time respectively by using an image acquisition device and a laser scanning device;

[0005] Inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes corresponding positions of various excavation areas on the first image data;

[0006] Acquiring, by the image acquisition device, second image data of the target tunnel in the second time period;

[0007] The second image data and the excavation area recognition result are input into a pre-trained construction process recognition model to obtain a construction process recognition result output by the construction process recognition model.

[0008] In some possible implementations, the excavation area includes a first excavation area, a second excavation area, a third excavation area, a fourth excavation area, a fifth excavation area, a sixth excavation area, and a seventh excavation area;

[0009] The first excavation area is located at the vault of the target tunnel, the second excavation area and the third excavation area are respectively located at the upper part of the two side walls of the target tunnel, the fourth excavation area and the fifth excavation area are respectively located at the lower part of the two side walls of the target tunnel, the sixth excavation area is located at the center of the target tunnel, at least part of the sixth excavation area is located between the second excavation area and the third excavation area, at least part of the sixth excavation area is located between the fourth excavation area and the fifth excavation area, and the seventh excavation area is located at the bottom ground of the target tunnel position;

[0010] Among them, the excavation area is divided into three construction steps, the sixth excavation area includes a first sub-position, a second sub-position and a third sub-position, the top surfaces of the third sub-position, the fourth excavation area and the fifth excavation area are located on the first step, the top surfaces of the second sub-position, the second excavation area and the third excavation area are located on the second step, and the top surface of the third sub-position is located on the third step.

[0011] In some possible implementations, the step of inputting the second image data and the excavation area recognition result into a pre-trained construction process recognition model to obtain the construction process recognition result output by the construction process recognition model includes:

[0012] The second image data is identified by a content identification module of the construction process identification model to determine first operation position data of each target construction object in the second image data, wherein the target construction object includes at least one of an engineering vehicle, a construction worker, and an internal support structure of a tunnel;

[0013] Inputting the first operation position data and the excavation area identification result into the position matching module of the construction process identification model to obtain the corresponding relationship between each excavation area in the target tunnel and the construction position of each target construction object;

[0014] The construction process identification result of each excavation area is determined according to the construction position correspondence.

[0015] In some possible implementations, the method further includes:

[0016] The first image data, the excavation area recognition result and the second image data are input into a pre-trained excavation area update model to obtain an updated excavation area recognition result output by the excavation area update model.

[0017] In some possible implementations, the image acquisition device and the laser scanning device are disposed at locations within a range of 20 cm from each other.

[0018] In some possible implementations, the step of acquiring first image data and first point cloud data of the target tunnel in the first time period by using an image acquisition device and a laser scanning device respectively includes:

[0019] Acquire first image data and first point cloud data of the target tunnel during a period when no construction or construction is suspended, respectively, by using an image acquisition device and a laser scanning device;

[0020] The step of acquiring the second image data of the target tunnel in the second time period by the image acquisition device comprises:

[0021] The second image data of the target tunnel during the construction period is acquired by the image acquisition device.

[0022] In some possible implementations, the method further includes:

[0023] Obtaining at least one historical construction process identification result and a current construction process identification result;

[0024] Inputting the historical construction process recognition results and the current construction process recognition results into a pre-trained construction process prediction model to obtain a construction process prediction result output by the construction process prediction model;

[0025] The engineering vehicles are dispatched according to the construction process prediction results.

[0026] In some possible implementations, the method further includes:

[0027] In the case where there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively;

[0028] Reconstructing a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel;

[0029] Marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result;

[0030] Performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of various excavation areas on the first sample image data as first training labels according to the first excavation area labeling result;

[0031] In the case where there are construction personnel and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively;

[0032] The second sample image data and the second sample point cloud data are used as first training samples, and the first training samples and the first training labels are used to train the excavation area recognition model, so that the excavation area recognition model determines the corresponding position of each excavation area on the input image data according to the input image data and point cloud data.

[0033] In some possible implementations, the method further includes:

[0034] In the case where there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively;

[0035] Reconstructing a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel;

[0036] Marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result;

[0037] Performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of each excavation area on the first sample image data according to the first excavation area annotation result, and using the internal 3D model of the sample tunnel, the first excavation area annotation result and the corresponding positions of each excavation area on the first sample image data together as second training labels;

[0038] In the case where there are construction personnel and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively;

[0039] The second sample image data and the second sample point cloud data are used as first training samples, and the first training samples and the second training labels are used to train the excavation area recognition model, so that the excavation area recognition model outputs a 3D model of the interior of the tunnel, a corresponding position of each excavation area in the 3D model of the interior of the sample tunnel, and a corresponding position of each excavation area on the input image data according to the input image data and point cloud data;

[0040] Wherein, in the step of inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, the excavation area recognition result includes a first recognition result and a second recognition result; the first recognition result includes a corresponding position of each excavation area on the first image data; the second recognition result includes a 3D model of the interior of the target tunnel and a corresponding position of each excavation area on the 3D model of the interior of the target tunnel;

[0041] The method further comprises:

[0042] A dynamic 3D construction progress display result of the target tunnel is generated according to the second recognition result corresponding to the data collected in multiple time periods.

[0043] The present application also provides an AI-based construction process management system, the system comprising an image acquisition device, a laser scanning device and a data processing device;

[0044] The image acquisition device is used to obtain first image data of the target tunnel during a non-construction period;

[0045] The laser scanning device is used to obtain first point cloud data of the target tunnel in an unconstructed state;

[0046] The data processing device is used to input the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes a corresponding position of each excavation area on the first image data;

[0047] The image acquisition device is also used to obtain second image data during the construction period of the target tunnel;

[0048] The data processing device is also used to input the second image data and the excavation area recognition result into a pre-trained construction process recognition model to obtain the construction process recognition result output by the construction process recognition model.

[0049] Compared with the prior art, this application has the following beneficial effects:

[0050] This application provides an AI-based construction process management method and system, which first identifies the excavation area by pre-collecting image data and point cloud data, and then determines the current construction process identification result together with the collected image and excavation area identification result during the construction process. In this way, the construction process stage of the tunnel can be determined more accurately, providing accurate data support for subsequent construction progress control and construction safety control. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 One of the step flow diagrams of an AI-based construction process management method provided in an embodiment of the present application;

[0053] Figure 2 A schematic diagram of an application tunnel environment of an AI-based construction process management method provided in an embodiment of the present application;

[0054] Figure 3The second step flow diagram of an AI-based construction process management method provided in an embodiment of the present application;

[0055] Figure 4 The third step flow diagram of a construction process management method based on AI provided in an embodiment of the present application;

[0056] Figure 5 A schematic diagram of an AI-based construction process management system provided in an embodiment of the present application;

[0057] Figure 6 A block diagram of a data processing device provided in an embodiment of the present application.

[0058] Icon: 1-first excavation area; 2-second excavation area; 3-third excavation area; 4-fourth excavation area; 5-fifth excavation area; 6-sixth excavation area; 7-seventh excavation area; 6-1-first sub-position; 6-2-second sub-position; 6-3-third sub-position. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0060] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0061] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0062] In the description of the present application, it should be noted that the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0063] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0064] In the description of this application, it should also be noted that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0065] See also Figure 1 , Figure 1 A schematic flowchart of a construction process management method based on AI is provided for this embodiment. The method may include the following steps.

[0066] Step S110, obtaining first image data and first point cloud data of the target tunnel in the first time period by using an image acquisition device and a laser scanning device respectively.

[0067] Specifically, on the one hand, the tunnel construction environment has the problem of poor lighting. When collecting tunnel data, if only image acquisition equipment is used, the collected image data may not accurately reflect the internal situation of the tunnel due to poor lighting. On the other hand, the tunnel construction environment also has the problem of small space. The concentration of engineering vehicles (such as excavators, trolleys, etc.) in the space is relatively large. If only laser scanning equipment is used for point cloud data collection, the collected point cloud data may contain more noise due to the large impact of engineering vehicles on the collection field of view.

[0068] In view of this, in this embodiment, an image acquisition device and a laser scanning device are used simultaneously to respectively obtain the first image data and the first point cloud data of the target tunnel in the first time period. In this way, when the collected data is subsequently processed, the image data and the point cloud data can complement each other, thereby better reflecting the internal conditions of the tunnel.

[0069] Optionally, in this embodiment, in order to enable subsequent image data and point cloud data to achieve point matching more conveniently and accurately, in this embodiment, the image acquisition device and the laser scanning device are set at positions within 20 cm of each other.

[0070] Optionally, the image acquisition device and the laser scanning device can be integrated into the same acquisition device and arranged at the same position.

[0071] Step S120, inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes corresponding positions of each excavation area on the first image data.

[0072] See also Figure 2 In the excavation and construction process of some large tunnels, in order to improve construction efficiency and construction flexibility, the "three-step seven-step excavation construction method" will be adopted. The target tunnel will be divided into multiple excavation areas. Multiple excavation areas are at multiple different steps, and multiple excavation areas can be staggered and constructed in parallel.

[0073] Specifically, the excavation area of ​​the target tunnel may include a first excavation area 1 , a second excavation area 2 , a third excavation area 3 , a fourth excavation area 4 , a fifth excavation area 5 , a sixth excavation area 6 and a seventh excavation area 7 .

[0074] The first excavation area 1 is located at the vault of the target tunnel, the second excavation area 2 and the third excavation area 3 are respectively located at the upper part of the two side walls of the target tunnel, the fourth excavation area 4 and the fifth excavation area 5 are respectively located at the lower part of the two side walls of the target tunnel, the sixth excavation area 6 is located at the center of the target tunnel, at least part of the sixth excavation area 6 is located between the second excavation area 2 and the third excavation area 3, at least part of the sixth excavation area 6 is located between the fourth excavation area 4 and the fifth excavation area 5, and the seventh excavation area 7 is located at the bottom ground of the target tunnel position.

[0075] Among them, the excavation area is divided into three construction steps, the sixth excavation area 6 includes a first sub-position 6-1, a second sub-position 6-2 and a third sub-position 6-3, the top surfaces of the third sub-position 6-3, the fourth excavation area 4 and the fifth excavation area 5 are located on the first step, the top surfaces of the second sub-position 6-2, the second excavation area 2 and the third excavation area 3 are located on the second step, and the top surface of the third sub-position 6-3 is located on the third step.

[0076] In this embodiment, the excavation area recognition model can perform excavation area recognition according to the input first image data and the first point cloud data, so as to determine the corresponding position of each excavation area in the first image data.

[0077] Among them, the data input into the excavation area identification model includes image data and point cloud data, which can avoid the problem of inaccurate recognition caused by unclear boundary images of various areas when only image data is used when the lighting inside the tunnel is poor. It can also avoid the problem of inaccurate recognition due to excessive occlusion when only point cloud data is used when there are many engineering vehicles inside the tunnel.

[0078] The training method of the excavation area recognition model will be explained in detail later.

[0079] Step S130: acquiring, by the image acquisition device, second image data of the target tunnel in a second time period.

[0080] Optionally, the second time period may be a time period during which the result of the current construction process needs to be determined.

[0081] In step S110, first image data and first point cloud data of the target tunnel during a period when no construction or construction is suspended can be obtained by an image acquisition device and a laser scanning device respectively.

[0082] That is, in step S110, the first time period may be a time period when the target tunnel is not under construction or when construction is suspended. During this time period, there are fewer construction personnel and engineering vehicles in the target tunnel or the engineering vehicles are in a stationary state, so that the noise in the first image data and the first point cloud data collected in step S110 can be reduced, thereby improving the accuracy of the excavation area recognition result in step S120.

[0083] In step S130, in order to accurately determine the current construction process stage, the image of the current construction state can be directly obtained as the second image data. That is, the second time period in step S130 can be the time period of construction.

[0084] Step S140: input the second image data and the excavation area recognition result into a pre-trained construction process recognition model to obtain a construction process recognition result output by the construction process recognition model.

[0085] Specifically, in step S140, the second image data can be identified by the content recognition module of the construction process recognition model to determine the first working position data of each target construction object in the second image data. The target construction object includes at least one of an engineering vehicle, construction personnel and an internal support structure of a tunnel.

[0086] The first operation position data and the excavation area identification result are input into the position matching module of the construction process identification model to obtain the corresponding relationship between each excavation area in the target tunnel and the construction position of each target construction object.

[0087] The construction process identification result of each excavation area is determined according to the construction position correspondence.

[0088] Among them, according to the corresponding relationship between each excavation area and the construction position of each target construction object, it can be determined whether each excavation area is under construction and the type of construction work. For example, when the excavator is located in a certain excavation area, it can be determined that the excavation area is performing excavation work. When there is a support mechanism in a certain excavation area, it can be determined that the excavation area is performing support work.

[0089] Based on the above design, this application provides an AI-based construction process management method and system, which first identifies the excavation area by pre-collecting image data and point cloud data, and then determines the current construction process identification result together with the collected image and excavation area identification result during the construction process. In this way, the construction process stage of the tunnel can be determined more accurately, providing accurate data support for subsequent construction progress control and construction safety control.

[0090] In some possible implementations, the solution provided in this embodiment may further include step S150.

[0091] Step S150: input the first image data, the excavation area recognition result and the second image data into a pre-trained excavation area update model to obtain an updated excavation area recognition result output by the excavation area update model.

[0092] In this way, when the excavation area boundary changes due to excavation operations during tunnel operation, the excavation area identification results can be updated in time, thereby making subsequent identification results more accurate.

[0093] In some possible implementations, the solution provided in this embodiment may further include the following steps.

[0094] Step S210, obtaining at least one historical construction process identification result and a current construction process identification result.

[0095] Step S220, inputting the historical construction process recognition results and the current construction process recognition results into a pre-trained construction process prediction model to obtain the construction process prediction results output by the construction process prediction model.

[0096] Step S230, dispatching engineering vehicles according to the construction process prediction results.

[0097] Specifically, in this embodiment, since different construction steps or different excavation areas can be constructed in parallel, but the space of the target tunnel is limited, it is necessary to reasonably arrange the entry and exit time of different engineering vehicles to improve the utilization efficiency of the engineering vehicles. In this embodiment, the subsequent construction process can be predicted based on the historical construction process identification results and the current construction process identification results, and the engineering vehicles can be dispatched in a more orderly manner according to the prediction results.

[0098] In some possible implementations, the method provided in this embodiment also includes a training step of an excavation area identification model.

[0099] Optionally, see Figure 3 , wherein a training step of the excavation area identification model may include the following steps.

[0100] Step S310, when there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively.

[0101] Step S320: Reconstruct a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel.

[0102] Step S330, marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result.

[0103] Step S340: performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of each excavation area on the first sample image data as first training labels according to the first excavation area labeling result.

[0104] Step S350, when there are construction workers and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively.

[0105] Step S360, using the second sample image data and the second sample point cloud data as the first training sample, and using the first training sample and the first training label to train the excavation area recognition model, so that the excavation area recognition model determines the corresponding position of each excavation area on the input image data according to the input image data and point cloud data.

[0106] Specifically, when collecting training samples for training the excavation area recognition model, the first training label can be obtained through steps S310 to S340, and the first training sample can be obtained through step S350.

[0107] Among them, in step S310, when there are no construction personnel and construction vehicles in the sample tunnel, the first sample image data and the first sample point cloud data of the sample tunnel can be respectively obtained by the image acquisition device and the laser scanning device. In this way, in step S320, a more accurate 3D model of the interior of the tunnel can be obtained based on the first sample point cloud data, eliminating the interference of construction personnel and construction vehicles in generating the 3D model of the interior of the tunnel.

[0108] Next, in step S330, the excavation area can be marked based on the 3D model of the tunnel. Since the excavation areas are constructed in parallel, there may be a certain boundary occlusion relationship between the excavation areas. Compared with marking the excavation areas only on the 2D image, the excavation areas marked on the 3D model can more accurately reflect the boundaries of each excavation area.

[0109] Then in step S340, the first sample image data and the first sample point cloud data are point matched to determine the correspondence between the first sample image data and each point in the first sample point cloud data, and then the corresponding position of each excavation area in the first sample image is accurately determined based on the correspondence between the first sample point cloud data and the reconstructed 3D model of the interior of the tunnel and the first position marking result.

[0110] In step S350, when there are construction workers and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel can be acquired by an image acquisition device and a laser scanning device respectively, and used as the first training sample.

[0111] Optionally, in this embodiment, when the first sample image data and the first sample point cloud data are collected and when the second sample image data and the second sample point cloud data are collected, the boundary range of each excavation area does not change significantly. The position of each excavation area marked and determined by steps S310 to S340 on the first sample image and the position on the second sample image are substantially the same.

[0112] Then in step S360, the excavation area recognition model is trained using the first training sample and the first training label, so that the excavation area recognition model outputs the same result as the first training label as much as possible according to the first training sample.

[0113] Among them, since there are construction workers and engineering vehicles in the first training sample, and the interference of construction workers and engineering vehicles is eliminated in the first training label, in the training process of the excavation area recognition model, the excavation area recognition model will gradually learn (i.e., adjust the model parameters) to eliminate the interference of construction workers and engineering vehicles in the input image data and point cloud data, so that when there is interference from construction workers and engineering vehicles in the input image data and point cloud data, more accurate excavation area recognition results can be obtained.

[0114] Optionally, see Figure 4 This embodiment also provides another training step for the excavation area identification model, which may specifically include the following steps.

[0115] Step S410, when there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively.

[0116] Step S420: Reconstruct a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel.

[0117] Step S430, marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result.

[0118] Step S440, performing point matching on the first sample image data and the first sample point cloud data, and determining the corresponding position of each excavation area on the first sample image data according to the first excavation area annotation result, and taking the internal 3D model of the sample tunnel, the first excavation area annotation result and the corresponding position of each excavation area on the first sample image data together as the second training label.

[0119] Step S450, when there are construction workers and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively.

[0120] Step S460, using the second sample image data and the second sample point cloud data as the first training sample, and using the first training sample and the second training label to train the excavation area recognition model, so that the excavation area recognition model outputs the internal 3D model of the tunnel, the corresponding position of each excavation area in the internal 3D model of the sample tunnel, and the corresponding position of each excavation area on the input image data according to the input image data and point cloud data.

[0121] Different from step S340, in step S440, the second training label includes not only the corresponding position of each excavation area on the first sample image data, but also the internal 3D model of the sample tunnel and the first excavation area labeling result.

[0122] Therefore, in step S460, the excavation area identification model not only needs to determine the corresponding position of each excavation area on the input image data based on the input image data and the point cloud data, but also needs to generate a 3D model of the interior of the tunnel based on the input image data and the point cloud data and mark each excavation area on the 3D model of the interior of the tunnel.

[0123] On this basis, the excavation area recognition result obtained in step S120 includes a first recognition result and a second recognition result. The first recognition result includes the corresponding position of each excavation area on the first image data. The second recognition result includes the tunnel internal 3D model of the target tunnel and the corresponding position of each excavation area on the internal 3D model of the target tunnel.

[0124] In this way, the tunnel internal 3D model data generated in steps S410 to S440 and the annotated first excavation area annotation results are better utilized in the training stage, so that the excavation area identification model also has the function of restoring the tunnel 3D model, thereby improving data utilization.

[0125] In this case, the method provided in this embodiment may further include step S510.

[0126] Step S510: Generate a dynamic 3D construction progress display result of the target tunnel according to the second recognition result corresponding to the data collected in multiple time periods.

[0127] That is, in this embodiment, since each time the current construction process identification result is obtained, the internal 3D model of the current target tunnel and the corresponding position of each excavation area on the internal 3D model of the target tunnel can also be obtained. In this way, the second identification results corresponding to the data collected in multiple time periods can be spliced ​​in time sequence to generate a dynamic 3D construction progress display result of the target tunnel, thereby more intuitively and conveniently displaying the progress of each construction process of the target tunnel, and providing accurate data support for subsequent construction progress control and construction safety control.

[0128] Optionally, in steps S310 to S340 or in steps S410 to S440, when reconstructing the 3D model of the interior of the tunnel based on the point cloud data, in order to improve the accuracy of the model, point cloud data from different perspectives can be collected for point matching and splicing to generate a 3D model of the interior of the tunnel, thereby reducing the impact of boundary occlusion of each excavation area on the reconstruction result.

[0129] See also Figure 5 This embodiment also provides an AI-based construction process management system, which includes an image acquisition device 200, a laser scanning device 300 and a data processing device 100.

[0130] The image acquisition device 200 is used to acquire first image data of a target tunnel during a non-construction period.

[0131] The laser scanning device 300 is used to obtain first point cloud data of the target tunnel in an unconstructed state.

[0132] The data processing device 100 is used to input the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, and the excavation area recognition result includes the corresponding positions of each excavation area on the first image data.

[0133] The image acquisition device 200 is also used to obtain second image data during the target tunnel construction period.

[0134] The data processing device 100 is further used to input the second image data and the excavation area recognition result into a pre-trained construction process recognition model to obtain the construction process recognition result output by the construction process recognition model.

[0135] Optionally, see Figure 6 , Figure 6 yes Figure 5 The data processing device 100 is a block diagram showing a data processing device 100. The data processing device 100 includes a machine-readable storage medium 120 and a processor 130.

[0136] The machine-readable storage medium 120 and the processor 130 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines.

[0137] The machine-readable storage medium 120 may be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The machine-readable storage medium 120 is used for machine executable instructions. After receiving the execution instruction, the processor 130 executes the machine executable instruction to implement the AI-based construction process management method provided in this embodiment.

[0138] The processor 130 may be an integrated circuit chip with signal processing capabilities. The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0139] In summary, the application provides an AI-based construction process management method and system, which first identifies the excavation area by pre-collecting image data and point cloud data, and then determines the current construction process identification result together with the collected image and excavation area identification result during the construction process. In this way, the construction process stage of the tunnel can be determined more accurately, providing accurate data support for subsequent construction progress control and construction safety control.

[0140] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0141] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0142] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program codes.

[0143] It should be noted that, in this article, relational terms such as first and second, etc. 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.

[0144] The above are only various implementations of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A construction process management method based on AI, characterized in that: The method comprises: Acquire first image data and first point cloud data of the target tunnel in the first period of time respectively by using an image acquisition device and a laser scanning device; Inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes corresponding positions of various excavation areas on the first image data; Acquiring, by the image acquisition device, second image data of the target tunnel in the second time period; The second image data is identified by a content identification module of a construction process identification model to determine first operation position data of each target construction object in the second image data, wherein the target construction object includes at least one of an engineering vehicle, a construction worker, and an internal support structure of a tunnel; Inputting the first operation position data and the excavation area identification result into the position matching module of the construction process identification model to obtain the corresponding relationship between each excavation area in the target tunnel and the construction position of each target construction object; Determine the construction process identification result of each excavation area according to the construction position correspondence relationship; Wherein, the method further comprises: In the case where there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; Reconstructing a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel; Marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result; Performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of various excavation areas on the first sample image data as first training labels according to the first excavation area labeling result; In the case where there are construction personnel and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; The second sample image data and the second sample point cloud data are used as first training samples, and the first training samples and the first training labels are used to train the excavation area recognition model, so that the excavation area recognition model determines the corresponding position of each excavation area on the input image data according to the input image data and point cloud data.

2. The method according to claim 1, characterized in that The excavation area includes a first excavation area, a second excavation area, a third excavation area, a fourth excavation area, a fifth excavation area, a sixth excavation area and a seventh excavation area; The first excavation area is located at the vault of the target tunnel, the second excavation area and the third excavation area are respectively located at the upper part of the two side walls of the target tunnel, the fourth excavation area and the fifth excavation area are respectively located at the lower part of the two side walls of the target tunnel, the sixth excavation area is located at the center of the target tunnel, at least part of the sixth excavation area is located between the second excavation area and the third excavation area, at least part of the sixth excavation area is located between the fourth excavation area and the fifth excavation area, and the seventh excavation area is located at the bottom ground of the target tunnel position; Among them, the excavation area is divided into three construction steps, the sixth excavation area includes a first sub-position, a second sub-position and a third sub-position, the top surfaces of the third sub-position, the fourth excavation area and the fifth excavation area are located on the first step, the top surfaces of the second sub-position, the second excavation area and the third excavation area are located on the second step, and the top surface of the third sub-position is located on the third step.

3. The method according to claim 1, characterized in that The method further comprises: The first image data, the excavation area recognition result and the second image data are input into a pre-trained excavation area update model to obtain an updated excavation area recognition result output by the excavation area update model.

4. The method according to claim 1, characterized in that: The image acquisition device and the laser scanning device are arranged at positions within a range of 20 cm from each other.

5. The method according to claim 1, characterized in that The step of respectively acquiring the first image data and the first point cloud data of the target tunnel in the first time period by using the image acquisition device and the laser scanning device comprises: Acquire first image data and first point cloud data of the target tunnel during a period when no construction or construction is suspended, respectively, by using an image acquisition device and a laser scanning device; The step of acquiring the second image data of the target tunnel in the second time period by the image acquisition device comprises: The second image data of the target tunnel during the construction period is acquired by the image acquisition device.

6. The method according to claim 1, characterized in that The method further comprises: Obtaining at least one historical construction process identification result and a current construction process identification result; Inputting the historical construction process recognition results and the current construction process recognition results into a pre-trained construction process prediction model to obtain a construction process prediction result output by the construction process prediction model; The engineering vehicles are dispatched according to the construction process prediction results.

7. A construction process management method based on AI, characterized in that: The method comprises: Acquire first image data and first point cloud data of the target tunnel in the first period of time respectively by using an image acquisition device and a laser scanning device; Inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes corresponding positions of various excavation areas on the first image data; Acquiring, by the image acquisition device, second image data of the target tunnel in the second time period; The second image data is identified by a content identification module of a construction process identification model to determine first operation position data of each target construction object in the second image data, wherein the target construction object includes at least one of an engineering vehicle, a construction worker, and an internal support structure of a tunnel; Inputting the first operation position data and the excavation area identification result into the position matching module of the construction process identification model to obtain the corresponding relationship between each excavation area in the target tunnel and the construction position of each target construction object; Determine the construction process identification result of each excavation area according to the construction position correspondence relationship; Wherein, the method further comprises: In the case where there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; Reconstructing a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel; Marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result; Performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of each excavation area on the first sample image data according to the first excavation area annotation result, and using the internal 3D model of the sample tunnel, the first excavation area annotation result and the corresponding positions of each excavation area on the first sample image data together as second training labels; In the case where there are construction personnel and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; The second sample image data and the second sample point cloud data are used as first training samples, and the first training samples and the second training labels are used to train the excavation area recognition model, so that the excavation area recognition model outputs a 3D model of the interior of the tunnel, a corresponding position of each excavation area in the 3D model of the interior of the sample tunnel, and a corresponding position of each excavation area on the input image data according to the input image data and point cloud data; Wherein, in the step of inputting the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, the excavation area recognition result includes a first recognition result and a second recognition result; the first recognition result includes a corresponding position of each excavation area on the first image data; the second recognition result includes a 3D model of the interior of the target tunnel and a corresponding position of each excavation area on the 3D model of the interior of the target tunnel; The method further comprises: A dynamic 3D construction progress display result of the target tunnel is generated according to the second recognition result corresponding to the data collected in multiple time periods.

8. An AI-based construction process management system, characterized in that: The system includes an image acquisition device, a laser scanning device and a data processing device; The image acquisition device is used to obtain first image data of the target tunnel during a non-construction period; The laser scanning device is used to obtain first point cloud data of the target tunnel in an unconstructed state; The data processing device is used to input the first image data and the first point cloud data into a pre-trained excavation area recognition model to obtain an excavation area recognition result generated by the excavation area recognition model, wherein the excavation area recognition result includes a corresponding position of each excavation area on the first image data; The image acquisition device is also used to obtain second image data during the construction period of the target tunnel; The data processing device is further used to identify the second image data through a content recognition module of a construction process recognition model, and determine first operation position data of each target construction object in the second image data, wherein the target construction object includes at least one of an engineering vehicle, a construction worker, and an internal support structure of a tunnel; Inputting the first operation position data and the excavation area identification result into the position matching module of the construction process identification model to obtain the corresponding relationship between each excavation area in the target tunnel and the construction position of each target construction object; Determine the construction process identification result of each excavation area according to the construction position correspondence relationship; Wherein, the data processing device is also used for: In the case where there are no construction workers and engineering vehicles in the sample tunnel, first sample image data and first sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; Reconstructing a 3D model according to the first sample point cloud data to obtain an internal 3D model of the sample tunnel; Marking the excavation area of ​​the internal 3D model to obtain a first excavation area marking result; Performing point matching on the first sample image data and the first sample point cloud data, and determining corresponding positions of various excavation areas on the first sample image data as first training labels according to the first excavation area labeling result; In the case where there are construction personnel and / or engineering vehicles in the sample tunnel, second sample image data and second sample point cloud data of the sample tunnel are acquired by an image acquisition device and a laser scanning device respectively; The second sample image data and the second sample point cloud data are used as first training samples, and the first training samples and the first training labels are used to train the excavation area recognition model, so that the excavation area recognition model determines the corresponding position of each excavation area on the input image data according to the input image data and point cloud data.

Images