A container-based deployment power line point cloud data cloud automatic processing system

Abstract

The application discloses a kind of power line point cloud data cloud end automatic processing system based on container deployment, it is related to power grid technical field, and the online processing of point cloud data is realized by cloud storage and container deployment service, by the cooperation of data storage module, parameter database module, spatial database cache module and data processing module, the efficiency of data storage, extraction and processing is improved, by first the container deployment of required data processing service is carried out, then the automatic flow of processing service is realized by the mode of message queue, by building kubernetes cluster, unified resource allocation is carried out to container deployment, so that existing resources are maximized, the configuration requirement of hardware is reduced, and finally the automatic online processing of power line point cloud data is realized.

Description

Technical Field

[0001] This invention relates to the field of power grid technology, and in particular to a cloud-based automatic processing system for power line point cloud data based on container deployment. Background Technology

[0002] By utilizing laser scanning technology and high-precision aerial surveying technology for quantitative measurement and qualitative analysis and early warning of the power transmission line environment, key distance measurement data and efficient obstacle measurement can be generated. Real-time online location and monitoring of power transmission line faults can be achieved, allowing for timely detection and elimination of line defects and major hidden dangers, saving a significant amount of manpower and resources, and improving the safety of inspection work.

[0003] Point cloud data collected by lidar inspection of power transmission line channels is used for detection and analysis after preprocessing. The traditional method is to select point cloud data on the local disk through client software, load it into the 3D scene in the interface, and then process it, such as mapping, thinning, point cloud classification, and detection.

[0004] Traditional methods for processing and analyzing point cloud data collected by lidar during power transmission line inspections have the following problems:

[0005] 1. Issues with data transmission between the client and the cloud

[0006] 1) The processing relies on the client program to complete, and the point cloud data is managed in the local disk. The process data and result data generated during the processing have data security risks and do not meet the power grid company's closed-loop requirements for data.

[0007] 2) In the traditional way, the unclassified point cloud, point cloud engineering (process data), and classified point cloud data generated during the point cloud data processing are stored in mobile hard drives or disk arrays. For grassroots units of power companies, the maintenance cost is high, the data utilization rate is low after the data is stored on the medium, and when point cloud data is needed for work, it is necessary to copy the point cloud data to the computer before it can be used.

[0008] 3) The local configuration (graphics card, memory, disk, etc.) is relatively high. The client software uses the local video memory, memory and CPU to render and load point cloud data from the disk on the local machine, which has high requirements for machine performance and occupies a large amount of local disk space.

[0009] 4) Existing point cloud processing systems and software tools store point cloud data in LAS files on network storage. During processing, the data needs to be downloaded from network storage, loaded and processed locally, and then uploaded again. This process has several drawbacks: firstly, downloading point cloud data consumes significant bandwidth; secondly, downloading data locally is insecure; and thirdly, changes during processing can generate multiple related point cloud datasets, leading to excessive network storage space consumption.

[0010] 2. Issues with high-concurrency rendering when processing massive amounts of data.

[0011] The existing inspection process only focuses on the process itself. In power applications, when processing large amounts of point cloud data (massive data) from hundreds to tens of thousands of kilometers, the 3dtiles technology used in the current technology has less JSON metadata. The tiles embed geometric and attribute information, which consumes a lot of bandwidth, resulting in slow page rendering, resource exhaustion, and inability to process data without high-concurrency rendering.

[0012] 3. The problem of low automation in handling service calls is that they are done manually.

[0013] 1) Low level of automation: During the data processing process of the client software, most tasks are completed manually. Due to different function settings of the client software, the operation is difficult.

[0014] 2) The existing processing flow provides inspection result storage services, point cloud data processing and analysis services, and inspection result application services. The calls between these services require manual intervention. It can only process the data and cannot allocate or transfer the data to be processed, thus failing to achieve large-scale data processing. Summary of the Invention

[0015] To address the aforementioned issues, this invention proposes a container-based automatic cloud processing system for power line point cloud data. This system enables online processing of point cloud data through cloud storage and container deployment services. The system improves the efficiency of data storage, retrieval, and processing by cooperating with data storage, parameter database, spatial database caching, and data processing modules. It first deploys the necessary data processing services in containers and then uses message queues to automate the flow of processing services. A Kubernetes cluster is built to uniformly allocate resources for container deployment, maximizing the utilization of existing resources and reducing hardware configuration requirements, ultimately achieving automatic online processing of power line point cloud data.

[0016] This invention proposes a container-based automatic cloud processing system for power line point cloud data, characterized by comprising the following modules: a data storage module, a parameter database module, a spatial database caching module, and a data processing module. Specifically,

[0017] Data storage module: Connected to the spatial database cache module, it is used to store the collected point cloud data or the point cloud data processed by the data processing module in the distributed object storage, and to perform spatial data compression on the stored point cloud data according to the spatial index;

[0018] Spatial database caching module: Connected to the data storage module and the data processing module, it is used to cache point cloud data when it is transferred between the data storage module and the data processing module;

[0019] Parameter database module: Connected to the data processing module, it is used to store the parameters required for the data processing process;

[0020] Data processing module: Connected to the spatial database cache module and parameter database module, it performs point cloud data archiving, classification, thinning, and detection based on the point cloud data in the spatial database cache module and the parameters in the parameter database module. It implements point cloud data archiving, classification, thinning, and detection through container deployment, and uses a message queue approach to generate data processing message commands according to task requirements. The point cloud data processing system controls task initiation. After the previous task is completed, the task status is sent to the point cloud data processing system, which then initiates the next task, achieving automated point cloud data processing.

[0021] Furthermore, the method for establishing the spatial index is as follows: indexing is performed according to the direction of the transmission line, the data is stored in real time in a distributed object storage, and spatial data compression is performed on it.

[0022] Furthermore, the task queue adopts a message queue mode, and its message queue parameters are as follows: select Rabbit as the message queue, select publish / subscribe mode, select direct-connect switch as the switch type, and select topic switch as the processing service.

[0023] Furthermore, the categorization, classification, thinning, and detection specifically refer to:

[0024] The process of marking out the data involves extracting each tower and marking its number, and then separating the two towers from the original point cloud data by a certain width.

[0025] The classification is based on the power grid specifications, which divides the point cloud data points into categories according to power equipment, such as poles, conductors, ground wires, jumpers, and insulator strings, and categories according to the features in the channel, such as buildings, low-voltage power lines, railway tracks, roads, rivers, and vegetation.

[0026] Thinning is the process of geometrically reducing the number of points in a cloud to a reasonable range.

[0027] The inspection involves analyzing whether the distance between the target object and the power transmission equipment meets the safety requirements of the power grid.

[0028] Furthermore, the automated processing of point cloud data specifically includes:

[0029] Step S10: Start the processing service: After the processing personnel confirm that there are no problems with the point cloud data, they start the processing process in the point cloud data processing system.

[0030] Step S11: Start the mapping service: After the automatic processing service is started, the point cloud data processing system sends a message to start the mapping service. After receiving the message, the mapping service starts the mapping program to map the point cloud data.

[0031] Step S12: Start the classification service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends the start classification service message. After receiving the message, the classification service starts the automatic classification program to classify the point cloud data.

[0032] Step S13: Start the thinning service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the thinning service. After receiving the message, the thinning service starts the thinning program to thin the point cloud data.

[0033] Step S14: Start the detection service: After the thinning process is completed, the thinning service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the detection service. After receiving the message, the detection service starts the detection program to detect the point cloud data.

[0034] Step S15: End Processing Service: After the detection program is completed, the detection service, as the publisher, sends the message to the point cloud data processing system through the switch. The point cloud data processing system then ends the processing flow and updates the processing status in the interactive interface.

[0035] Furthermore, the container deployment method is as follows:

[0036] 1) Download the base image to be used, and use the Dockerfile file to create a Docker image that can run independently, and upload it to the image repository;

[0037] 2) Pull the image from the image repository address and configure the resources required for container deployment;

[0038] 3) After the configuration parameters are set, the cluster allocates the resources required for the deployment and processing services, schedules them to nodes with sufficient resources, and starts them.

[0039] 4) If access to the processing service is required, an access proxy is generated through kube-proxy, and the entry address is provided by the Service;

[0040] 5) Configure resources for container deployment by setting up a Kubernetes cluster.

[0041] A method for using a container-based power line point cloud data cloud-based automatic processing system as described in any of the preceding claims, characterized by comprising the following steps:

[0042] Step S20: Data storage: Store the collected point cloud data in a distributed object storage, and perform spatial data compression on the stored point cloud data according to the spatial index;

[0043] Step S21: Target data extraction: Download the target data to be processed from the data storage module to the spatial database cache module, and then extract the target data from the spatial database cache module and load it into the data processing module;

[0044] Step S22: Parameter extraction: Load the parameters needed for data processing from the parameter database module into the data processing module;

[0045] Step S23: Data Processing: According to the set data processing services, each data processing service is started sequentially using a message queue to process the point cloud data.

[0046] Step S24: Processing result caching: Upload the processed point cloud data to the spatial database caching module;

[0047] Step S25: Processing Result Storage: After all data processing services are completed, the data processing results cached in the spatial database caching module are uploaded to the data storage module.

[0048] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0049] First, by combining cloud storage and container deployment, all point cloud data is processed in the cloud, reducing security risks and improving the efficiency of point cloud data storage, retrieval, and processing. By building a Kubernetes cluster, unified resource allocation for container deployment is achieved, maximizing the utilization of existing resources, reducing hardware configuration requirements, and solving the problem of data transmission between the client and the cloud.

[0050] Secondly, by pre-deploying data processing services into containers, then setting and selecting the required data processing services, loading the containers corresponding to each data processing service, and using message queues to automate each processing step, the system automatically allocates and transfers the data to be processed, thereby achieving automated processing of large-scale data and solving the problem of low automation caused by manual invocation of processing services.

[0051] Third, by compressing the stored point cloud data according to spatial index, the data download and upload time is shortened, data storage space is saved, and the problem of high-concurrency rendering of massive data is solved. Attached Figure Description

[0052] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0053] Figure 1 This is a schematic diagram of the cloud-based automatic processing system for power line point cloud data based on container deployment according to the present invention.

[0054] Figure 2 This is a flowchart illustrating the usage method of the container-based power line point cloud data cloud automatic processing system of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Those skilled in the art should understand that the following specific embodiments or implementation methods are a series of optimized configurations listed to further explain the specific content of the invention. These configuration methods can be combined or used in conjunction with each other, unless the invention explicitly states that some or a specific embodiment or implementation method cannot be associated with or used in conjunction with other embodiments or implementation methods. Furthermore, the following specific embodiments or implementation methods are merely optimized configurations and are not intended to limit the scope of protection of the invention.

[0057] The specific embodiments of the present invention will now be described with reference to the accompanying drawings (tables). Specific Implementation Method 1

[0059] To address the problems of low data transmission efficiency and high resource consumption between the client and the cloud in existing technologies, inability to perform high-concurrency rendering when processing massive amounts of data, and low automation due to manual invocation of processing services, this invention achieves online processing of point cloud data through cloud storage and container deployment services. By cooperating with data storage and retrieval modules, parameter database modules, spatial database caching modules, and data processing modules, the efficiency of data storage, retrieval, and processing is improved, and the hardware configuration requirements are reduced. Task transmission is achieved through message queues to realize data allocation and data flow. By compressing the stored point cloud data according to spatial indexes, the data download and upload time is shortened, and data storage space is saved, ultimately realizing automatic online processing of power line point cloud data.

[0060] Figure 1 This is a schematic diagram of the cloud-based automatic processing system for power line point cloud data based on container deployment according to the present invention.

[0061] This invention proposes a container-based automatic cloud processing system for power line point cloud data, characterized by comprising the following modules: a data storage module, a parameter database module, a spatial database caching module, and a data processing module.

[0062] 1. Data storage module: Connected to the spatial database cache module, it stores the collected point cloud data or the point cloud data processed by the data processing module in the distributed object storage, and performs spatial data compression on the stored point cloud data according to the spatial index.

[0063] Point cloud data is stored as file objects in a distributed object storage system. After uploading, the stored point cloud data is compressed according to its spatial index. The spatial index is created with full consideration of the characteristics of power transmission lines, indexing along the transmission line direction. This index data is editable and renders quickly. Users can edit attributes, set category codes, and delete data by selecting points. Point cloud attribute information is set and stored in the distributed object storage in real time, and spatial data compression is applied to quickly load point cloud data during processing, improving the speed of point cloud retrieval and analysis.

[0064] 2. Spatial Database Cache Module: Connected to the data storage module and data processing module, it is used to cache point cloud data during the transfer between the data storage module and the data processing module.

[0065] The spatial database caching module serves as a connection channel between the data storage module and the data processing module, used for temporary data storage. Although point cloud data acquisition and processing are completed in the cloud, avoiding the problem of excessive bandwidth consumption during cloud-to-local data transmission, data processing or acquisition still requires a certain amount of time, and frequent data processing consumes significant resources. The spatial database caching module stores this data, and when the same data is requested again, it is retrieved directly from the cache instead of from the data storage module, thereby reducing resource consumption and further improving response speed.

[0066] The spatial database caching module implements data caching in two directions. First, it downloads target data from the data storage module and caches it in the spatial database caching module, while the data processing module retrieves point cloud data from the spatial database caching module for processing. Second, the point cloud data processed by the data processing module is cached in the spatial database caching module and then uploaded to the data storage module for point cloud data storage.

[0067] 3. Parameter Database Module: Connected to the data processing module, it is used to store the parameters required for the data processing process.

[0068] The data processing module retrieves point cloud data from the spatial database cache module and the required parameters from the parameter database module according to the service task requirements, and completes the corresponding processing services.

[0069] 4. Data Processing Module: Connected to the spatial database cache module and parameter database module, it performs point cloud data archiving, classification, thinning, and detection based on the point cloud data in the spatial database cache module and the parameters in the parameter database module. It utilizes container deployment for point cloud data archiving, classification, thinning, and detection, and employs a message queue approach. Data processing message commands are generated according to task requirements, and the point cloud data processing system controls task initiation. After the previous task is completed, its status is sent to the point cloud data processing system, which then initiates the next task, achieving automated point cloud data processing.

[0070] The data processing module includes point cloud data processing services, specifically: mapping, classification, thinning, and detection. These services include, but are not limited to, the above-mentioned services. Additional services can be added as needed to achieve automated online processing of point clouds for different functions.

[0071] The mapping service extracts each tower and marks its number. Then, it segments the area between two towers from the original point cloud data according to a certain width, saving the resulting segment as a new point cloud data file. For example, when an aircraft carrying a lidar device scans a power transmission line from the air, it scans multiple tower sections (#1-#10) at a time. Based on the provided tower coordinates, the point cloud data for #1-#10 is mapped and divided into #1-#2, #2-#3, ..., #9-#10.

[0072] The classification service, based on the specifications of the power grid, categorizes point cloud data points into power equipment such as poles, conductors, ground wires, jumpers, and insulator strings, and into features within the channel such as buildings, low-voltage power lines, railway tracks, roads, rivers, and vegetation.

[0073] Point thinning service geometrically reduces the number of point clouds to a reasonable range. During data acquisition, variations in LiDAR frequency and speed can result in tens of millions, or even hundreds of millions, of point cloud data points. This increased point density can negatively impact the efficiency of data analysis and rendering. Before analysis, the point cloud data is loaded into the point cloud processing software client, and each point is thinned using spatial, random, or Octree methods, thus geometrically reducing the number of point clouds to a reasonable range.

[0074] The inspection service analyzes whether the distance between the target object and the power transmission equipment meets the safety requirements of the power grid.

[0075] The automated processing of point cloud data specifically includes:

[0076] Step S10: Start the processing service: After the processing personnel confirm that there are no problems with the point cloud data, they start the processing process in the point cloud data processing system.

[0077] Step S11: Start the mapping service: After the automatic processing service is started, the point cloud data processing system sends a message to start the mapping service. After receiving the message, the mapping service starts the mapping program to map the point cloud data.

[0078] Step S12: Start the classification service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends the start classification service message. After receiving the message, the classification service starts the automatic classification program to classify the point cloud data.

[0079] Step S13: Start the thinning service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the thinning service. After receiving the message, the thinning service starts the thinning program to thin the point cloud data.

[0080] Step S14: Start the detection service: After the thinning process is completed, the thinning service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the detection service. After receiving the message, the detection service starts the detection program to detect the point cloud data. After the detection program is completed, the detection service, as the publisher, sends the message to the point cloud data processing system through the switch and updates the processing status in the interactive interface.

[0081] Step S15: End Processing Service: After the detection program is completed, the detection service, as the publisher, sends the message to the point cloud data processing system through the switch. The point cloud data processing system then ends the processing flow and updates the processing status in the interactive interface.

[0082] All of the above processing services are implemented through container deployment. The required processing services are selected from other services such as grading, classification, thinning, and detection, and each selected processing service is pre-deployed into a container before the processing service is started.

[0083] Container deployment refers to packaging software code and all necessary components (such as libraries, frameworks, and other dependencies) together and isolating them in their own "containers." After packaging the application into a container that can be moved between platforms and infrastructures, the application can run successfully simply by moving the container to a location. In this invention, service data is stored as images, and multiple instances of the processing service image and cache service image are deployed and run in containers. Service calls between these instances are implemented through a scheduling service.

[0084] This invention manages container deployments by building a Kubernetes cluster. A Kubernetes cluster mainly consists of a master node and worker nodes, each with different components installed. The master node primarily includes the following service components: APIServer, Scheduler, ControllerManager, and Etcd. The node primarily includes the following components: Kubelet and KubeProxy.

[0085] master: The control plane of the cluster, responsible for cluster decision-making (management);

[0086] ApiServer: The sole entry point for resource operations, receiving user-input commands and providing mechanisms for authentication, authorization, API registration, and discovery;

[0087] Scheduler: Responsible for cluster resource scheduling, scheduling Pods to the corresponding node nodes according to the predefined scheduling strategy;

[0088] ControllerManager: Responsible for maintaining the state of the cluster, such as application deployment schedules, fault detection, automatic scaling, and rolling updates;

[0089] Etcd: Responsible for storing information about various resource objects in the cluster, stored in a key / value manner, and all Kubernetes cluster data is stored here;

[0090] node: The data plane of the cluster, responsible for providing the runtime environment (doing the work) for containers;

[0091] Kubelet: Responsible for maintaining the lifecycle of containers, that is, creating, updating and destroying containers by controlling Docker. It checks the health status of nodes at a fixed frequency and reports it to the API Server. This status is recorded in the status of the Node object.

[0092] KubeProxy: Responsible for providing service discovery and load balancing within the cluster.

[0093] Deploying a processing service, such as a classification service, includes the following processes:

[0094] 1) Build the image: Download the base image, such as daocloud.io / library / java:8u40-jdk, and use the application of the category module to create a Docker image that can run independently using the Dockerfile you wrote, and upload it to the image repository.

[0095] 2) Configuration parameters: Pull the image through the image repository address, configure the resources required by the container, such as the number of replicas. You can set it according to your needs. Here it is set to 2. The minimum CPU cores are reserved and the maximum is set to 2 cores. The minimum memory used is 200M and the maximum is 1.5G. The external port and internal port are both set to 9080. You can also set the external port and internal port to be different.

[0096] 3) Resource Scheduling: After the parameters are set, the cluster allocates the resources required for deploying and processing services, scheduling them to nodes with sufficient resources and starting them. The installation request for the classification service is sent to the apiServer component on the master node. The apiServer component calls the scheduler component to determine which node the service should be installed on. The scheduler component reads information about each node from etcd, selects a node according to a certain algorithm, and sends the result back to the apiServer. The apiServer calls the controller-manager to schedule the node to install the classification service. After receiving the instruction, the kubelet notifies Docker, and then Docker starts a pod for the classification service, enabling the classification service to run.

[0097] 4) Create an access proxy: kube-proxy acts as a proxy for pod access, enabling access to categorized module services. A Service aggregates multiple pods providing the same service and provides a unified entry point. Accessing the Service's entry point leads to the subsequent pod services.

[0098] 5) Service Maintenance: ControllerManager maintains the load balancing cluster services. When one of the category service pods is in use or its node crashes, the Service will forward requests to another available pod. If access requests increase, the cluster will adjust based on the resource utilization of existing pods, automatically deploying one or more new pods (based on an existing cluster node with 32GB of memory and 8 CPUs, a single node can run a maximum of about 30 category pods), thus ensuring the concurrency requirements of the category module. When requests decrease, the number of pods will also be automatically reduced, saving server resources and achieving dynamic adjustment. Category services deployed via Deployment will detect updates or changes to their configuration files and automatically restart to update.

[0099] After the processing services are deployed, they need to be executed sequentially. This invention uses a message queue to realize the automatic flow of each processing service, thereby achieving automated online processing of point cloud data.

[0100] Message queues enable the automatic initiation of each processing step, thereby automating data processing.

[0101] The publish / subscribe pattern is a pattern where a message sent by a producer is received by multiple consumers. Because a single message is consumed and processed by multiple consumers, it is also called the broadcast pattern or the one-to-many pattern.

[0102] To automate the process, Rabbit was chosen as the message queue, the publish / subscribe mode was selected, the topic exchange was selected as the processing service, and messages were handed over to queues that conform to the routing pattern. Specific Implementation Method Two

[0104] The present invention also proposes a method for using a container-based power line point cloud data cloud automatic processing system as described in any of the specific embodiments. Figure 2 This is a flowchart illustrating the usage method of the container-based power line point cloud data cloud automatic processing system of the present invention. It includes the following steps:

[0105] Step S20: Data storage: Store the collected point cloud data in a distributed object storage, and perform spatial data compression on the stored point cloud data according to the spatial index;

[0106] Step S21: Target data extraction: Download the target data to be processed from the data storage module to the spatial database cache module, and then extract the target data from the spatial database cache module and load it into the data processing module;

[0107] Step S22: Parameter extraction: Load the parameters needed for data processing from the parameter database module into the data processing module;

[0108] Step S23: Data Processing: According to the set data processing services, each data processing service is started sequentially using a message queue to process the point cloud data.

[0109] Step S24: Processing result caching: Upload the processed point cloud data to the spatial database caching module;

[0110] Step S25: Processing Result Storage: After all data processing services are completed, the data processing results cached in the spatial database caching module are uploaded to the data storage module.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A container-based automatic cloud processing system for power line point cloud data, characterized in that, It includes the following modules: data storage module, parameter database module, spatial database caching module, and data processing module. Specifically, Data storage module: connected to the spatial database cache module, used to store the collected point cloud data or the point cloud data processed by the data processing module in the distributed object storage, and to perform spatial data compression on the stored point cloud data according to the spatial index; wherein, the method for establishing the spatial index is: indexing according to the direction of the transmission line, storing in the distributed object storage in real time, and performing spatial data compression on it; Spatial database caching module: Connected to the data storage module and the data processing module, it is used to cache point cloud data when it is transferred between the data storage module and the data processing module; The parameter database module connects to the data processing module and stores the parameters required for data processing. The data processing module connects to the spatial database cache module and the parameter database module. Based on the point cloud data in the spatial database cache module and the parameters in the parameter database module, it performs point cloud data mapping, classification, thinning, and detection. Point cloud data mapping, classification, thinning, and detection are implemented through container deployment. A message queue is used to generate data processing message commands according to task requirements. The point cloud data processing system controls the start of tasks. After the previous task is completed, the task status is sent to the point cloud data processing system, which then starts the next task, achieving automated processing of point cloud data. The mapping, classification, thinning, and detection specifically involve: The process of marking out the data involves extracting each tower and marking its number, and then separating the two towers from the original point cloud data by a certain width. The classification is based on the power grid specifications, which divides the point cloud data points into categories according to power equipment, such as poles, conductors, ground wires, jumpers, and insulator strings, and categories according to the features in the channel, such as buildings, low-voltage power lines, railway tracks, roads, rivers, and vegetation. Thinning is the process of geometrically reducing the number of points in a cloud to a reasonable range. The inspection involves analyzing whether the distance between the target object and the power transmission equipment meets the safety requirements of the power grid.

2. The container-based power line point cloud data cloud automatic processing system according to claim 1, characterized in that, The parameters for the message queue are as follows: select Rabbit as the message queue, select publish / subscribe mode as the mode, select direct-connect switch as the switch type, and select topic switch as the processing service.

3. The container-based power line point cloud data cloud automatic processing system according to claim 1, characterized in that, Automated point cloud data processing specifically includes: Step S10: Start the processing service: After the processing personnel confirm that there are no problems with the point cloud data, they start the processing process in the point cloud data processing system. Step S11: Start the mapping service: After the automatic processing service is started, the point cloud data processing system sends a message to start the mapping service. After receiving the message, the mapping service starts the mapping program to map the point cloud data. Step S12: Start the classification service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the classification service. After receiving the message, the classification service starts the automatic classification program to classify the point cloud data. Step S13: Start the thinning service: After the classification process is completed, the classification service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the thinning service. After receiving the message, the thinning service starts the thinning program to thin the point cloud data. Step S14: Start the detection service: After the thinning process is completed, the thinning service, as the publisher, sends the message to the point cloud data processing system through the switch. After receiving the message, the point cloud data processing system updates the processing status in the interactive interface and sends a message to start the detection service. After receiving the message, the detection service starts the detection program to detect the point cloud data. After the detection program is completed, the detection service, as the publisher, sends the message to the point cloud data processing system through the switch and updates the processing status in the interactive interface. Step S15: End Processing Service: After the detection program is completed, the detection service, as the publisher, sends the message to the point cloud data processing system through the switch. The point cloud data processing system then ends the processing flow and updates the processing status in the interactive interface.

4. The container-based power line point cloud data cloud automatic processing system according to claim 1, characterized in that, The container deployment method is as follows: 1) Download the base image to be used, and use the Dockerfile file to create a Docker image that can run independently, and upload it to the image repository; 2) Pull the image from the image repository address and configure the resources required for container deployment; 3) After the configuration parameters are set, the cluster allocates the resources required for the deployment and processing services, schedules them to nodes with sufficient resources, and starts them. 4) If access to the processing service is required, an access proxy is generated through kube-proxy, and the entry address is provided by the Service; 5) Configure resources for container deployment by setting up a Kubernetes cluster.

5. A method of using a container-based power line point cloud data cloud automatic processing system as described in any one of claims 1-4, characterized in that, Includes the following steps, Step S20: Data storage: Store the collected point cloud data in a distributed object storage, and perform spatial data compression on the stored point cloud data according to the spatial index; Step S21: Target data extraction: Download the target data to be processed from the data storage module to the spatial database cache module, and then extract the target data from the spatial database cache module and load it into the data processing module; Step S22: Parameter extraction: Load the parameters needed for data processing from the parameter database module into the data processing module; Step S23: Data Processing: According to the set data processing services, each data processing service is started sequentially using a message queue to process the point cloud data. Step S24: Processing result caching: Upload the processed point cloud data to the spatial database caching module; Step S25: Processing Result Storage: After all data processing services are completed, the data processing results cached in the spatial database caching module are uploaded to the data storage module.

Images