NDT algorithm-based nuclear power plant inspection robot laser radar real-time point cloud matching method and system

Abstract

The invention relates to the field of nuclear power plant inspection device automatic navigation, in particular to a nuclear power plant inspection robot laser radar real-time point cloud matching method and system based on an NDT algorithm. The system comprises a multi-dimensional point cloud analysis tool, an environment information interaction module, a sensor interaction module and an environment information generation module. The method comprises the following steps: S1, loading an environment; s2, performing real-time scanning; s3, optimization calculation is carried out; and S4, generating a result. The terrain feature point matching effect of the inspection device in the autonomous navigation process is improved, and the navigation quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of automatic navigation of nuclear power plant inspection devices, and in particular to a real-time point cloud matching method and system of laser radar of a nuclear power plant inspection robot based on an NDT algorithm. Background Art

[0002] When the inspection robot performs navigation tasks in a nuclear power plant building, it needs to use lidar to complete map construction and autonomous navigation and obstacle avoidance in indoor closed environments. If the point cloud data does not match well, it will be difficult to achieve ideal positioning accuracy, which can easily lead to problems such as navigation failure.

[0003] In nuclear power plant buildings, it is necessary to improve the matching accuracy of point cloud data collected by the lidar to help nuclear power plant inspection robots improve the accuracy of environmental information feedback from the lidar during autonomous navigation, and form an autonomous navigation terrain feature matching solution for this special scenario. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a real-time point cloud matching method and system for a laser radar of a nuclear power plant inspection robot based on an NDT algorithm, thereby improving the accuracy of environmental information feedback from the laser radar.

[0005] A nuclear power plant inspection robot laser radar real-time point cloud matching system based on NDT algorithm, including:

[0006] Multi-dimensional point cloud analysis tool, which performs point cloud matching calculation based on the data of the environmental information interaction module and the sensor interaction module;

[0007] The environmental information interaction module is used to obtain detailed inspection environmental information, store the read data in the environmental information library, and perform data interaction with other functional modules through the data transmission interface;

[0008] The sensor interaction module is used to obtain real-time laser radar data through the sensor data interface. This module performs data interaction with other functional modules through the data transmission interface;

[0009] The environment information generation module is used to generate and export the final normalized result data. This module performs data interaction with other functional modules through the data transmission interface.

[0010] In a specific embodiment of the present invention, the multi-dimensional point cloud analysis tool includes: a global map analysis submodule, a point cloud matching calculation submodule, an optimization result feedback submodule and a first data transmission interface;

[0011] The global map parsing submodule includes a first process information library, a map file parsing operator, and a data normalization operator; by loading a preset nuclear island global map and performing parsing, map identification data information is obtained and an environmental information vector is generated;

[0012] The point cloud matching calculation submodule includes the second process information library, point cloud parsing operation operator, NDT operator, and overall probability optimization operator; it performs point cloud data loading, target point cloud grid area division, and corresponding area mean and covariance calculation operations, and finally feeds back the overall probability optimization result of the global NDT distribution to generate a result data vector;

[0013] The optimization result feedback submodule includes a third process information base and a result information normalization operator; it is used to store result data and generate a result matrix;

[0014] The first data transmission interface includes a data loading and data exporting interface, which is used to perform data communication with other functional modules in the system.

[0015] In a specific implementation of the present invention, the environment information interaction module includes: a first environment data loading interface, an environment information library, and a second data transmission interface.

[0016] In a specific embodiment of the present invention, the sensor interaction module includes: a device interface, a sensor detailed information library, a sensor data interface and a third data transmission interface.

[0017] In a specific implementation of the present invention, the environment information generation module includes: a second environment data loading interface, a point cloud result parsing submodule and a fourth data transmission interface.

[0018] The present invention provides a real-time point cloud matching method of a laser radar of a nuclear power plant inspection robot based on an NDT algorithm, comprising the following steps:

[0019] Step S1: environment loading;

[0020] Step S2: real-time scanning;

[0021] Step S3: optimization calculation;

[0022] Step S4: Result generation.

[0023] In a specific embodiment of the present invention, the step S1 specifically includes:

[0024] Step S1-1: The first environment data loading interface in the environment information interaction module executes an inspection environment map information reading operation;

[0025] Step S1-2: transferring the obtained inspection environment map information to the environment information database;

[0026] Step S1-3: calling the data transmission interface to transmit the data from the environment information base in the environment information interaction module to the first process information base in the global map parsing submodule in the multi-dimensional point cloud parsing tool.

[0027] In a specific embodiment of the present invention, the step S2 specifically includes:

[0028] Step S2-1: calling the device interface of the sensor interaction module to perform a data reading operation; Step S2-2: performing structured matching on the read device data, and transferring the data in batches to the sensor detailed information library of the sensor interaction module to complete the initialization of the device detailed parameters;

[0029] Step S2-3: calling the sensor data interface to open the sensor real-time value data acquisition channel;

[0030] Step S2-4: Execute data transmission to the process information library in the global map parsing submodule and the point cloud matching calculation submodule in the multi-dimensional point cloud parsing tool through the third data transmission interface of the sensor interaction module.

[0031] In a specific embodiment of the present invention, step S3 specifically includes:

[0032] Step S3-1: Load the data stored in the first process information library in the global map parsing submodule, obtain map identification data information, and generate an environment information vector;

[0033] Step S3-2: Loading the data stored in the second process information library in the point cloud matching calculation submodule;

[0034] Step S3-3: Call the point cloud parsing operator to process the stored data in step S3-1 and step S3-2, perform point cloud structured data loading, target point cloud grid area division and corresponding area mean and covariance calculation operations, if the optimization constraint preset conditions are not met, return to the target point cloud grid area division step to re-execute the matching optimization calculation; finally feedback the overall probability optimization result of the global NDT distribution, and generate the result data vector V G-Map ;

[0035] Step S3-4: Execute result data storage and generate result matrix M r ;

[0036] Step S3-5: Execute the result information data transmission operation to the environment information generation module through the data transmission interface.

[0037] In a specific embodiment of the present invention, step S4 specifically includes:

[0038] Step S4-1: Calling the second environment data loading interface through the environment information generation module to obtain global map information, and at the same time calling the data transmission interface to obtain the point cloud matching optimization calculation result fed back in step S3-3;

[0039] Step S4-2: the point cloud result parsing submodule is called by the environment information generation module to parse the point cloud matching optimization calculation result obtained in step S4-1 and assemble and convert it into a standardized data format;

[0040] Step S4-3: Feedback the optimization result parsing data file information completed in step S4-2 through the data transmission interface.

[0041] Compared with the prior art, the real-time point cloud matching method and system of the nuclear power plant inspection robot laser radar based on the NDT algorithm of the present invention can optimize and improve the regional matching probability of the acquired point cloud data by using a packaged multi-dimensional point cloud analysis tool, thereby improving the terrain feature point matching effect of the inspection device during autonomous navigation and improving the navigation quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram showing the system structure;

[0043] Figure 2 It shows the structure of multi-dimensional point cloud analysis tool;

[0044] Figure 3 Representation method flow chart. DETAILED DESCRIPTION

[0045] In order to further understand the present invention, the embodiments of the present invention are described below in conjunction with examples, but it should be understood that these descriptions are only for further illustrating the features and advantages of the present invention, rather than limiting the present invention.

[0046] The embodiment of the present invention discloses a nuclear power plant inspection robot laser radar real-time point cloud matching system based on NDT algorithm, such as Figure 1 As shown, it includes: a multi-dimensional point cloud parsing tool, an environmental information interaction module, a sensor interaction module and an environmental information generation module;

[0047] Multi-dimensional point cloud analysis tool, used to perform point cloud matching calculation based on the data of the environmental information interaction module and the sensor interaction module;

[0048] like Figure 2 As shown, the multi-dimensional point cloud parsing tool includes: a global map parsing submodule, a point cloud matching calculation submodule, an optimization result feedback submodule and a first data transmission interface;

[0049] The global map parsing submodule includes a first process information library, a map file parsing operator, and a data normalization operator; by loading a preset nuclear island global map and performing parsing, map identification data information is obtained and an environmental information vector is generated;

[0050] The point cloud matching calculation submodule includes the second process information library, point cloud analysis operation operator, NDT operator, and overall probability optimization operator; it performs point cloud data loading, target point cloud grid area division, and corresponding area mean and covariance calculation operations, and finally feeds back the overall probability optimization result of the global NDT distribution to generate the result data vector V pc-r ;

[0051] The optimization result feedback submodule includes the third process information library and the result information normalization operator; it is used to store the result data and generate the result matrix M r ;

[0052] The first data transmission interface includes a data loading and data exporting interface, which is used to perform data communication with other functional modules in the system.

[0053] The environmental information interaction module is used to obtain detailed inspection environmental information, store the read data in the environmental information library, and perform data interaction with other functional modules through the data transmission interface;

[0054] The environmental information interaction module includes: a first environmental data loading interface, an environmental information library and a second data transmission interface.

[0055] The sensor interaction module is used to obtain real-time laser radar data through the sensor data interface. This module can perform data interaction with other functional modules through the data transmission interface;

[0056] The sensor interaction module includes: a device interface, a sensor detailed information library, a sensor data interface and a third data transmission interface.

[0057] The environment information generation module is used to generate and export the final normalized result data. This module can perform data interaction with other functional modules through the data transmission interface.

[0058] The environment information generation module includes: a second environment data loading interface, a point cloud result parsing submodule and a fourth data transmission interface.

[0059] The embodiment of the present invention also discloses a real-time point cloud matching method of a laser radar of a nuclear power plant inspection robot based on an NDT algorithm, such as Figure 3 As shown, the following steps are included:

[0060] Step S1: environment loading;

[0061] The specific steps include:

[0062] Step S1-1: The first environment data loading interface in the environment information interaction module executes an inspection environment map information reading operation;

[0063] Step S1-2: transferring the obtained inspection environment map information to the environment information database;

[0064] Step S1-3: calling the data transmission interface to transmit data from the environment information library in the environment information interaction module to the process information library in the global map parsing submodule in the multi-dimensional point cloud parsing tool.

[0065] Step S2: real-time scanning;

[0066] Specifically include:

[0067] Step S2-1: calling the device interface of the sensor interaction module to perform a data reading operation; the device refers to a device for scanning the environment map carried by the robot including a laser radar; the read data refers to the original map point cloud data obtained by scanning the environment map carried by the robot including the laser radar;

[0068] Step S2-2: perform structured matching on the read device data, and transfer the data in batches to the sensor detailed information library of the sensor interaction module to complete the initialization of the device detailed parameters;

[0069] Step S2-3: calling the sensor data interface to open the sensor real-time value data acquisition channel;

[0070] The sensor interface can adapt to the raw data generated by various sensors including distance sensors, visual sensors, and lidar. In this example, it is lidar point cloud data.

[0071] Step S2-4: Execute data transmission to the process information library in the global map parsing submodule and the point cloud matching calculation submodule in the multi-dimensional point cloud parsing tool through the third data transmission interface of the sensor interaction module.

[0072] Step S3: optimization calculation;

[0073] Specifically include:

[0074] Step S3-1: Load the data stored in the first process information library in the global map parsing submodule, obtain map identification data information, and generate an environment information vector;

[0075] Step S3-2: Loading the data stored in the second process information library in the point cloud matching calculation submodule;

[0076] Step S3-3: Call the point cloud parsing operator to process the stored data in step S3-1 and step S3-2, perform point cloud structured data loading, target point cloud grid area division and corresponding area mean and covariance calculation operations, if the optimization constraint preset conditions are not met, return to the target point cloud grid area division step to re-execute the matching optimization calculation; finally feedback the overall probability optimization result of the global NDT distribution, and generate the result data vector V G-Map ;

[0077] Step S3-4: Execute result data storage and generate result matrix M r ;

[0078] Step S3-5: Execute the result information data transmission operation to the environment information generation module through the data transmission interface.

[0079] Step S4: Result generation.

[0080] Specifically include:

[0081] Step S4-1: Calling the second environment data loading interface through the environment information generation module to obtain global map information, and at the same time calling the data transmission interface to obtain the point cloud matching optimization calculation result fed back in step S3-3;

[0082] Step S4-2: the point cloud result parsing submodule is called by the environment information generation module to parse the point cloud matching optimization calculation result obtained in step S4-1 and assemble and convert it into a standardized data format;

[0083] Step S4-3: Feedback the optimization result parsing data file information completed in step S4-2 through the data transmission interface.

[0084] The above embodiments are only used to help understand the method and core idea of ​​the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

[0085] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A nuclear power plant inspection robot laser radar real-time point cloud matching system based on NDT algorithm, characterized in that: include: Multi-dimensional point cloud analysis tool, which performs point cloud matching calculation based on the data of the environmental information interaction module and the sensor interaction module; The environmental information interaction module is used to obtain detailed inspection environmental information, store the read data in the environmental information library, and perform data interaction with other functional modules through the data transmission interface; The sensor interaction module is used to obtain real-time data of the laser radar through the sensor data interface. This module performs data interaction with other functional modules through the data transmission interface; The environment information generation module is used to generate and export the final normalized result data. This module performs data interaction with other functional modules through the data transmission interface.

2. The real-time point cloud matching system of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 1 is characterized in that: The multi-dimensional point cloud analysis tool includes: a global map analysis submodule, a point cloud matching calculation submodule, an optimization result feedback submodule and a first data transmission interface; The global map parsing submodule includes a first process information library, a map file parsing operator, and a data normalization operator; by loading a preset nuclear island global map and performing parsing, map identification data information is obtained and an environmental information vector is generated; The point cloud matching calculation submodule includes the second process information library, point cloud parsing operation operator, NDT operator, and overall probability optimization operator; it performs point cloud data loading, target point cloud grid area division, and corresponding area mean and covariance calculation operations, and finally feeds back the overall probability optimization result of the global NDT distribution to generate a result data vector; The optimization result feedback submodule includes a third process information base and a result information normalization operator; it is used to store result data and generate a result matrix; The first data transmission interface includes a data loading and data exporting interface, which is used to perform data communication with other functional modules in the system.

3. The real-time point cloud matching system of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 1 is characterized in that: The environmental information interaction module includes: a first environmental data loading interface, an environmental information library and a second data transmission interface.

4. The real-time point cloud matching system of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 1 is characterized in that: The sensor interaction module includes: a device interface, a sensor detailed information library, a sensor data interface and a third data transmission interface.

5. The nuclear power plant inspection robot laser radar real-time point cloud matching system based on NDT algorithm according to claim 1 is characterized in that: The environment information generation module includes: a second environment data loading interface, a point cloud result parsing submodule and a fourth data transmission interface.

6. A real-time point cloud matching method for laser radar of nuclear power plant inspection robot based on NDT algorithm, characterized in that: The following steps are involved: Step S1: environment loading; Step S2: real-time scanning; Step S3: optimization calculation; Step S4: Result generation.

7. The real-time point cloud matching method of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 6 is characterized in that: The step S1 specifically includes: Step S1-1: The first environment data loading interface in the environment information interaction module executes an inspection environment map information reading operation; Step S1-2: transferring the obtained inspection environment map information to the environment information database; Step S1-3: calling the data transmission interface to transmit the data from the environment information base in the environment information interaction module to the first process information base in the global map parsing submodule in the multi-dimensional point cloud parsing tool.

8. The real-time point cloud matching method of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 6 is characterized in that: The step S2 specifically includes: Step S2-1: calling the device interface of the sensor interaction module to perform a data reading operation; Step S2-2: performing structured matching on the read device data, and transferring the data in batches to the sensor detailed information library of the sensor interaction module to complete the initialization of the device detailed parameters; Step S2-3: calling the sensor data interface to open the sensor real-time value data acquisition channel; Step S2-4: Execute data transmission to the process information library in the global map parsing submodule and the point cloud matching calculation submodule in the multi-dimensional point cloud parsing tool through the third data transmission interface of the sensor interaction module.

9. The real-time point cloud matching method of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 6 is characterized in that: The step S3 specifically includes: Step S3-1: Load the data stored in the first process information library in the global map parsing submodule, obtain map identification data information, and generate an environment information vector; Step S3-2: Loading the data stored in the second process information library in the point cloud matching calculation submodule; Step S3-3: Call the point cloud parsing operator to process the stored data in step S3-1 and step S3-2, perform point cloud structured data loading, target point cloud grid area division and corresponding area mean and covariance calculation operations, if the optimization constraint preset conditions are not met, return to the target point cloud grid area division step to re-execute the matching optimization calculation; finally feedback the overall probability optimization result of the global NDT distribution, and generate the result data vector V G-Map ; Step S3-4: Execute result data storage and generate result matrix M r ; Step S3-5: Execute the result information data transmission operation to the environment information generation module through the data transmission interface.

10. The real-time point cloud matching method of laser radar for nuclear power plant inspection robot based on NDT algorithm according to claim 9 is characterized in that: The step S4 specifically includes: Step S4-1: Calling the second environment data loading interface through the environment information generation module to obtain global map information, and at the same time calling the data transmission interface to obtain the point cloud matching optimization calculation result fed back in step S3-3; Step S4-2: the point cloud result parsing submodule is called by the environment information generation module to parse the point cloud matching optimization calculation result obtained in step S4-1 and assemble and convert it into a standardized data format; Step S4-3: Feedback the optimization result parsing data file information completed in step S4-2 through the data transmission interface.

Images