Total station automatic monitoring implementation method and system combined with convolutional neural network

Abstract

The invention discloses a total station automatic monitoring implementation method and system combined with a convolutional neural network, and the method comprises the steps: continuously obtaining the coordinate information of a monitoring target in real time based on the total station automatic monitoring system, dividing the monitoring target into different sections based on the spatial position relation between the monitoring target and a total station, and carrying out the automatic monitoring of the total station. Based on historical measurement data, a convolutional neural network is used for performing model training to predict a current measurement result, deviation conditions of a prediction result and an actual result in a comparison judgment section are comprehensively analyzed and judged, and effective filtering of error early warning of the automatic monitoring system can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of total station automated monitoring, and in particular to a method and system for implementing total station automated monitoring in combination with a convolutional neural network. Background Art

[0002] my country has complex geological conditions, and the ecological environment has deteriorated in recent years. Serious natural disasters such as earthquakes, landslides, and mud-rock flows have occurred frequently. At the same time, with the rapid development of my country's water conservancy, transportation, construction and other industries, traditional manual photogrammetry and leveling elevation measurement methods cannot meet the current high-frequency, long-term, and time-sensitive requirements for safety deformation monitoring of large-scale projects and their surrounding environments.

[0003] The total station measurement robot can measure angles with an accuracy of 0.5 seconds, has high distance measurement accuracy, and can measure the local coordinate system of the station with an accuracy of millimeters. Combined with the Internet of Things, embedded and edge computing technologies can customize microcontrollers with computing and storage functions for the total station. The monitoring system built on this basis can achieve unattended automated monitoring and meet the current urgent needs of safe deformation monitoring. However, the measurement results of the total station measurement robot are greatly affected by the natural environment, and false warnings may occur when used for field automated monitoring. To address this problem, there are currently two main solutions: equipping the total station with an electronic temperature barometer, correcting the external environmental parameters each time the data is collected, or actively manually reviewing when the warning is triggered. These two solutions have the following problems: 1. Simply correcting the external environmental parameters during collection cannot effectively eliminate errors in rainy and foggy weather. 2. Simply using manual review will increase labor costs and deviate from the original intention of automated monitoring. Summary of the invention

[0004] The present application provides a method and system for realizing automatic monitoring of a total station in combination with a convolutional neural network, so as to solve the problems existing in the prior art solutions, such as the inability to effectively eliminate errors, possible false warnings, and high costs.

[0005] According to the first aspect, an embodiment provides a method for realizing automatic monitoring of a total station combined with a convolutional neural network, the method comprising:

[0006] Determine the deployment plan based on the actual situation of the monitoring target, and deploy monitoring prisms, reference prisms and total stations;

[0007] Based on the spatial position relationship between the monitoring target and the total station, the sections to which all monitoring prisms belong are divided;

[0008] Performing automated data collection through a total station, judging whether an early warning is triggered based on the data collection results, and if an early warning is triggered, taking the section to which the monitoring prism generating the early warning belongs as the judgment section, and obtaining the historical collection data of all monitoring prisms in the judgment section before the current collection;

[0009] The constructed convolutional neural network model is trained based on the obtained historical collection data, and the trained convolutional neural network model is used to predict the results of this data collection;

[0010] The deviation between the model prediction result and the actual result in the judgment section is calculated, and combined with the discrete situation of the deviation value in the spatial position, a comprehensive judgment is made on whether the warning is true and effective.

[0011] Furthermore, based on the spatial position relationship between the monitoring target and the total station, the sections to which all monitoring prisms belong are divided, specifically including:

[0012] Calculate the distance between each monitoring prism and the total station according to the total station coordinates and the monitoring prism coordinates;

[0013] Mark the monitoring prism farthest from the total station as prism A, calculate the distance from prism A to all other monitoring prisms, and divide all monitoring prisms within the preset distance range from prism A and prism A into section one; then mark the monitoring prism that is beyond the preset distance range but closest to prism A as prism B, calculate the distance from prism B to all other monitoring prisms outside section one, and divide all monitoring prisms within the preset distance range from prism B and prism B into section two; and so on, until all monitoring prisms have completed section division.

[0014] Furthermore, the total station is used to collect data automatically, and the results of this data collection are used to determine whether an early warning has been triggered, including:

[0015] Obtain the distance, horizontal angle, and zenith angle from the total station to each monitoring prism, and perform three-dimensional adjustment on the collected results with the collected data from the total station to the reference prism as the restriction condition to obtain the corrected results;

[0016] Based on the results, various change indicators of each monitoring prism in the three directions of X, Y, and Z axes of the monitoring coordinate system are calculated. According to the design plan of the monitoring project, it is determined whether each indicator exceeds the limit and whether any monitoring prism has generated an early warning.

[0017] Furthermore, the constructed convolutional neural network model is trained based on the obtained historical collection data, and the trained convolutional neural network model is used to predict the results of this data collection, specifically including:

[0018] The DenseNet convolutional neural network model is trained using historical collection data, and the trained DenseNet convolutional neural network model is used to predict the results of this data collection.

[0019] Further, calculating the deviation value between the model prediction result and the actual result in the judgment section specifically includes:

[0020] Calculate the spatial distance between the adjustment value of the acquisition result of all monitoring prisms in the judgment section and the predicted value of the convolutional neural network model, and calculate the standard deviation of the spatial distance. The standard deviation calculation process is as follows:

[0021]

[0022] Where S is the standard deviation, n is the number of monitoring prisms involved in the calculation, and x i is the spatial distance between the adjustment value of the prism i collection result and the predicted value of the convolutional neural network model, The average value of the spatial distance between the adjustment values ​​of all prism acquisition results and the predicted values ​​of the convolutional neural network model calculated this time;

[0023] If the standard deviation is lower than the preset threshold, the prediction result is valid, otherwise it is invalid.

[0024] Furthermore, combined with the discreteness of the deviation value in the spatial position, a comprehensive judgment is made on whether this warning is real and effective, including:

[0025] When the prediction results of the judgment section are valid, the spatial distance between the acquisition results of all non-warning prisms in the judgment section and the prediction results is taken as the Y value, and the spatial distance from each non-warning prism to the total station is taken as the X value to construct a two-dimensional coordinate system and calculate the linear regression equation. The average length of the vertical line from the scattered points of all non-warning prisms in the two-dimensional coordinate system to the linear regression equation is recorded as If the length of the vertical line from the scatter point of the warning prism in the two-dimensional coordinate system to the linear regression equation is less than the preset multiple If the corresponding point warning is valid, it is a false alarm warning otherwise.

[0026] Furthermore, combined with the discreteness of the deviation value in the spatial position, a comprehensive judgment is made on whether this warning is real and effective, including:

[0027] When the prediction results of the judgment section are invalid, the three-dimensional coordinates of all monitoring prisms in the judgment interval are projected two-dimensionally along the Z direction of the monitoring coordinate system, and the minimum circumscribed circle formed by the projection results of all warning prisms is calculated. If the proportion of non-warning prisms in the minimum circumscribed circle exceeds half of all prisms, the warning is invalid, otherwise it is valid.

[0028] According to the second aspect, an embodiment provides a total station automated monitoring implementation system combined with a convolutional neural network, the system comprising:

[0029] The preparation module is used to determine the deployment plan according to the actual situation of the monitoring target and deploy the monitoring prism, reference prism and total station;

[0030] The section division module is used to divide the sections to which all monitoring prisms belong based on the spatial position relationship between the monitoring target and the total station;

[0031] The monitoring module is used to perform automatic data collection through the total station, and determine whether a warning is triggered according to the data collection results. If a warning is triggered, the section to which the monitoring prism that generates the warning belongs is used as the judgment section, and the historical collection data of all monitoring prisms in the judgment section before the current collection is obtained;

[0032] The model prediction module is used to train the constructed convolutional neural network model based on the obtained historical collection data, and use the trained convolutional neural network model to predict the results of this data collection;

[0033] The judgment module is used to calculate the deviation value between the model prediction result and the actual result in the judgment section, and combine the discrete situation of the deviation value in the spatial position to comprehensively judge whether the warning is true and effective.

[0034] Furthermore, the segment division module is specifically used for:

[0035] Calculate the distance between each monitoring prism and the total station according to the total station coordinates and the monitoring prism coordinates;

[0036] Mark the monitoring prism farthest from the total station as prism A, calculate the distance from prism A to all other monitoring prisms, and divide all monitoring prisms within the preset distance range from prism A and prism A into section one; then mark the monitoring prism that is beyond the preset distance range but closest to prism A as prism B, calculate the distance from prism B to all other monitoring prisms outside section one, and divide all monitoring prisms within the preset distance range from prism B and prism B into section two; and so on, until all monitoring prisms have completed section division.

[0037] Furthermore, the model prediction module is specifically used for:

[0038] The DenseNet convolutional neural network model is trained using historical collection data, and the trained DenseNet convolutional neural network model is used to predict the results of this data collection.

[0039] The present application provides a total station automatic monitoring implementation method and system combined with convolutional neural network, determines the layout plan according to the actual situation of the monitoring target, and arranges the monitoring prism, reference prism and total station; divides the sections to which all monitoring prisms belong based on the spatial position relationship between the monitoring target and the total station; performs automatic data collection through the total station, judges whether there is a triggered warning according to the data collection result, and if the warning is triggered, the section to which the monitoring prism that generates the warning belongs is used as the judgment section, and the historical collection data of all monitoring prisms in the judgment section before the current collection is obtained; trains the constructed convolutional neural network model based on the obtained historical collection data, and uses the trained convolutional neural network model to predict the data collection result; calculates the deviation value between the model prediction result and the actual result in the judgment section, and comprehensively judges whether the warning is true and effective in combination with the discrete situation of the deviation value in the spatial position. The present invention is based on equipment and technology such as total station measurement robot and convolutional neural network. Compared with the traditional total station automatic monitoring system, it has the advantage of being able to effectively eliminate misjudged warnings; it has high engineering application value for the safe deformation monitoring of large-scale projects and their surrounding environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a method for implementing total station automated monitoring in combination with a convolutional neural network provided by an embodiment of the present invention;

[0041] Figure 2 A specific implementation flow chart of a method for realizing total station automated monitoring combined with a convolutional neural network provided by an embodiment of the present invention;

[0042] Figure 3 A schematic diagram of the logical structure of a total station automated monitoring implementation system combined with a convolutional neural network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below by specific embodiments in conjunction with the accompanying drawings. Wherein similar elements in different embodiments adopt associated similar element numbers. In the following embodiments, many detailed descriptions are for making the present application better understood. However, those skilled in the art can easily recognize that some features can be omitted in different situations, or can be replaced by other elements, materials, methods. In some cases, some operations related to the present application are not shown or described in the specification, this is to avoid the core part of the present application being overwhelmed by too much description, and for those skilled in the art, it is not necessary to describe these related operations in detail, and they can fully understand the related operations according to the description in the specification and the general technical knowledge in the art.

[0044] In addition, the features, operations or characteristics described in the specification can be combined in any appropriate manner to form various implementations. At the same time, the steps or actions in the method description can also be interchanged or adjusted in a manner that is obvious to those skilled in the art. Therefore, the various sequences in the specification and the drawings are only for the purpose of clearly describing a certain embodiment and are not meant to be a required sequence, unless otherwise specified that a certain sequence must be followed.

[0045] The first embodiment of the present invention provides a method for realizing total station automatic monitoring combined with a convolutional neural network. Figure 1 and Figure 2 Provide detailed explanation.

[0046] like Figure 1 As shown, in step S100, a deployment plan is determined according to the actual situation of the monitoring target, and monitoring prisms, reference prisms and total stations are deployed.

[0047] In this embodiment, the total station automated monitoring system is mainly composed of three parts: a total station measurement robot, a monitoring prism, and a reference prism. The monitoring prism is a prism installed at a point of unknown three-dimensional coordinates to be monitored, and the reference prism is a prism installed in a stable area at a point of known three-dimensional coordinates. The total station measurement robot periodically obtains the relative spatial position relationship between the total station and the monitoring prism and the reference prism, and these relative spatial position relationships are uniformly calculated to eliminate measurement errors, thereby obtaining the three-dimensional coordinates of the monitoring prism. Timed monitoring of deformation in the monitoring area is achieved.

[0048] The total station automated monitoring specification is used for safety monitoring in the fields of geological disasters, railways, highways, construction, etc. According to the different types of monitoring targets, combined with the corresponding specifications, actual project conditions, and historical manual measurement data, the prism distribution locations, total station installation locations, and reference prism distribution locations in the monitoring area are selected.

[0049] Use a stable foundation that meets the requirements of the specification to deploy the total station, and use the corresponding prisms to deploy the monitoring prism and the reference prism. Perform free station setting and collect the coordinates of each prism in the custom coordinate system. The monitoring prism coordinates are used as the initial values ​​of the monitoring prism of this system, and the reference prism coordinates are used as the known values ​​of the reference prism.

[0050] like Figure 1 As shown, in step S200, based on the spatial position relationship between the monitoring target and the total station, the sections to which all monitoring prisms belong are divided.

[0051] Specifically, firstly, the distance between each monitoring prism and the total station is calculated based on the total station coordinates and the monitoring prism coordinates. The distance calculation formula is as follows:

[0052]

[0053] Where S ij is the spatial distance from point i to point j. (X i ,Y i ,Z i ) is the three-dimensional coordinate of point i.

[0054] Mark the monitoring prism farthest from the total station as prism A, calculate the distance from prism A to all other monitoring prisms, and divide all monitoring prisms within one hundred meters from prism A and prism A into section one; then mark the monitoring prism that is more than one hundred meters away but closest to prism A as prism B, calculate the distance from prism B to all other monitoring prisms outside section one, and divide all monitoring prisms within one hundred meters from prism B and prism B into section two; and so on, until all monitoring prisms have completed section division.

[0055] like Figure 1 As shown, in step S300, automatic data collection is performed using a total station, and a determination is made as to whether an early warning has been triggered based on the results of this data collection. If an early warning has been triggered, the section to which the monitoring prism that generated the early warning belongs is used as the judgment section, and historical collection data of all monitoring prisms in the judgment section before this collection is obtained.

[0056] Specifically, automated data collection is performed, that is, the distance, horizontal angle, and zenith angle to each prism are obtained through a total station, and three-dimensional adjustment is performed on the collected results with the collected data from the total station to the reference prism as the restriction condition to obtain the corrected results.

[0057] Based on the results, calculate the cumulative change, single change, cumulative change rate, single change rate and other indicators of each monitoring prism in the X, Y and Z directions. According to the design plan of the monitoring project, determine whether each indicator exceeds the limit and then determine whether any monitoring prism has issued an early warning.

[0058] like Figure 1 As shown, in step S400, the constructed convolutional neural network model is trained based on the obtained historical collection data, and the trained convolutional neural network model is used to predict the results of this data collection.

[0059] Specifically, when an early warning is triggered, all prisms in the judgment section where the early warning prism is located are extracted, and the 20 historical collection data before this collection result are used to train the DenseNet convolutional neural network model. The DenseNet model can effectively alleviate the gradient vanishing problem, and has high feature utilization and fewer network parameters. It is very suitable for coordinate automatic monitoring systems with spatial relationships and possible mutations.

[0060] like Figure 1As shown, in step S500, the deviation value between the model prediction result and the actual result in the judgment section is calculated, and combined with the discrete situation of the deviation value in the spatial position, a comprehensive judgment is made as to whether the warning is true and effective.

[0061] Specifically, the spatial distance between the adjustment value of the acquisition result of all monitoring prisms in the judgment section and the predicted value of the convolutional neural network model is calculated, and the standard deviation of the spatial distance is calculated. The standard deviation calculation process is as follows:

[0062]

[0063] Where S is the standard deviation, n is the number of monitoring prisms involved in the calculation, and x i is the spatial distance between the adjustment value of the prism i collection result and the predicted value of the convolutional neural network model, The average value of the spatial distance between the adjustment values ​​of all prism acquisition results and the predicted values ​​of the convolutional neural network model calculated this time;

[0064] According to experimental experience, if the standard deviation is less than 4, the prediction result is valid, otherwise it is invalid.

[0065] When the prediction results of the judgment section are valid, the spatial distance between the acquisition results of all non-warning prisms in the judgment section and the prediction results is taken as the Y value, and the spatial distance from each non-warning prism to the total station is taken as the X value to construct a two-dimensional coordinate system and calculate the linear regression equation. The average length of the vertical line from the scattered points of all non-warning prisms in the two-dimensional coordinate system to the linear regression equation is recorded as If the length of the vertical line from the scatter point of the warning prism in the two-dimensional coordinate system to the linear regression equation is less than the preset multiple If the corresponding point warning is valid, it is a false alarm warning otherwise.

[0066] When the prediction results of the judgment section are invalid, the three-dimensional coordinates of all monitoring prisms in the judgment interval are projected two-dimensionally along the Z direction of the monitoring coordinate system, and the minimum circumscribed circle formed by the projection results of all warning prisms is calculated. If the proportion of non-warning prisms in the minimum circumscribed circle exceeds half of all prisms, the warning is invalid, otherwise it is valid.

[0067] The method for realizing total station automated monitoring combined with convolutional neural network in this embodiment aims at the deficiencies of the existing technical solutions, and continuously and in real time obtains the coordinate information of the monitored target based on the total station automated monitoring system. Different judgment sections are divided based on the spatial position relationship between the monitored target and the total station, and the model is trained using convolutional neural network based on historical measurement data to predict the current measurement results. The deviation between the predicted results and the actual results in each judgment section is compared and comprehensively analyzed and judged, so as to realize the effective filtering of the error warning of the automated monitoring system.

[0068] Corresponding to the above-disclosed method for realizing total station automatic monitoring in combination with convolutional neural network, the embodiment of the present invention further discloses a system for realizing total station automatic monitoring in combination with convolutional neural network, such as Figure 3 As shown, it specifically includes:

[0069] The preparation module is used to determine the deployment plan according to the actual situation of the monitoring target and deploy the monitoring prism, reference prism and total station;

[0070] The section division module is used to divide the sections to which all monitoring prisms belong based on the spatial position relationship between the monitoring target and the total station;

[0071] The monitoring module is used to perform automatic data collection through the total station, and determine whether a warning is triggered according to the data collection results. If a warning is triggered, the section to which the monitoring prism that generates the warning belongs is used as the judgment section, and the historical collection data of all monitoring prisms in the judgment section before the current collection is obtained;

[0072] The model prediction module is used to train the constructed convolutional neural network model based on the obtained historical collection data, and use the trained convolutional neural network model to predict the results of this data collection;

[0073] The judgment module is used to calculate the deviation value between the model prediction result and the actual result in the judgment section, and combine the discrete situation of the deviation value in the spatial position to comprehensively judge whether the warning is true and effective.

[0074] It should be noted that for the detailed description of a total station automated monitoring implementation system combined with a convolutional neural network provided in an embodiment of the present invention, reference can be made to the relevant description of a total station automated monitoring implementation method combined with a convolutional neural network provided in an embodiment of the present application, which will not be repeated here.

[0075] Those skilled in the art will appreciate that all or part of the functions of the various methods in the above-mentioned embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above-mentioned embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, and the storage medium can include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to implement the above-mentioned functions. For example, the program is stored in the memory of the device, and when the program in the memory is executed by the processor, all or part of the above-mentioned functions can be implemented. In addition, when all or part of the functions in the above-mentioned embodiments are implemented by computer programs, the program can also be stored in a storage medium such as a server, another computer, disk, optical disk, flash disk or mobile hard disk, and can be downloaded or copied and saved in the memory of the local device, or the system of the local device is updated, and when the program in the memory is executed by the processor, all or part of the functions in the above-mentioned embodiments can be implemented.

[0076] The above specific examples are used to illustrate the present invention, which is only used to help understand the present invention and is not intended to limit the present invention. For those skilled in the art, according to the concept of the present invention, some simple deductions, modifications or substitutions can be made.

Claims

1. A method for realizing total station automated monitoring combined with convolutional neural network, characterized in that: The method comprises: Determine the deployment plan based on the actual situation of the monitoring target, and deploy monitoring prisms, reference prisms and total stations; Based on the spatial position relationship between the monitoring target and the total station, the sections to which all monitoring prisms belong are divided; Performing automated data collection through a total station, judging whether an early warning is triggered based on the data collection results, and if an early warning is triggered, taking the section to which the monitoring prism generating the early warning belongs as the judgment section, and obtaining the historical collection data of all monitoring prisms in the judgment section before the current collection; The constructed convolutional neural network model is trained based on the obtained historical collection data, and the trained convolutional neural network model is used to predict the results of this data collection; The deviation between the model prediction result and the actual result in the judgment section is calculated, and combined with the discrete situation of the deviation value in the spatial position, a comprehensive judgment is made on whether the warning is true and effective.

2. The method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 1, characterized in that: Based on the spatial position relationship between the monitoring target and the total station, the sections to which all monitoring prisms belong are divided, including: Calculate the distance between each monitoring prism and the total station according to the total station coordinates and the monitoring prism coordinates; Mark the monitoring prism farthest from the total station as prism A, calculate the distance from prism A to all other monitoring prisms, and divide all monitoring prisms within the preset distance range from prism A and prism A into section one; then mark the monitoring prism that is beyond the preset distance range but closest to prism A as prism B, calculate the distance from prism B to all other monitoring prisms outside section one, and divide all monitoring prisms within the preset distance range from prism B and prism B into section two; and so on, until all monitoring prisms have completed section division.

3. The method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 1, characterized in that: The total station is used to collect data automatically, and the results of this data collection are used to determine whether an early warning has been triggered, including: Obtain the distance, horizontal angle, and zenith angle from the total station to each monitoring prism, and perform three-dimensional adjustment on the collected results with the collected data from the total station to the reference prism as the restriction condition to obtain the corrected results; Based on the results, various change indicators of each monitoring prism in the three directions of X, Y, and Z axes of the monitoring coordinate system are calculated. According to the design plan of the monitoring project, it is determined whether each indicator exceeds the limit and whether any monitoring prism has generated an early warning.

4. The method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 1, characterized in that: The constructed convolutional neural network model is trained based on the historical collection data, and the trained convolutional neural network model is used to predict the results of this data collection, including: The DenseNet convolutional neural network model is trained using historical collection data, and the trained DenseNet convolutional neural network model is used to predict the results of this data collection.

5. The method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 1, characterized in that: Calculating the deviation between the model prediction result and the actual result in the judgment section specifically includes: Calculate the spatial distance between the adjustment value of the acquisition result of all monitoring prisms in the judgment section and the predicted value of the convolutional neural network model, and calculate the standard deviation of the spatial distance. The standard deviation calculation process is as follows: Where S is the standard deviation, n is the number of monitoring prisms involved in the calculation, and x i is the spatial distance between the adjustment value of the prism i collection result and the predicted value of the convolutional neural network model, The average value of the spatial distance between the adjustment values ​​of all prism acquisition results and the predicted values ​​of the convolutional neural network model calculated this time; If the standard deviation is lower than the preset threshold, the prediction result is valid, otherwise it is invalid.

6. A method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 5, characterized in that: Combined with the discreteness of the deviation value in the spatial position, a comprehensive judgment is made on whether the warning is real and effective, including: When the prediction results of the judgment section are valid, the spatial distance between the acquisition results of all non-warning prisms in the judgment section and the prediction results is taken as the Y value, and the spatial distance from each non-warning prism to the total station is taken as the X value to construct a two-dimensional coordinate system and calculate the linear regression equation. The average length of the vertical line from the scattered points of all non-warning prisms in the two-dimensional coordinate system to the linear regression equation is recorded as If the length of the vertical line from the scatter point of the warning prism to the linear regression equation in the two-dimensional coordinate system is less than the preset multiple If the corresponding point warning is valid, it is a false alarm warning otherwise.

7. The method for realizing total station automated monitoring in combination with a convolutional neural network as claimed in claim 5, characterized in that: Combined with the discreteness of the deviation value in the spatial position, a comprehensive judgment is made on whether the warning is real and effective, including: When the prediction results of the judgment section are invalid, the three-dimensional coordinates of all monitoring prisms in the judgment interval are projected two-dimensionally along the Z direction of the monitoring coordinate system, and the minimum circumscribed circle formed by the projection results of all warning prisms is calculated. If the proportion of non-warning prisms in the minimum circumscribed circle exceeds half of all prisms, the warning is invalid, otherwise it is valid.

8. A total station automated monitoring implementation system combined with a convolutional neural network, characterized in that: The system comprises: The preparation module is used to determine the deployment plan according to the actual situation of the monitoring target and deploy the monitoring prism, reference prism and total station; The section division module is used to divide the sections to which all monitoring prisms belong based on the spatial position relationship between the monitoring target and the total station; The monitoring module is used to perform automatic data collection through the total station, and determine whether a warning is triggered according to the data collection results. If a warning is triggered, the section to which the monitoring prism that generates the warning belongs is used as the judgment section, and the historical collection data of all monitoring prisms in the judgment section before the current collection is obtained; The model prediction module is used to train the constructed convolutional neural network model based on the obtained historical collection data, and use the trained convolutional neural network model to predict the results of this data collection; The judgment module is used to calculate the deviation value between the model prediction result and the actual result in the judgment section, and combine the discrete situation of the deviation value in the spatial position to comprehensively judge whether the warning is true and effective.

9. The total station automated monitoring implementation system combined with a convolutional neural network as claimed in claim 8, characterized in that: The segment division module is specifically used for: Calculate the distance between each monitoring prism and the total station according to the total station coordinates and the monitoring prism coordinates; Mark the monitoring prism farthest from the total station as prism A, calculate the distance from prism A to all other monitoring prisms, and divide all monitoring prisms within the preset distance range from prism A and prism A into section one; then mark the monitoring prism that is beyond the preset distance range but closest to prism A as prism B, calculate the distance from prism B to all other monitoring prisms outside section one, and divide all monitoring prisms within the preset distance range from prism B and prism B into section two; and so on, until all monitoring prisms have completed section division.

10. The total station automated monitoring implementation system combined with a convolutional neural network as claimed in claim 8, characterized in that: The model prediction module is specifically used for: The DenseNet convolutional neural network model is trained using historical collection data, and the trained DenseNet convolutional neural network model is used to predict the results of this data collection.

Images