Scene control method and system for unmanned tower crane

Abstract

The present invention discloses a scene control method and system for an unmanned tower crane, which constructs a first layer of images according to a laser point cloud set; constructs a second layer of images according to a screen image set; and constructs an overall image of an operating area according to the first layer of images and the second layer of images, thereby ensuring the accuracy of the overall image of the operating area. At the same time, according to the image optimization logic and the online optimization of multiple sub-images triggered by multiple sub-images, the online optimization of the overall image is realized, thereby constructing a working space between the tower crane part and the operating area according to the multiple sub-images, and defining the safety area of ​​the tower crane part based on the working space and the operating trajectory of the tower crane part, and defining the corresponding working scene according to the safety area, the working space, and the optimized multiple sub-images, thereby realizing multi-dimensional control of the safety area, the working space, and the optimized multiple sub-images, and ensuring the accuracy of the working scene, and being compatible with the overall consideration of the tower crane part, the laser radar, and the camera.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned tower cranes, and in particular to a scene control method and system for unmanned tower cranes. Background Art

[0002] With the development of science and technology, unmanned tower cranes are applied to construction sites and adopt intelligent control. At this time, the unmanned tower crane is equipped with a tower crane part, which serves as the working part of the unmanned tower crane. In the existing technology, unmanned tower cranes are applied to different work scenes. The working scenes of unmanned tower cranes are generally identified through captured images and only rely on cameras for control. It is incompatible with the overall consideration of the tower crane part, laser radar, and camera, which affects the accuracy of the work scene. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention provides a scene control method and system for an unmanned tower crane, which traverses the unmanned tower crane and locates the tower crane, laser radar and camera in the unmanned tower crane; based on the laser radar, a laser point cloud set of the operating area corresponding to the tower crane is collected, and a first layer of images is constructed according to the laser point cloud set; based on the camera, a screen image set of the operating area corresponding to the tower crane is collected, and a second layer of images is constructed according to the screen image set; an overall image of the operating area is constructed according to the first layer image and the second layer image, multi-dimensional control is performed based on the first layer image and the second layer image, and the accuracy of the overall image of the operating area is guaranteed.

[0004] At the same time, corresponding position parameters are defined based on the relative positions of the tower crane, laser radar and camera; multiple sub-images are divided based on the overall image of the operation area, and image optimization logic is defined according to the multiple sub-images, position parameters and the operation trajectory of the tower crane. The online optimization of the multiple sub-images is triggered according to the image optimization logic and the multiple sub-images, thereby realizing the online optimization of the overall image. The working space from the tower crane to the operation area is constructed according to the multiple sub-images, and the safety area of ​​the tower crane is defined based on the working space and the operation trajectory of the tower crane. The corresponding working scenes are defined according to the safety area, the working space and the optimized multiple sub-images, thereby realizing multi-dimensional control of the safety area, the working space and the optimized multiple sub-images, ensuring the accuracy of the working scene, and being compatible with the overall consideration of the tower crane, the laser radar and the camera.

[0005] The embodiment of the present invention provides a scene control method for an unmanned tower crane, which is applied to the scene control scenario of an unmanned tower crane;

[0006] The scene control method of the unmanned tower crane comprises:

[0007] Traverse the unmanned tower crane and locate the crane, lidar and camera in the unmanned tower crane;

[0008] Based on the laser radar, a laser point cloud set of the operation area corresponding to the tower crane is collected, and a first layer of image is constructed according to the laser point cloud set;

[0009] Based on the image set of the operation area corresponding to the tower crane part collected by the camera, a second layer of images is constructed according to the image set of the operation area;

[0010] An overall image of the operation area is constructed based on the first layer image and the second layer image, and corresponding position parameters are defined based on the relative positions of the crane, the laser radar, and the camera;

[0011] Divide the overall image of the operation area into multiple sub-images, define image optimization logic according to the multiple sub-images, position parameters and the operation trajectory of the tower crane part, and trigger online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images;

[0012] Collect multiple optimized sub-images, construct a workspace from the tower crane part to the operation area based on the multiple sub-images, define a safe area for the tower crane part based on the workspace and the operation trajectory of the tower crane part, and define a corresponding work scene based on the safe area, the workspace, and the multiple optimized sub-images.

[0013] Optionally, traversing the unmanned tower crane and locating the tower crane part, the laser radar and the camera in the unmanned tower crane include:

[0014] Collect the location of the unmanned tower crane and locate the unmanned tower crane according to the location of the unmanned tower crane;

[0015] Dividing the unmanned tower crane into multiple working modules based on the traversal of the unmanned tower crane;

[0016] Based on the identification of multiple working modules, the tower crane part, the laser radar and the camera are defined, and the tower crane part, the laser radar and the camera in the unmanned tower crane are located;

[0017] Associate the laser radar and the camera, and define the matching coefficient according to the model of the laser radar and the model of the camera;

[0018] The operating area corresponding to the tower crane part is defined according to the matching coefficient, the range detected by the laser radar, the range detected by the camera and the orientation of the tower crane part.

[0019] Optionally, the laser point cloud set of the operating area corresponding to the tower crane part is collected based on the laser radar, and the first layer of image is constructed according to the laser point cloud set, including:

[0020] The operation area corresponding to the fixed tower crane part;

[0021] Define the laser radar collection area based on the laser radar collection direction, the direction of the tower crane part and the operation area;

[0022] Define multiple laser point clouds according to the laser radar collection area and the laser radar collection mode;

[0023] Constructing a laser point cloud set of the operation area corresponding to the tower crane part according to the multiple laser point clouds;

[0024] Based on the screening of the laser point cloud set in the operation area, the abnormal point cloud is fixed, and the corresponding point cloud optimization mode is defined according to the abnormal point cloud and the outer contour point cloud;

[0025] An optimized laser point cloud set is defined according to the point cloud optimization mode and the laser point cloud set of the operation area, and a first layer of image is constructed according to the optimized laser point cloud set, the service life of the laser radar and the point cloud recognition logic.

[0026] Optionally, the step of collecting a screen image set of the operation area corresponding to the tower crane part based on the camera and constructing a second layer of images according to the screen image set includes:

[0027] The operation area corresponding to the fixed tower crane part;

[0028] Define the camera's acquisition area based on the camera's acquisition direction, the crane's orientation, and the operating area;

[0029] Define multiple screens according to the camera's acquisition area and the camera's acquisition mode;

[0030] Constructing a picture set of the operation area corresponding to the tower crane part according to the multiple pictures;

[0031] Based on the screening of the picture set of the operation area, the abnormal area is fixed, and the corresponding picture optimization mode is defined according to the abnormal area and the splicing track between multiple pictures;

[0032] An optimized picture set is defined according to the picture optimization mode and the picture set of the operation area, and a second layer of images is constructed according to the optimized picture set, the focus parameters of the camera and the picture recognition logic.

[0033] Optionally, the step of constructing an overall image of the operation area according to the first layer image and the second layer image, and defining corresponding position parameters based on the relative positions of the tower crane, the laser radar and the camera, includes:

[0034] Freeze the first layer image and the second layer image;

[0035] Perform image matching on the first layer image and the second layer image, and define corresponding image matching coefficients;

[0036] Constructing an overall image of the operation area according to the image matching coefficient, the area portion in the first layer image, and the area portion in the second layer image;

[0037] Freeze the crane, the laser radar, and the camera, and define the relative positions of the crane, the laser radar, and the camera;

[0038] The corresponding position parameters are defined based on the relative positions of the tower crane, the lidar and the camera, and the operating space where the tower crane is located.

[0039] Optionally, dividing the overall image of the operation area into multiple sub-images, defining image optimization logic according to the multiple sub-images, position parameters, and operation trajectories of the tower crane, and triggering online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images includes:

[0040] Forming a plurality of sub-images according to the division of the overall image of the operation area;

[0041] The abnormal part is defined based on the recognition of multiple sub-images, and the laser radar and camera are triggered to conduct targeted re-inspections according to the location of the abnormal part to optimize the abnormal part.

[0042] Optionally, the dividing of the overall image of the operation area into multiple sub-images, defining an image optimization logic according to the multiple sub-images, position parameters, and an operation trajectory of the tower crane, and triggering online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images, further includes:

[0043] Associating multiple sub-images, position parameters, and operation trajectories of crane parts;

[0044] Define image optimization logic according to multiple sub-images, position parameters, and operation trajectories of the crane parts;

[0045] Freeze the image optimization logic and associate the image optimization logic with multiple sub-images;

[0046] Online optimization of multiple sub-images is triggered according to the image optimization logic and the multiple sub-images.

[0047] Optionally, the collecting of the optimized multiple sub-images, constructing a workspace from the tower crane part to the operating area according to the multiple sub-images, defining a safety area of ​​the tower crane part based on the workspace and the operating trajectory of the tower crane part, and defining corresponding working scenes according to the safety area, the workspace, and the optimized multiple sub-images include:

[0048] Acquiring multiple optimized sub-images;

[0049] Constructing a workspace between the tower crane part and the operation area according to the multiple sub-images and the operation trajectory of the tower crane part;

[0050] Freeze the working space between the tower crane part and the operation area, and associate the operation trajectory based on the working space and the tower crane part;

[0051] The safety area of ​​the tower crane part is defined based on the working space and the operation trajectory of the tower crane part.

[0052] Optionally, the collecting of the optimized multiple sub-images, constructing a workspace from the tower crane to the operating area according to the multiple sub-images, defining a safety area of ​​the tower crane based on the workspace and the operating trajectory of the tower crane, and defining a corresponding working scene according to the safety area, the workspace, and the optimized multiple sub-images, further includes:

[0053] Associating safe areas, workspaces, and optimized multiple sub-images;

[0054] Defining a first scene parameter based on the safety area and the workspace, and defining a second scene parameter based on the safety area and the optimized plurality of sub-images;

[0055] The corresponding working scene is defined according to the second scene parameter, the first scene parameter and the scene matching table.

[0056] In addition, an embodiment of the present invention further provides a scene control system for an unmanned tower crane, the scene control system for the unmanned tower crane comprising:

[0057] The traversal module is used to traverse the unmanned tower crane and locate the tower crane, laser radar and camera in the unmanned tower crane;

[0058] The first layer image module is used to collect a laser point cloud set of the operation area corresponding to the tower crane part based on the laser radar, and construct a first layer image according to the laser point cloud set;

[0059] The second layer image module is used to collect a set of screen images of the operation area corresponding to the tower crane part based on the camera, and construct a second layer image according to the set of screen images;

[0060] A position module is used to construct an overall image of the operation area based on the first layer image and the second layer image, and define corresponding position parameters based on the relative positions of the tower crane, the laser radar and the camera;

[0061] An optimization module, used for dividing a plurality of sub-images based on an overall image of an operation area, defining an image optimization logic according to the plurality of sub-images, position parameters and an operation track of a tower crane part, and triggering online optimization of the plurality of sub-images according to the image optimization logic and the plurality of sub-images;

[0062] The work scene module is used to collect multiple optimized sub-images, construct a workspace between the tower crane part and the operation area based on the multiple sub-images, and define a safe area of ​​the tower crane part based on the workspace and the operation trajectory of the tower crane part, and define a corresponding work scene based on the safe area, the workspace, and the multiple optimized sub-images.

[0063] In an embodiment of the present invention, through the method in the embodiment of the present invention, the unmanned tower crane is traversed, and the tower crane part, laser radar and camera in the unmanned tower crane are located; based on the laser radar, a laser point cloud set of the operating area corresponding to the tower crane part is collected, and a first layer of images is constructed according to the laser point cloud set; based on the camera, a screen image set of the operating area corresponding to the tower crane part is collected, and a second layer of images is constructed according to the screen image set; based on the first layer image and the second layer image, an overall image of the operating area is constructed, and multi-dimensional control is performed based on the first layer image and the second layer image, and the accuracy of the overall image of the operating area is guaranteed.

[0064] At the same time, corresponding position parameters are defined based on the relative positions of the tower crane, laser radar and camera; multiple sub-images are divided based on the overall image of the operation area, and image optimization logic is defined according to the multiple sub-images, position parameters and the operation trajectory of the tower crane. The online optimization of the multiple sub-images is triggered according to the image optimization logic and the multiple sub-images, thereby realizing the online optimization of the overall image. The working space from the tower crane to the operation area is constructed according to the multiple sub-images, and the safety area of ​​the tower crane is defined based on the working space and the operation trajectory of the tower crane. The corresponding working scenes are defined according to the safety area, the working space and the optimized multiple sub-images, thereby realizing multi-dimensional control of the safety area, the working space and the optimized multiple sub-images, ensuring the accuracy of the working scene, and being compatible with the overall consideration of the tower crane, the laser radar and the camera. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] 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, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 It is a flow chart of a scene control method of an unmanned tower crane in an embodiment of the present invention;

[0067] Figure 2 is a flow chart of S11 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0068] Figure 3is a flow chart of S12 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0069] Figure 4 is a flow chart of S13 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0070] Figure 5 is a flow chart of S14 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0071] Figure 6 is a flow chart of S15 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0072] Figure 7 is a flow chart of S16 in the scene control method of the unmanned tower crane in the embodiment of the present invention;

[0073] Figure 8 It is a schematic diagram of the structural composition of the scene control system of the unmanned tower crane in the embodiment of the present invention;

[0074] Fig. 9 The figure is a hardware diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] See also Figures 1 to 9 , a scene control method for an unmanned tower crane is applied to a scene control scenario of an unmanned tower crane; the scene control method for an unmanned tower crane includes:

[0077] Step S11: traverse the unmanned tower crane and locate the tower crane part, laser radar and camera in the unmanned tower crane;

[0078] Step S12: collecting a laser point cloud set of the operation area corresponding to the tower crane part based on the laser radar, and constructing a first layer of image according to the laser point cloud set;

[0079] Step S13: Based on the camera collecting a screen image set of the operation area corresponding to the tower crane part, a second layer of images is constructed according to the screen image set;

[0080] Step S14: constructing an overall image of the operation area according to the first layer image and the second layer image, and defining corresponding position parameters based on the relative positions of the tower crane, the laser radar and the camera;

[0081] Step S15: dividing the overall image of the operation area into multiple sub-images, defining image optimization logic according to the multiple sub-images, position parameters, and the operation trajectory of the tower crane part, and triggering online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images;

[0082] Step S16: Collect multiple optimized sub-images, construct a workspace from the tower crane part to the operating area based on the multiple sub-images, and define a safe area of ​​the tower crane part based on the workspace and the operating trajectory of the tower crane part, and define a corresponding work scene based on the safe area, the workspace, and the multiple optimized sub-images.

[0083] In an embodiment of the present invention, through the method in the embodiment of the present invention, the unmanned tower crane is traversed, and the tower crane part, laser radar and camera in the unmanned tower crane are located; based on the laser radar, a laser point cloud set of the operating area corresponding to the tower crane part is collected, and a first layer of images is constructed according to the laser point cloud set; based on the camera, a screen image set of the operating area corresponding to the tower crane part is collected, and a second layer of images is constructed according to the screen image set; based on the first layer image and the second layer image, an overall image of the operating area is constructed, and multi-dimensional control is performed based on the first layer image and the second layer image, and the accuracy of the overall image of the operating area is guaranteed.

[0084] At the same time, corresponding position parameters are defined based on the relative positions of the tower crane, laser radar and camera; multiple sub-images are divided based on the overall image of the operation area, and image optimization logic is defined according to the multiple sub-images, position parameters and the operation trajectory of the tower crane. The online optimization of the multiple sub-images is triggered according to the image optimization logic and the multiple sub-images, thereby realizing the online optimization of the overall image. The working space from the tower crane to the operation area is constructed according to the multiple sub-images, and the safety area of ​​the tower crane is defined based on the working space and the operation trajectory of the tower crane. The corresponding working scenes are defined according to the safety area, the working space and the optimized multiple sub-images, thereby realizing multi-dimensional control of the safety area, the working space and the optimized multiple sub-images, ensuring the accuracy of the working scene, and being compatible with the overall consideration of the tower crane, the laser radar and the camera.

[0085] refer to Figure 2 , in step S11, traverse the unmanned tower crane and locate the tower crane part, laser radar and camera in the unmanned tower crane;

[0086] In the specific implementation process of the present invention, the specific steps may be:

[0087] S111: collecting the location of the unmanned tower crane, and locating the unmanned tower crane according to the location of the unmanned tower crane;

[0088] S112: Dividing the unmanned tower crane into multiple working modules based on the traversal of the unmanned tower crane;

[0089] S113: defining a tower crane part, a laser radar, and a camera based on the identification of multiple working modules, and locating the tower crane part, the laser radar, and the camera in the unmanned tower crane;

[0090] S114: Associating the laser radar and the camera, and defining a matching coefficient according to the model of the laser radar and the model of the camera;

[0091] S115: Define the operation area corresponding to the tower crane part according to the matching coefficient, the range detected by the laser radar, the range detected by the camera, and the orientation of the tower crane part.

[0092] In an embodiment of the present application, the location of the unmanned tower crane is collected, and the unmanned tower crane is positioned according to the location of the unmanned tower crane, so as to perform positioning detection on the unmanned tower crane, so as to trigger the traversal of the unmanned tower crane.

[0093] At this time, the unmanned tower crane is traversed, and multiple working modules in the unmanned tower crane are divided based on the traversal of the unmanned tower crane. Multiple working modules are introduced, and multiple working modules are synchronously identified. Then, based on the identification of multiple working modules, the tower crane part, lidar and camera are defined, and the tower crane part, lidar and camera in the unmanned tower crane are located, so as to facilitate further control of the tower crane part, lidar and camera.

[0094] Therefore, the laser radar and the camera are associated, and a matching coefficient is defined according to the model of the laser radar and the model of the camera, so as to facilitate matching of the model of the laser radar and the model of the camera, thereby achieving dynamic matching of the model of the laser radar and the model of the camera, and defining the matching coefficient. The matching degree between the laser radar and the camera is presented through the matching coefficient, thereby ensuring the interactive effect of the laser radar and the camera.

[0095] Furthermore, the operating area corresponding to the tower crane is defined according to the matching coefficient, the range detected by the laser radar, the range detected by the camera and the orientation of the tower crane. The matching coefficient, the range detected by the laser radar, the range detected by the camera and the orientation of the tower crane are introduced, and multi-dimensional control is performed on the matching coefficient, the range detected by the laser radar, the range detected by the camera and the orientation of the tower crane, thereby ensuring the accuracy of the operating area corresponding to the tower crane.

[0096] refer to Figure 3In step S12, based on the laser radar collecting the laser point cloud set of the operation area corresponding to the tower crane part, the first layer of image is constructed according to the laser point cloud set;

[0097] In the specific implementation process of the present invention, the specific steps may be:

[0098] S121: the operation area corresponding to the fixed tower crane part;

[0099] S122: defining a collection area of ​​the laser radar based on a collection direction of the laser radar, a direction of the tower crane, and an operation area;

[0100] S123: defining a plurality of laser point clouds according to a collection area of ​​the laser radar and a collection mode of the laser radar;

[0101] S124: constructing a laser point cloud set of the operation area corresponding to the tower crane part according to the multiple laser point clouds;

[0102] S125: fixing the abnormal point cloud based on the screening of the laser point cloud set in the operation area, and defining a corresponding point cloud optimization mode according to the abnormal point cloud and the outer contour point cloud;

[0103] S126: defining an optimized laser point cloud set according to the point cloud optimization mode and the laser point cloud set of the operation area, and constructing a first layer of image according to the optimized laser point cloud set, the service life of the laser radar and the point cloud recognition logic.

[0104] In an embodiment of the present application, the operating area corresponding to the tower crane part is frozen, and the operating area corresponding to the tower crane part is further processed. At this time, the collecting area of ​​the laser radar is defined based on the collecting direction of the laser radar, the orientation of the tower crane part and the operating area, which is compatible with the overall consideration of the collecting direction of the laser radar, the orientation of the tower crane part and the operating area, and multi-dimensional control is performed through the collecting direction of the laser radar, the orientation of the tower crane part and the operating area, thereby ensuring the collecting area of ​​the laser radar and realizing further management and control of the collecting area of ​​the laser radar.

[0105] Furthermore, multiple laser point clouds are defined according to the collection area and collection mode of the laser radar, and multiple laser point clouds are introduced, so as to construct a laser point cloud set of the operating area corresponding to the tower crane part according to the multiple laser point clouds, and the laser point cloud set of the operating area is controlled as a whole.

[0106] Therefore, based on the screening of the laser point cloud set in the operating area, the abnormal point cloud is fixed to facilitate the control of the abnormal point cloud, so as to define the corresponding point cloud optimization mode according to the abnormal point cloud and the outer contour point cloud, realize the multi-dimensional control of the abnormal point cloud and the outer contour point cloud, and ensure the accuracy of the point cloud optimization mode. At the same time, according to the point cloud optimization mode and the laser point cloud set of the operating area, the optimized laser point cloud set is defined, and the first layer of images is constructed according to the optimized laser point cloud set, the service life of the laser radar and the point cloud recognition logic, so as to ensure the accuracy of the first layer of images. At this time, the first layer of images serves as the primary image of the overall image of the operating area.

[0107] refer to Figure 4 In step S13, based on the picture image set of the operation area corresponding to the tower crane part collected by the camera, a second layer of images is constructed according to the picture image set;

[0108] In the specific implementation process of the present invention, the specific steps may be:

[0109] S131: the operation area corresponding to the fixed tower crane part;

[0110] S132: defining a collection area of ​​the camera based on a collection direction of the camera, a direction of the tower crane, and an operation area;

[0111] S133: defining a plurality of images according to a collection area of ​​the camera and a collection mode of the camera;

[0112] S134: constructing a picture set of the operation area corresponding to the tower crane part according to the multiple pictures;

[0113] S135: freezing an abnormal area based on the screening of the image set of the operation area, and defining a corresponding image optimization mode according to the abnormal area and the splicing trajectory between the multiple images;

[0114] S136: defining an optimized picture set according to the picture optimization mode and the picture set of the operation area, and constructing a second layer of images according to the optimized picture set, the focus parameters of the camera, and the picture recognition logic.

[0115] In an embodiment of the present application, the operation area corresponding to the tower crane part is frozen, and the operation area corresponding to the tower crane part is further processed. At this time, the camera's acquisition area is defined based on the camera's acquisition direction, the orientation of the tower crane part, and the operation area. The camera's acquisition direction, the orientation of the tower crane part, and the operation area are introduced, and multi-dimensional control of the camera's acquisition direction, the orientation of the tower crane part, and the operation area is achieved, so that multiple pictures can be defined according to the camera's acquisition area and the camera's acquisition mode, thereby ensuring the accuracy of multiple pictures.

[0116] Furthermore, a picture set of the operation area corresponding to the tower crane part is constructed based on multiple pictures, and a picture set of the operation area is introduced. The picture set of the operation area is considered as a whole. At the same time, the abnormal area is fixed based on the screening of the picture set of the operation area, so as to further control the abnormal area, thereby defining the corresponding picture optimization mode according to the splicing trajectory between the abnormal area and multiple pictures, and ensuring the accuracy of the picture optimization mode

[0117] Therefore, an optimized picture set is defined according to the picture optimization mode and the picture set of the operation area, and the optimized picture set is considered as a whole. At the same time, a second-layer image is constructed according to the optimized picture set, the focus parameters of the camera, and the picture recognition logic. The overall consideration of the optimized picture set, the focus parameters of the camera, and the picture recognition logic is compatible, and multi-dimensional control of the optimized picture set, the focus parameters of the camera, and the picture recognition logic is achieved, ensuring the accuracy of the second-layer image. At this time, the second-layer image serves as a high-level image of the overall image of the operation area, so as to facilitate the association of the second-layer image with the first-layer image.

[0118] refer to Figure 5 , S14: constructing an overall image of the operation area according to the first layer image and the second layer image, and defining corresponding position parameters based on the relative positions between the tower crane part, the laser radar and the camera;

[0119] In the specific implementation process of the present invention, the specific steps may be:

[0120] S141: freeze the first layer image and the second layer image;

[0121] S142: performing image matching on the first layer image and the second layer image, and defining corresponding image matching coefficients;

[0122] S143: constructing an overall image of the operation area according to the image matching coefficient, the area portion in the first layer image, and the area portion in the second layer image;

[0123] S144: freeze the tower crane part, the laser radar and the camera, and define the relative positions among the tower crane part, the laser radar and the camera;

[0124] S145: Define corresponding position parameters based on the relative positions among the tower crane, the laser radar and the camera, and the operating space where the tower crane is located.

[0125] In an embodiment of the present application, an unmanned tower crane is traversed, and a tower crane, a laser radar, and a camera in the unmanned tower crane are located; a laser point cloud set of an operating area corresponding to the tower crane is collected based on the laser radar, and a first layer of images is constructed based on the laser point cloud set; a screen image set of the operating area corresponding to the tower crane is collected based on the camera, and a second layer of images is constructed based on the screen image set; an overall image of the operating area is constructed based on the first layer of images and the second layer of images, and multi-dimensional control is performed based on the first layer of images and the second layer of images, and the accuracy of the overall image of the operating area is guaranteed.

[0126] At this time, the first layer image and the second layer image are frozen, and subsequent processing is performed on the first layer image and the second layer image. The first layer image and the second layer image are introduced, so as to perform image matching on the first layer image and the second layer image, and define corresponding image matching coefficients. Multi-dimensional control is performed based on the first layer image and the second layer image, and the image matching coefficients are output to facilitate further processing of the image matching coefficients.

[0127] Therefore, the overall image of the operating area is constructed according to the image matching coefficient, the regional part in the first layer image, and the regional part in the second layer image, so as to be compatible with the image matching coefficient, the regional part in the first layer image, and the regional part in the second layer image, thereby performing multi-dimensional control on the image matching coefficient, the regional part in the first layer image, and the regional part in the second layer image, thereby ensuring the accuracy of the overall image of the operating area.

[0128] Furthermore, the tower crane part, the laser radar and the camera are frozen, and the relative positions of the tower crane part, the laser radar and the camera are defined; based on the relative positions of the tower crane part, the laser radar and the camera and the operating space where the tower crane part is located, corresponding position parameters are defined, so as to control the position parameters so as to be compatible with the overall consideration of the relative positions of the tower crane part, the laser radar and the camera and the operating space where the tower crane part is located.

[0129] refer to Figure 6 , S15: dividing a plurality of sub-images based on the overall image of the operation area, defining an image optimization logic according to the plurality of sub-images, position parameters, and the operation trajectory of the tower crane part, and triggering online optimization of the plurality of sub-images according to the image optimization logic and the plurality of sub-images;

[0130] In the specific implementation process of the present invention, the specific steps may be:

[0131] S151: forming a plurality of sub-images according to the division of the overall image of the operation area;

[0132] S152: defining an abnormal part according to the recognition of the multiple sub-images, and triggering a laser radar and a camera for targeted review according to the location of the abnormal part, so as to optimize the abnormal part;

[0133] S153: Associating multiple sub-images, position parameters, and operation trajectories of the crane parts;

[0134] S154: defining image optimization logic according to the plurality of sub-images, the position parameters, and the operation trajectory of the crane part;

[0135] S155: freeze the image optimization logic, and associate the image optimization logic with the multiple sub-images;

[0136] S156: triggering online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images.

[0137] In an embodiment of the present application, multiple sub-images are formed according to the division of the overall image of the operating area, so as to introduce multiple sub-images and realize further processing of the multiple sub-images. At the same time, abnormal parts are defined according to the identification of the multiple sub-images, and the laser radar and the camera are triggered for targeted review according to the location of the abnormal parts to optimize the abnormal parts, thereby realizing autonomous optimization of the multiple sub-images and making full use of the laser radar and the camera for targeted review to ensure the integrity and accuracy of the multiple sub-images.

[0138] Therefore, multiple sub-images, position parameters and the operating trajectories of the tower crane are associated, and multi-dimensional control is performed on the multiple sub-images, position parameters and the operating trajectories of the tower crane, so as to define the image optimization logic according to the multiple sub-images, position parameters and the operating trajectories of the tower crane, thereby performing overall control on the multiple sub-images, position parameters and the operating trajectories of the tower crane, which is compatible with the overall consideration of the multiple sub-images, position parameters and the operating trajectories of the tower crane, and ensures the accuracy of the image optimization logic.

[0139] Furthermore, the image optimization logic is frozen, and the image optimization logic and multiple sub-images are associated, so as to dynamically associate the image optimization logic and the multiple sub-images, thereby triggering online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images, so as to ensure further processing of the multiple sub-images.

[0140] refer to Figure 7 , S16: collecting multiple optimized sub-images, constructing a workspace from the tower crane part to the operation area according to the multiple sub-images, and defining a safe area of ​​the tower crane part based on the workspace and the operation trajectory of the tower crane part, and defining a corresponding work scene according to the safe area, the workspace, and the multiple optimized sub-images;

[0141] In the specific implementation process of the present invention, the specific steps may be:

[0142] S161: Acquire multiple optimized sub-images; construct a workspace between the tower crane part and the operation area according to the multiple sub-images and the operation trajectory of the tower crane part;

[0143] S162: freeze the workspace between the tower crane part and the operation area, and associate the operation trajectory based on the workspace and the tower crane part;

[0144] S163: defining a safety area of ​​the tower crane part based on the workspace and the operation trajectory of the tower crane part;

[0145] S164: Associating the safe area, the workspace, and the optimized multiple sub-images;

[0146] S165: defining a first scene parameter based on the safety area and the workspace, and defining a second scene parameter based on the safety area and the optimized plurality of sub-images;

[0147] S166: Define a corresponding working scene according to the second scene parameter, the first scene parameter and the scene matching table.

[0148] In the specific implementation process of the present invention, corresponding position parameters are defined based on the relative positions among the tower crane, the laser radar and the camera; a plurality of sub-images are divided based on the overall image of the operation area, image optimization logic is defined according to the plurality of sub-images, position parameters and the operation trajectory of the tower crane, and online optimization of the plurality of sub-images is triggered according to the image optimization logic and the plurality of sub-images, thereby realizing online optimization of the overall image, thereby constructing a working space from the tower crane to the operation area according to the plurality of sub-images, defining a safe area of ​​the tower crane based on the working space and the operation trajectory of the tower crane, defining corresponding working scenes according to the safe area, the working space and the plurality of optimized sub-images, realizing multi-dimensional control of the safe area, the working space and the plurality of optimized sub-images, ensuring the accuracy of the working scenes, and being compatible with the overall consideration of the tower crane, the laser radar and the camera.

[0149] At this time, multiple optimized sub-images are collected; a working space between the tower crane part and the operating area is constructed according to the multiple sub-images and the operating trajectory of the tower crane part, and multi-dimensional control is performed on the multiple sub-images and the operating trajectory of the tower crane part, so as to introduce the working space between the tower crane part and the operating area, thereby further processing the working space between the tower crane part and the operating area.

[0150] Furthermore, the working space between the tower crane part and the operating area is frozen, and the operating trajectory based on the working space and the tower crane part is associated, so as to facilitate overall consideration of the working space and the operating trajectory of the tower crane part, thereby defining the safety area of ​​the tower crane part based on the working space and the operating trajectory of the tower crane part, introducing the safety area of ​​the tower crane part, and realizing precise control of the safety area of ​​the tower crane part.

[0151] Therefore, the safety area, workspace, and multiple optimized sub-images are associated and considered as a whole, so as to define the first scene parameters based on the safety area and the workspace, and define the second scene parameters based on the safety area and the multiple optimized sub-images, and then define the corresponding work scene according to the second scene parameters, the first scene parameters, and the scene matching table, thereby achieving multi-dimensional control of the safety area, workspace, and multiple optimized sub-images, ensuring the accuracy of the work scene, and being compatible with the overall consideration of the tower crane, lidar, and camera.

[0152] In another embodiment, the overall structure of the generalizable representation learning method framework proposed by the present invention includes synthetic scene generation and joint target learning. Synthetic scene generation is based on a simple and effective generative Bayesian network to model the scene-object distribution, and the generated synthetic scene is also enhanced by noise and occlusion. During the training process, two pre-tasks are jointly learned: a coarse-to-fine contrast learning task and a point cloud reconstruction task considering occlusion.

[0153] 1. Synthetic scene generation

[0154] We design a generative Bayesian network to model the scene-object distribution in the real world. , , are random variables representing scene type, object category, and object instance, respectively. We model the scene generator as a conditional categorical distribution, where the joint probability is given by:

[0155] ;

[0156] Prior probability and conditional probability And the conditional probability Following a categorical distribution:

[0157] ;

[0158] , , are the scene categories, object categories, and the total number of instances of each object category, respectively. When Take the kth scene type. If it is 0, it is not the kth scene type. Similarly, For In the th scene, take the lth object category. If it is 0, it is not the lth object category. , as well as

[0159] is the parameter of the category distribution (explanation: yes We learn these parameters from the actual dataset. Specifically, and Taking the maximum log-likelihood value of the joint probability, we can get:

[0160] , ;

[0161] in Indicates k The total number of occurrences of scenes, and Indicated in k In the scene l The total number of occurrences of object categories. As you can see, It is k The normalized value of the frequency of occurrence of a scene, and It is in k The normalized value of the frequency of occurrence of the lth object category in the scene. Then, we use the learned joint probability distribution to extract samples. First, we extract Extract a sample of a scene category from , then according to Extract N object categories Samples of Extract the corresponding object instance .

[0162] To generate synthetic scenes with sufficiently diverse real-world patterns, we perform occlusion simulation on point clouds. To increase the diversity, a distance-based occlusion simulation method is used. Specifically,

[0163] For each scene, a random observation point is selected in the 3D space of the scene, and for each object in the scene, 0% to 50% of the farthest points are randomly removed. This method improves our framework through its efficiency and occlusion diversity, i.e., each object has a random degree of incompleteness. The importance of this occlusion simulation is that it increases the realism and challenge of synthetic scenes. In the real world, we often cannot see all objects completely due to occlusions between objects, changes in perspective, and the shape and position of objects. By simulating this occlusion in synthetic scenes, we can make the trained model more robust and able to handle various occlusion situations that may occur in real applications. In addition, by randomly selecting the proportion of removed points, we are also able to simulate occlusions of different severity to different extents, which helps the model learn how to deal with different degrees of occlusion information. This diversity helps improve the generalization ability of the model and enables it to better adapt to different scenes and tasks.

[0164] 2. Joint Goal Learning

[0165] Our joint objective learning approach aims to enable representations that capture rich geometric information for efficient transfer to downstream tasks. Specifically, we propose a coarse-to-fine contrastive learning task to learn consistent fine-grained object representations, and an occlusion-considered reconstruction loss to preserve detailed geometric information.

[0166] 2.2.1 Coarse-to-fine contrastive learning task

[0167] Object-level contrastive learning is a method that uses object-level average features to calculate contrastive loss when dealing with synthetic scenes with occlusion. Given two corresponding scenes A and B with occlusion, we can first extract their features through the backbone network g(⋅), and then project φo(⋅) through a shared multi-layer perceptron (MLP) to obtain the projected output features X=φo(g(RX)), where X∈{A,B}. Next, in order to obtain object-level features, the output features of each instance are average pooled to obtain the average features at the instance level. The object-level contrastive loss can then be given by:

[0168] ;

[0169] in, K Represents the number of instances in each scene, Represents a collection of object features of different categories, and is a positive sample pair, is a negative sample pair, is the scale parameter.

[0170] Although object-level contrastive learning helps represent object-level geometric and semantic information, it may fail to capture fine-grained details of object parts, which are crucial in improving the ability of pre-trained models in certain downstream tasks. To overcome this problem, an object-aware point-level contrastive loss based on a relaxed point matching scheme is proposed. With our point-level contrastive learning, the model can produce consistent representations for specific parts of the same object. Specifically, point-level contrastive learning focuses on points or local regions inside objects instead of just the entire object. This approach allows the model to learn finer structures and features inside objects, which is particularly important for tasks that require accurate object part information, such as fine-grained classification, shape analysis, or 3D reconstruction. In the relaxed point matching scheme, we do not require each point to strictly correspond to the same point on another object, but allow a certain degree of flexibility. This flexibility enables the model to handle small deformations, occlusions, or viewpoint changes between objects, thereby generating more robust and consistent point-level representations. By combining object-level and point-level contrastive learning, our method is able to capture both the overall structure and fine-grained details of objects, providing more comprehensive and accurate feature representations for downstream tasks.

[0171] Point cloud reconstruction task considering occlusion

[0172] With the availability of complete synthetic scene-level point cloud data and motivated by the need to encourage the backbone network to preserve detailed geometric information, we propose a novel pre-training task: occlusion-aware reconstruction. To this end, we design a decoder to reconstruct the complete scene. Visualizations of the proposed decoder are shown in the supplementary material.

[0173] Specifically, given a complete point cloud of a scene X (X belongs to the set {A, B}) And the point cloud after occlusion , and the point-by-point features of the seed points , where s=256 is the feature dimension. We then use a shared MLP φc(·) to predict the coordinates and feature offsets: ;

[0174] Here, express The non-homogeneous coordinates, that is, the coordinates of a point in three-dimensional space, usually contain x、 y, z Three servings. and Represent the coarse-grained reconstruction result and coarse-grained features, respectively. These are usually obtained through some initial, less refined reconstruction process, serving as the starting point or basis for subsequent more refined reconstruction.

[0175] In this context, one might first estimate the general shape and features of a scene using a coarse-grained network or method, and then further refine these results using a fine-grained network or method to obtain a more accurate reconstruction.

[0176] The full dense point cloud is downsampled to obtain the true labels for the detailed point cloud and the coarse-grained point cloud respectively, and the reconstruction loss is constructed as the sum of the Chamfer distances of the coarse-grained point cloud and the detailed point cloud respectively. The reconstruction pipeline acts like an occlusion-aware point cloud variational autoencoder. This property enables the backbone network to maintain rich geometric information, thereby improving the effectiveness of the transfer results for specific downstream tasks. These tasks usually involve datasets with partially represented objects but require accurate localization, which is a common requirement in tasks such as 3D object detection. In this setting, by training the model to reconstruct the point cloud of occluded parts, the model learns to infer the complete scene structure from limited and possibly incomplete information. This ability is very useful for many scenarios in real-world applications, where sensor data is often corrupted by occlusion, noise or other factors. By using the occlusion-aware reconstruction task as pre-training, we can help the model learn more robust and general feature representations that may be more useful in subsequent tasks.

[0177] In an embodiment of the present invention, through the method in the embodiment of the present invention, the unmanned tower crane is traversed, and the tower crane part, laser radar and camera in the unmanned tower crane are located; based on the laser radar, a laser point cloud set of the operating area corresponding to the tower crane part is collected, and a first layer of images is constructed according to the laser point cloud set; based on the camera, a screen image set of the operating area corresponding to the tower crane part is collected, and a second layer of images is constructed according to the screen image set; based on the first layer image and the second layer image, an overall image of the operating area is constructed, and multi-dimensional control is performed based on the first layer image and the second layer image, and the accuracy of the overall image of the operating area is guaranteed.

[0178] At the same time, corresponding position parameters are defined based on the relative positions of the tower crane, laser radar and camera; multiple sub-images are divided based on the overall image of the operation area, and image optimization logic is defined according to the multiple sub-images, position parameters and the operation trajectory of the tower crane. The online optimization of the multiple sub-images is triggered according to the image optimization logic and the multiple sub-images, thereby realizing the online optimization of the overall image. The working space from the tower crane to the operation area is constructed according to the multiple sub-images, and the safety area of ​​the tower crane is defined based on the working space and the operation trajectory of the tower crane. The corresponding working scenes are defined according to the safety area, the working space and the optimized multiple sub-images, thereby realizing multi-dimensional control of the safety area, the working space and the optimized multiple sub-images, ensuring the accuracy of the working scene, and being compatible with the overall consideration of the tower crane, the laser radar and the camera.

[0179] See also Figure 8 , Figure 8 It is a schematic diagram of the structural composition of the scene control system of the unmanned tower crane in the embodiment of the present invention.

[0180] like Figure 8 As shown, a scene control system of an unmanned tower crane, the scene control system of the unmanned tower crane comprises:

[0181] A traversal module 21 is used to traverse the unmanned tower crane and locate the tower crane part, laser radar and camera in the unmanned tower crane;

[0182] The first layer image module 22 is used to collect a laser point cloud set of the operation area corresponding to the tower crane part based on the laser radar, and construct a first layer image according to the laser point cloud set;

[0183] The second layer image module 23 is used to collect a set of screen images of the operation area corresponding to the tower crane part based on the camera, and construct a second layer image according to the set of screen images;

[0184] A position module 24 is used to construct an overall image of the operation area according to the first layer image and the second layer image, and define corresponding position parameters based on the relative positions of the tower crane, the laser radar and the camera;

[0185] The optimization module 25 is used to divide the overall image of the operation area into multiple sub-images, define the image optimization logic according to the multiple sub-images, the position parameters and the operation trajectory of the tower crane part, and trigger the online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images;

[0186] The work scene module 26 is used to collect multiple optimized sub-images, construct a work space between the tower crane part and the operation area based on the multiple sub-images, and define a safe area of ​​the tower crane part based on the work space and the operation trajectory of the tower crane part, and define a corresponding work scene based on the safe area, the work space, and the multiple optimized sub-images.

[0187] See also Fig. 9 , refer to the following Fig. 9 The electronic intelligent device 40 according to this embodiment of the present invention is described. Fig. 9 The electronic intelligent device 40 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0188] like Fig. 9As shown, the electronic intelligent device 40 is in the form of a general computing intelligent device. The components of the electronic intelligent device 40 may include but are not limited to: at least one processing unit 41, at least one storage unit 42, and a bus 43 connecting different system components (including the storage unit 42 and the processing unit 41).

[0189] The storage unit stores program codes, which can be executed by the processing unit 41, so that the processing unit 41 executes the steps according to various exemplary embodiments of the present invention described in the above “Embodiment Method” section of this specification.

[0190] The storage unit 42 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 421 and / or a cache memory unit 422 , and may further include a read-only memory unit (ROM) 423 .

[0191] The storage unit 42 may also include a program / utility 424 having a set (at least one) of program modules 425, such program modules 425 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0192] Bus 43 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0193] The electronic smart device 40 may also communicate with one or more external smart devices (e.g., keyboards, pointing smart devices, Bluetooth smart devices, etc.), one or more smart devices that enable a user to interact with the electronic smart device 40, and / or any smart device that enables the electronic smart device 40 to communicate with one or more other computing smart devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 44. Furthermore, the electronic smart device 40 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 45. Fig. 9 As shown, the network adapter 45 communicates with other modules of the electronic intelligent device 40 via the bus 43. It should be understood that although Fig. 9 Not shown, other hardware and / or software modules may be used in conjunction with the electronic intelligent device 40, including but not limited to: microcode, intelligent device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup planning systems.

[0194] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing intelligent device (which can be a personal computer, a server, a terminal device, or a network intelligent device, etc.) to execute the method according to the implementation of the present disclosure.

[0195] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium can include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. In addition, it stores computer program instructions, and when the computer program instructions are executed by a computer, the computer executes the above method.

[0196] In addition, the scene control method and system of the unmanned tower crane provided by the embodiment of the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principle and implementation mode of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation mode and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A scene control method for an unmanned tower crane, characterized in that: Applied to scene control scenarios of unmanned tower cranes; The scene control method of the unmanned tower crane comprises: Traverse the unmanned tower crane and locate the crane, lidar and camera in the unmanned tower crane; Based on the laser radar, a laser point cloud set of the operation area corresponding to the tower crane is collected, and a first layer of image is constructed according to the laser point cloud set; Based on the image set of the operation area corresponding to the tower crane part collected by the camera, a second layer of images is constructed according to the image set of the operation area; An overall image of the operation area is constructed based on the first layer image and the second layer image, and corresponding position parameters are defined based on the relative positions of the crane, the laser radar, and the camera; Divide the overall image of the operation area into multiple sub-images, define image optimization logic according to the multiple sub-images, position parameters and operation trajectories of the crane part, and trigger online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images; Collect multiple optimized sub-images, construct a workspace from the crane part to the operation area based on the multiple sub-images, define a safe area of ​​the crane part based on the workspace and the operation trajectory of the crane part, and define a corresponding work scene based on the safe area, the workspace, and the multiple optimized sub-images; The traversal of the unmanned tower crane and positioning of the tower crane part, the laser radar and the camera in the unmanned tower crane include: Collect the location of the unmanned tower crane and locate the unmanned tower crane according to the location of the unmanned tower crane; Dividing the unmanned tower crane into multiple working modules based on the traversal of the unmanned tower crane; Defining the crane part, the laser radar and the camera based on the identification of multiple working modules, and locating the crane part, the laser radar and the camera in the unmanned tower crane; Associate the laser radar and the camera, and define the matching coefficient according to the model of the laser radar and the model of the camera; The operating area corresponding to the tower crane part is defined according to the matching coefficient, the range detected by the laser radar, the range detected by the camera, and the orientation of the tower crane part.

2. The scene control method of an unmanned tower crane according to claim 1, characterized in that: The laser point cloud set of the operation area corresponding to the tower crane part is collected based on the laser radar, and the first layer of image is constructed according to the laser point cloud set, including: The operation area corresponding to the fixed tower crane part; Define the laser radar collection area based on the laser radar collection direction, the direction of the tower crane part and the operation area; Define multiple laser point clouds according to the laser radar collection area and the laser radar collection mode; Constructing a laser point cloud set of the operation area corresponding to the tower crane part according to the multiple laser point clouds; Based on the screening of the laser point cloud set in the operation area, the abnormal point cloud is fixed, and the corresponding point cloud optimization mode is defined according to the abnormal point cloud and the outer contour point cloud; An optimized laser point cloud set is defined according to the point cloud optimization mode and the laser point cloud set of the operation area, and a first layer of image is constructed according to the optimized laser point cloud set, the service life of the laser radar and the point cloud recognition logic.

3. The scene control method of an unmanned tower crane according to claim 2, characterized in that: The step of collecting a screen image set of the operation area corresponding to the tower crane part based on the camera and constructing a second layer of images according to the screen image set includes: The operation area corresponding to the fixed tower crane part; Define the camera's acquisition area based on the camera's acquisition direction, the crane's orientation, and the operating area; Define multiple screens according to the camera's acquisition area and the camera's acquisition mode; Constructing a picture set of the operation area corresponding to the tower crane part according to the multiple pictures; Based on the screening of the picture set of the operation area, the abnormal area is fixed, and the corresponding picture optimization mode is defined according to the abnormal area and the splicing track between multiple pictures; An optimized picture set is defined according to the picture optimization mode and the picture set of the operation area, and a second layer of images is constructed according to the optimized picture set, the focus parameters of the camera and the picture recognition logic.

4. The scene control method of an unmanned tower crane according to claim 3, characterized in that: The overall image of the operation area is constructed according to the first layer image and the second layer image, and corresponding position parameters are defined based on the relative positions of the tower crane, the laser radar and the camera, including: Freeze the first layer image and the second layer image; Perform image matching on the first layer image and the second layer image, and define corresponding image matching coefficients; Constructing an overall image of the operation area according to the image matching coefficient, the area portion in the first layer image, and the area portion in the second layer image; Freeze the crane, the laser radar, and the camera, and define the relative positions of the crane, the laser radar, and the camera; The corresponding position parameters are defined based on the relative positions of the tower crane, the lidar and the camera, and the operating space where the tower crane is located.

5. The scene control method of an unmanned tower crane according to claim 4, characterized in that: The method of dividing the overall image of the operation area into multiple sub-images, defining the image optimization logic according to the multiple sub-images, the position parameters and the operation trajectory of the tower crane part, and triggering the online optimization of the multiple sub-images according to the image optimization logic and the multiple sub-images includes: Forming a plurality of sub-images according to the division of the overall image of the operation area; The abnormal part is defined based on the recognition of multiple sub-images, and the laser radar and camera are triggered to conduct targeted re-inspections according to the location of the abnormal part to optimize the abnormal part.

6. The scene control method of an unmanned tower crane according to claim 5, characterized in that: The method further includes dividing the overall image of the operation area into a plurality of sub-images, defining an image optimization logic according to the plurality of sub-images, position parameters, and the operation trajectory of the crane part, and triggering online optimization of the plurality of sub-images according to the image optimization logic and the plurality of sub-images. Associating multiple sub-images, position parameters, and operation trajectories of crane parts; Define image optimization logic according to multiple sub-images, position parameters, and operation trajectories of the crane parts; Freeze the image optimization logic and associate the image optimization logic with multiple sub-images; Online optimization of multiple sub-images is triggered according to the image optimization logic and the multiple sub-images.

7. The scene control method of an unmanned tower crane according to claim 6, characterized in that: The method includes collecting the optimized multiple sub-images, constructing a workspace between the crane part and the operation area according to the multiple sub-images, defining a safe area of ​​the crane part based on the workspace and the operation trajectory of the crane part, and defining a corresponding work scene according to the safe area, the workspace, and the optimized multiple sub-images, including: Collect multiple optimized sub-images; construct a workspace between the tower crane part and the operation area according to the multiple sub-images and the operation trajectory of the tower crane part; Freeze the working space between the tower crane part and the operation area, and associate the operation trajectory based on the working space and the tower crane part; The safety area of ​​the tower crane part is defined based on the working space and the operation trajectory of the tower crane part.

8. The scene control method of an unmanned tower crane according to claim 7, characterized in that: The method further comprises: collecting the optimized multiple sub-images, constructing a workspace between the tower crane part and the operation area according to the multiple sub-images, defining a safe area of ​​the tower crane part based on the workspace and the operation trajectory of the tower crane part, and defining a corresponding work scene according to the safe area, the workspace and the optimized multiple sub-images. Associating safe areas, workspaces, and optimized multiple sub-images; Defining a first scene parameter based on the safe area and the workspace, and defining a second scene parameter based on the safe area and the optimized plurality of sub-images; The corresponding working scene is defined according to the second scene parameter, the first scene parameter and the scene matching table.

9. A scene control system for an unmanned tower crane, characterized in that: The scene control system of the unmanned tower crane is applied to the scene control method of the unmanned tower crane as claimed in any one of claims 1 to 8, and the scene control system of the unmanned tower crane includes: The traversal module is used to traverse the unmanned tower crane and locate the tower crane, laser radar and camera in the unmanned tower crane; The first layer image module is used to collect a laser point cloud set of the operation area corresponding to the tower crane part based on the laser radar, and construct a first layer image according to the laser point cloud set; The second layer image module is used to collect a set of screen images of the operation area corresponding to the tower crane part based on the camera, and construct a second layer image according to the set of screen images; A position module is used to construct an overall image of the operation area based on the first layer image and the second layer image, and define corresponding position parameters based on the relative positions of the tower crane, the laser radar and the camera; An optimization module, used for dividing a plurality of sub-images based on an overall image of an operation area, defining an image optimization logic according to the plurality of sub-images, position parameters and an operation track of a tower crane part, and triggering online optimization of the plurality of sub-images according to the image optimization logic and the plurality of sub-images; A work scene module is used to collect multiple optimized sub-images, construct a work space from the tower crane to the operation area based on the multiple sub-images, define a safe area for the tower crane based on the work space and the operation trajectory of the tower crane, and define a corresponding work scene based on the safe area, the work space, and the multiple optimized sub-images; The traversal of the unmanned tower crane and positioning of the tower crane part, the laser radar and the camera in the unmanned tower crane include: Collect the location of the unmanned tower crane and locate the unmanned tower crane according to the location of the unmanned tower crane; Dividing the unmanned tower crane into multiple working modules based on the traversal of the unmanned tower crane; Defining the crane part, the laser radar and the camera based on the identification of multiple working modules, and locating the crane part, the laser radar and the camera in the unmanned tower crane; Associate the laser radar and the camera, and define the matching coefficient according to the model of the laser radar and the model of the camera; The operating area corresponding to the tower crane part is defined according to the matching coefficient, the range detected by the laser radar, the range detected by the camera, and the orientation of the tower crane part.

Images