Method and system for collecting radial growth data of tree

Abstract

The invention discloses a tree radial growth data acquisition method and system, and belongs to the technical field of intelligent monitoring. The method comprises the following steps: acquiring first radial growth data through a measuring instrument, and then collecting historical remote sensing and environmental monitoring data and corresponding radial growth data to construct a growth data set; then, a BP neural network model fused with a MapReduce algorithm is adopted, the model is trained through parallelization preprocessing and operation of a Map stage and a Reduce stage, and an optimization analysis model is obtained; inputting remote sensing and environment monitoring data acquired in real time into the model to obtain second radial growth data; and finally, calculating according to a given growth anomaly degree formula and correcting the radial growth data of the tree according to a result. The comprehensive utilization of the multi-source data enables the data to be more comprehensive and accurate, the advanced data analysis model can accurately mine the data relationship, and the calculation of the growth abnormality and the data correction mechanism can guarantee the reliability of the radial growth data of the tree.

Description

Technical Field

[0001] The present invention relates to the field of intelligent monitoring technology, and more particularly to a method and system for collecting radial growth data of trees. Background Art

[0002] As the main body of the terrestrial ecosystem, real-time monitoring of tree growth data is of great significance to the development and maintenance of my country's forestry ecology. Among them, the deformation measurement of tree radial growth is an important indicator for monitoring the tree growth process.

[0003] Traditional tree measurement devices mostly use manual measurement and recording, and record the measurement results in a computer. This survey method is relatively backward, and it takes a lot of manpower to conduct a measurement, and the work efficiency is not high.

[0004] Therefore, how to propose a method and system for collecting tree radial growth data to achieve efficient and accurate tree radial growth data monitoring is an urgent problem that technical personnel in this field need to solve. Summary of the invention

[0005] In view of this, the present invention provides a method and system for collecting radial growth data of trees, so as to improve the accuracy and reliability of radial data monitoring of tree growth.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] In one aspect, the present invention provides a method for collecting radial growth data of trees, comprising the following steps:

[0008] Acquiring first radial growth data based on a measuring instrument;

[0009] Obtain historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set;

[0010] Constructing a data analysis model, and using the growth data set to train the data analysis model to obtain an optimized analysis model;

[0011] Inputting the remote sensing data and environmental monitoring data collected in real time into the optimization analysis model, and outputting the second radial growth data;

[0012] The growth abnormality is calculated according to the first radial growth data and the second radial growth data, and the tree radial growth data is corrected according to the growth abnormality.

[0013] Preferably, the data analysis model is a BP neural network model integrated with the MapReduce algorithm.

[0014] Preferably, using the growth data set to train the data analysis model to obtain an optimized analysis model includes:

[0015] The BP neural network is preprocessed in parallel; the preprocessing in parallel includes setting the number of samples contained in each row of data in the growing data set read by MapReduce according to the batch gradient;

[0016] In the Map stage, the samples contained in each row are split and formed into a new array, and then the BP neural network is trained using the mini-batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity;

[0017] In the Reduce phase, the output weights of the Map phase are taken as input, and their arithmetic mean is calculated as output. The Map and Reduce phases are repeated until the maximum number of training times is reached or the specified accuracy is achieved.

[0018] Preferably, the samples contained in each row are split and formed into a new array, including:

[0019] Use the split function to split the sample into a list of multiple lines;

[0020] Use the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

[0021] Preferably, the calculation formula of the growth abnormality degree E is as follows:

[0022]

[0023] Wherein, α and β are weight coefficients, m1 is the first radial growth data, and m2 is the second radial growth data.

[0024] Preferably, correcting the tree radial growth data according to the growth abnormality comprises:

[0025] Selecting the first radial growth data or the second radial growth data as the growth data m to be corrected according to the weight coefficient;

[0026] Calculate the correction value r based on the growth abnormality E;

[0027] The growth data to be corrected m is corrected according to the correction value r to obtain the final tree radial growth data.

[0028] On the other hand, the present invention further provides a system for collecting radial growth data of trees, which is used to implement the above-mentioned method for collecting radial growth data of trees, comprising:

[0029] A data acquisition module, for acquiring first radial growth data based on a measuring instrument;

[0030] The data set construction module obtains historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set;

[0031] A model training module constructs a data analysis model and uses the growth data set to train the data analysis model to obtain an optimized analysis model;

[0032] A data output module, inputting the real-time collected remote sensing data and environmental monitoring data into the optimization analysis model, and outputting the second radial growth data;

[0033] The growth data output module calculates the growth abnormality according to the first radial growth data and the second radial growth data, and corrects the radial growth data of the tree according to the growth abnormality.

[0034] Preferably, the data analysis model is a BP neural network model integrated with the MapReduce algorithm.

[0035] Preferably, the model training module includes:

[0036] A preprocessing unit performs parallel preprocessing on the BP neural network; the parallel preprocessing includes setting the number of samples contained in each row of data in the growing data set read by MapReduce according to the batch gradient;

[0037] Map output unit, in the Map stage, the samples contained in each row are split and formed into a new array, and then the BP neural network is trained using the mini-batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity;

[0038] Reduce output unit, in the Reduce phase, takes the output weights of the Map phase as input and calculates their arithmetic mean as output, repeating the Map and Reduce phases until the maximum number of training times is reached or the specified accuracy is achieved.

[0039] Preferably, the pre-processing unit comprises:

[0040] The sample segmentation subunit uses the split function to split the sample into a list of multiple lines;

[0041] The array conversion subunit uses the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

[0042] Through the above technical solutions, it can be known that compared with the prior art, the present invention discloses a method and system for collecting radial growth data of trees, which obtains first radial growth data through a measuring instrument, obtains second radial growth data using a BP neural network that integrates a MapReduce algorithm, and then calculates the growth anomaly through the first radial growth data and the second radial growth data to correct the data. The present invention integrates multi-source data, including measurement, remote sensing and environmental data, to make the data more comprehensive and accurate; adopts advanced data analysis models to effectively mine the relationship between data and improve prediction accuracy; the calculation of growth anomaly and data correction mechanism can timely correct possible measurement deviations and ensure the reliability of the final tree radial growth data. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0044] Figure 1 A flow chart of the method provided by the present invention;

[0045] Figure 2 This is a system architecture diagram provided by the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] On the one hand, the embodiment of the present invention provides a method for collecting radial growth data of trees, such as Figure 1 As shown, the following steps are included:

[0048] S1. Acquire first radial growth data based on a measuring instrument.

[0049] Select a suitable tree radial growth measuring instrument, such as a high-precision electronic diameter measuring instrument. Select a disease-free, relatively flat location on the tree trunk as the measuring point and mark it. Install the measuring instrument at the measuring point according to the operating specifications, start the measuring instrument to obtain the radial growth data at the initial moment, and record the measurement time, tree number, measurement location and other related information.

[0050] S2. Obtain historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set.

[0051] Obtain historical remote sensing images of the corresponding time period in the study area from satellite remote sensing data providers or relevant institutions, including multispectral data, and extract features related to tree growth, such as the Normalized Difference Vegetation Index (NDVI).

[0052] Obtain historical environmental data from local meteorological stations, environmental monitoring stations, etc., including temperature, precipitation, soil moisture, light intensity, etc., to ensure that the time series of these data correspond to the tree radial growth data and remote sensing data.

[0053] The radial growth data of the same or similar trees in the corresponding time period were obtained from the historical archives of the forestry department and previous research materials.

[0054] The historical remote sensing data, historical environmental monitoring data and corresponding radial growth data collected above were organized in a unified format, and each piece of data was labeled with information such as time, location, and tree number to form a growth data set.

[0055] S3. Build a data analysis model, use the growth data set to train the data analysis model, and obtain an optimized analysis model.

[0056] The data analysis model is a BP neural network model integrated with the MapReduce algorithm.

[0057] The BP neural network is a three-layer neural network, including an input layer, a hidden layer and an output layer. The number of remote sensing data and environmental monitoring data represents the number of input layer nodes of the neural network, and the output layer of the neural network is the radial growth data of trees.

[0058] The data analysis model is trained using the growth data set to obtain an optimized analysis model, including:

[0059] S31. Parallel preprocessing of the BP neural network; parallel preprocessing includes setting the number of samples contained in each row of data in MapReduce reading the growth data set according to the batch gradient;

[0060] S32. In the Map stage, the samples contained in each row are divided and formed into a new array, and then the BP neural network is trained using a small batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity;

[0061] The samples contained in each row are split and formed into a new array, including:

[0062] S321. Use the split function to split the sample into a list of multiple lines;

[0063] S322. Use the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

[0064] S33. In the Reduce stage, the output weights of the Map stage are taken as input, and their arithmetic mean is calculated as output. The Map and Reduce stages are repeated until the maximum number of training times is reached or the specified accuracy is reached.

[0065] This embodiment also includes using time as the horizontal axis and the predicted tree radial growth data as the vertical axis to perform curve fitting, compare the predicted growth curve with the actual growth curve, and adjust the batch gradient according to the comparison result.

[0066] S4. Input the remote sensing data and environmental monitoring data collected in real time into the optimization analysis model, and output the second radial growth data.

[0067] S5. Calculate the growth abnormality according to the first radial growth data and the second radial growth data, and correct the tree radial growth data according to the growth abnormality.

[0068] The calculation formula of growth abnormality E is as follows:

[0069]

[0070] Wherein, α and β are weight coefficients, m1 is the first radial growth data, and m2 is the second radial growth data.

[0071] Correction of tree radial growth data based on growth anomaly, including:

[0072] S51. Select the first radial growth data or the second radial growth data as the growth data m to be corrected according to the weight coefficient.

[0073] Specifically, when α>β, the first radial growth data m1 is selected as the growth data m to be corrected, that is, m=m1.

[0074] When α<β, the second radial growth data m2 is selected as the growth data m to be corrected, that is, m=m2.

[0075] When α=β, the selection is made based on random selection or based on current environmental conditions (such as whether the recent weather is stable, etc.).

[0076] S52. Determine a correction level based on the growth abnormality, and obtain a correction value r according to the correction level.

[0077] When 0≤E<0.1, the correction level is defined as 1, indicating that the growth abnormality is very low, and r=0 at this time.

[0078] When 0.1≤E<0.3, the correction level is defined as 2, indicating that the growth is abnormal to a certain extent. At this time, r=k1·E, and k1 is a coefficient determined based on experience.

[0079] When E≥0.3, the correction level is defined as 3, indicating that the growth abnormality is more significant, and at this time r=k2·E+1) n -1]. Where k2 and n are coefficients determined based on experience.

[0080] S53. The growth data m to be corrected is corrected according to the correction value r to obtain the final tree radial growth data M as follows:

[0081] M=m+r.

[0082] On the other hand, the present invention also provides a system for collecting radial growth data of trees, such as Figure 2 As shown, the system is used to implement the above-mentioned method for collecting tree radial growth data, including:

[0083] A data acquisition module, for acquiring first radial growth data based on a measuring instrument;

[0084] The data set construction module obtains historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set;

[0085] Model training module, builds data analysis model, uses growth data set to train data analysis model, and obtains optimized analysis model;

[0086] The data output module inputs the real-time collected remote sensing data and environmental monitoring data into the optimization analysis model and outputs the second radial growth data;

[0087] The growth data output module calculates the growth abnormality according to the first radial growth data and the second radial growth data, and corrects the radial growth data of the tree according to the growth abnormality.

[0088] Preferably, the data analysis model is a BP neural network model integrated with the MapReduce algorithm.

[0089] Preferably, the model training module includes:

[0090] The preprocessing unit performs parallel preprocessing on the BP neural network; the parallel preprocessing includes setting the number of samples contained in each row of data in the growing data set read by MapReduce according to the batch gradient;

[0091] Map output unit, in the Map stage, the samples contained in each row are split and formed into a new array, and then the BP neural network is trained using the mini-batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity;

[0092] Reduce output unit, in the Reduce phase, takes the output weights of the Map phase as input and calculates their arithmetic mean as output, repeating the Map and Reduce phases until the maximum number of training times is reached or the specified accuracy is achieved.

[0093] Preferably, the pre-processing unit comprises:

[0094] The sample segmentation subunit uses the split function to split the sample into a list of multiple lines;

[0095] The array conversion subunit uses the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

[0096] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0097] 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 method for collecting tree radial growth data, characterized in that: The following steps are involved: Acquiring first radial growth data based on a measuring instrument; Obtain historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set; Constructing a data analysis model, and using the growth data set to train the data analysis model to obtain an optimized analysis model; Inputting the remote sensing data and environmental monitoring data collected in real time into the optimization analysis model, and outputting the second radial growth data; The growth abnormality is calculated according to the first radial growth data and the second radial growth data, and the tree radial growth data is corrected according to the growth abnormality.

2. The method for collecting radial growth data of trees according to claim 1, characterized in that: The data analysis model is a BP neural network model integrated with the MapReduce algorithm.

3. The method for collecting radial growth data of trees according to claim 2, characterized in that: Using the growth data set to train the data analysis model to obtain an optimized analysis model includes: The BP neural network is preprocessed in parallel; the preprocessing in parallel includes setting the number of samples contained in each row of data in the growing data set read by MapReduce according to the batch gradient; In the Map stage, the samples contained in each row are split and formed into a new array, and then the BP neural network is trained using the mini-batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity; In the Reduce phase, the output weights of the Map phase are taken as input, and their arithmetic mean is calculated as output. The Map and Reduce phases are repeated until the maximum number of training times is reached or the specified accuracy is achieved.

4. The method for collecting radial growth data of trees according to claim 3, characterized in that: Split the samples contained in each row into a new array, including: Use the split function to split the sample into a list of multiple lines; Use the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

5. The method for collecting radial growth data of trees according to claim 1, characterized in that: The calculation formula of the growth abnormality degree E is as follows: Wherein, α and β are weight coefficients, m1 is the first radial growth data, and m2 is the second radial growth data.

6. The method for collecting radial growth data of trees according to claim 5, characterized in that: Correcting the tree radial growth data according to the growth anomaly includes: Selecting the first radial growth data or the second radial growth data as the growth data m to be corrected according to the weight coefficient; Calculate the correction value r based on the growth abnormality E; The growth data to be corrected m is corrected according to the correction value r to obtain the final tree radial growth data.

7. A system for collecting radial growth data of trees, characterized in that: include: A data acquisition module, for acquiring first radial growth data based on a measuring instrument; The data set construction module obtains historical remote sensing data, historical environmental monitoring data and corresponding radial growth data to construct a growth data set; A model training module constructs a data analysis model and uses the growth data set to train the data analysis model to obtain an optimized analysis model; A data output module, inputting the real-time collected remote sensing data and environmental monitoring data into the optimization analysis model, and outputting the second radial growth data; The growth data output module calculates the growth abnormality according to the first radial growth data and the second radial growth data, and corrects the radial growth data of the tree according to the growth abnormality.

8. A system for collecting radial growth data of trees according to claim 7, characterized in that: The data analysis model is a BP neural network model integrated with the MapReduce algorithm.

9. A system for collecting radial growth data of trees according to claim 8, characterized in that: The model training module includes: A preprocessing unit performs parallel preprocessing on the BP neural network; the parallel preprocessing includes setting the number of samples contained in each row of data in the growing data set read by MapReduce according to the batch gradient; Map output unit, in the Map stage, the samples contained in each row are split and formed into a new array, and then the BP neural network is trained using the mini-batch gradient descent algorithm, and the average weight is updated and output according to the batch processing quantity; Reduce output unit, in the Reduce phase, takes the output weights of the Map phase as input and calculates their arithmetic mean as output, repeating the Map and Reduce phases until the maximum number of training times is reached or the specified accuracy is achieved.

10. The system for collecting radial growth data of trees according to claim 9, characterized in that: The pre-processing unit comprises: The sample segmentation subunit uses the split function to split the sample into a list of multiple lines; The array conversion subunit uses the numpy.array function to convert the list into an array form; the number of array rows is the batch gradient, and the number of columns is the number of independent variables of the data.

Images