Movable intelligent observation method for accurate acquisition of phenological period of tree body

Abstract

The invention discloses a movable intelligent observation method for accurate acquisition of a tree body phenological period, relates to the technical field of mobile observation, and is used for solving the problem that the timeliness and judgment accuracy of tree body phenological period identification are reduced. The method comprises the following steps: acquiring leaf image data of each section within basic observation time, screening, marking and dividing sections based on a leaf brightness change trend and in combination with a phenological analysis period, adaptively generating observation frequency according to section heights, executing continuous image acquisition, and extracting leaf region area and contour boundary data; furthermore, phenological change evolution characteristics are constructed through conjoint analysis of leaf area missing proportion and coverage area change, phenological change prompt is triggered accordingly, layered, dynamic and traceable observation of phenological changes in different space sections of the tree body is achieved, and accuracy of phenological period identification and engineering applicability are improved.

Description

Technical Field

[0001] This invention relates to the field of mobile observation technology, and more specifically, to a mobile intelligent observation method for accurately acquiring tree phenological periods. Background Technology

[0002] Tree phenology is an important parameter reflecting plant growth rhythm, environmental adaptability and climate response characteristics. It has important application value in forestry resource monitoring, refined management of urban greening, ecological environment assessment and climate change research. Existing tree phenology monitoring methods mainly include regular manual observation, fixed-point video monitoring and macroscopic phenology identification methods based on remote sensing images.

[0003] The existing technology has the following shortcomings: Currently, existing technologies mostly adopt a unified collection strategy with fixed observation points or whole-tree scale, which cannot conduct zonal and refined observations on the phenological differences of different branch distribution directions and height sections of the tree. They also lack a continuous evolution analysis mechanism that adaptively adjusts the observation frequency based on segment characteristics and integrates multi-dimensional leaf morphological parameters. This leads to a decrease in the timeliness and accuracy of tree phenological period identification, an increase in the misjudgment rate of phenological changes, and an increase in intervention costs. Therefore, a mobile intelligent observation method for accurate acquisition of tree phenological periods is proposed.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a mobile intelligent observation method for accurately acquiring tree phenological periods. This method utilizes a segmentation strategy based on tree branch distribution, a multi-dimensional image analysis method based on leaf brightness and contour features, and an adaptive observation frequency adjustment mechanism to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a mobile intelligent observation method for accurately acquiring tree phenological stages, comprising the following steps: Step S1: Identify the branch distribution direction information of the poplar tree to be tested, divide the poplar tree to be tested into sections based on the branch distribution direction information, set the basic observation time, and collect leaf image data of each section by using a mobile observation device during the basic observation time. Step S2: Extract leaf brightness information based on leaf image data and evaluate the leaf brightness change trend. Retrieve the phenological analysis period of the poplar tree to be tested, and analyze the phenological candidate status of the segment based on the leaf brightness change trend. Step S3: Use phenological candidate states to filter and mark the division of segments, collect the segment height of the marked division segments and generate the observation frequency, and perform continuous image acquisition and detection on the marked division segments based on the observation frequency to obtain leaf area data and leaf outline boundary data. Step S4: Use the leaf outline boundary data to generate the leaf region missing ratio, and combine the leaf region missing ratio and leaf region area data to evaluate the phenological change evolution characteristics of the marked and divided segments, and determine whether to trigger the phenological change prompt based on the phenological change evolution characteristics.

[0007] In a preferred embodiment, in step S1, the initial identification of the overall tree structure of the poplar tree under test is performed by moving the observation device to obtain information on the branch distribution direction. Tree image data covering the trunk and canopy area is acquired through the image acquisition unit; Perform morphological feature extraction on the branch line regions in tree image data that present a linear or bifurcated structure; The pixel connectivity length in the branch line region is taken as the branch length. If the branch length is greater than the preset length threshold, the branch line region is taken as the target branch. The branch distribution direction information includes the spatial orientation angle of each target branch relative to the central axis of the trunk and the branch height of the target branch in the canopy.

[0008] In a preferred embodiment, in step S1, the target branches are sorted from low to high according to the numerical value of their height, and the sorting results are processed into layers according to a preset height layering interval, so that the target branches located in the same preset height layering interval are merged into the same division area. For each target branch within a defined region, directional segmentation is performed based on its corresponding branch spatial orientation angle to obtain the segmented area. The basic observation time is preset, and the tree area images of each segment are detected by the image acquisition unit; The image regions in the tree region image that are connected to the target branch and present a sheet-like structure are identified and processed, and the sheet-like structure image regions are used as leaf image data.

[0009] In a preferred embodiment, in step S2, the leaf image data is grayscaled to obtain a leaf brightness image, and the pixel brightness values ​​in the leaf brightness image are statistically analyzed to obtain leaf brightness information, which includes the average brightness value of the leaf brightness image. The average brightness values ​​corresponding to the brightness images of each leaf obtained within the basic observation time in the same segment are arranged in the order of acquisition time, and the difference between the average brightness values ​​at adjacent acquisition times is obtained to obtain the change in brightness between adjacent times. The average value of each adjacent brightness change is used to obtain the trend of blade brightness change; The phenological analysis period corresponding to the poplar variety to be tested was retrieved from the phenological parameter database.

[0010] In a preferred embodiment, in step S2, if the current time point corresponding to the basic observation time falls within the phenological analysis period, then the phenological analysis period is marked. Access the phenological parameter database to obtain the baseline interval of brightness change for the corresponding period of phenological analysis; The brightness variation trend of the divided sections is compared with the benchmark interval of brightness variation in each stage, and the brightness deviation coefficient is calculated. If the brightness deviation coefficient is less than the preset brightness deviation threshold, then the phenological candidate state of the segment is determined to be the priority candidate state. Conversely, if the phenological candidate states for the segmented area are not found to be true, they are considered as candidate states to be reviewed.

[0011] In a preferred embodiment, in step S3, when the phenological candidate state is a priority candidate state, the corresponding division segment is marked; The height of the marked sections in the vertical direction of the poplar tree to be measured is obtained by moving the observation equipment, and the section height coefficient is obtained after standardizing the section height. Access the observation parameter configuration table to retrieve the observation frequency benchmark, and calculate the observation frequency by combining the observation frequency benchmark with the section height coefficient. Based on the observation frequency, the corresponding image acquisition time interval is determined, the acquisition cycle is preset, and the marked segments are continuously acquired and detected according to the image acquisition time interval to obtain leaf image data.

[0012] In a preferred embodiment, in step S3, the leaf image data is subjected to color space conversion processing, and the color feature range of the leaf in the color space is preset, and the pixels of the image that fall into the preset color feature range are marked. The ratio of the number of marked pixels to the total number of pixels in the leaf image data is used as the leaf area data. Perform connectivity analysis on each marked pixel to obtain the leaf region, and filter boundary pixels based on the leaf region; By utilizing the spatial coordinate relationship of boundary pixels in the image, the boundary pixels are connected and sorted point by point to form a pixel sequence extending along the outer edge of the leaf region. The boundary curve formed by the pixel sequence serves as the blade profile boundary data.

[0013] In a preferred embodiment, in step S4, contour closure detection and contour filling processing are performed on the blade contour boundary data to generate the complete target area of ​​the blade. The number of pixels within the complete area of ​​the leaf target that were not identified as leaf areas was counted, and the ratio of this number of pixels to the total number of pixels within the complete area of ​​the leaf target was calculated to obtain the leaf area missing ratio. The proportion of missing parts in each leaf region was sorted in chronological order, and the average change in missing parts was calculated based on the sorting results. The area data of each leaf region are sorted in chronological order, and the average area change is calculated based on the sorting results.

[0014] In a preferred embodiment, in step S4, the average area change range and the average missing change range are standardized to obtain the average area change coefficient and the average missing change coefficient, respectively. The evolution characteristics of phenological changes were calculated by combining the average area change coefficient and the average missing change coefficient. If the evolution characteristics of phenological changes exceed the preset phenological change evolution threshold, then a phenological change prompt is triggered. Conversely, if the phenological change warning is not triggered, it will be determined that the warning will not be issued.

[0015] The technical effects and advantages of this invention are as follows: This invention identifies the distribution direction of tree branches, divides the tree under test into sections, collects leaf image data of each section within a basic observation period, filters and marks the sections based on the leaf brightness change trend and combined with the phenological analysis period, adaptively generates the observation frequency according to the section height, performs continuous image acquisition, extracts leaf area and outline boundary data, and further constructs phenological change evolution characteristics through joint analysis of the leaf area missing ratio and coverage area changes, thereby triggering phenological change prompts. This method realizes hierarchical, dynamic, and traceable observation of phenological changes in different spatial sections of the tree, improving the accuracy of phenological period identification and engineering applicability. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the implementation of a mobile intelligent observation method for accurately acquiring tree phenological periods according to the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the steps of a mobile intelligent observation method for accurately acquiring tree phenological periods according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention identifies the distribution direction of tree branches, divides the tree under test into sections, collects leaf image data of each section within a basic observation period, filters and marks the sections based on the leaf brightness change trend and combined with the phenological analysis period, adaptively generates the observation frequency according to the section height, performs continuous image acquisition, extracts leaf area and outline boundary data, and further constructs phenological change evolution characteristics through joint analysis of the leaf area missing ratio and coverage area changes, thereby triggering phenological change prompts. This method realizes hierarchical, dynamic, and traceable observation of phenological changes in different spatial sections of the tree.

[0020] Example 1, such as Figures 1 to 2 As shown, a mobile intelligent observation method for accurately acquiring tree phenological stages includes the following steps: Step S1: Identify the branch distribution direction information of the poplar tree to be tested, divide the poplar tree to be tested into sections based on the branch distribution direction information, set the basic observation time, and collect leaf image data of each section by using a mobile observation device during the basic observation time. Step S2: Extract leaf brightness information based on leaf image data and evaluate the leaf brightness change trend. Retrieve the phenological analysis period of the poplar tree to be tested, and analyze the phenological candidate status of the segment based on the leaf brightness change trend. Step S3: Use phenological candidate states to filter and mark the division of segments, collect the segment height of the marked division segments and generate the observation frequency, and perform continuous image acquisition and detection on the marked division segments based on the observation frequency to obtain leaf area data and leaf outline boundary data. Step S4: Use the leaf outline boundary data to generate the leaf region missing ratio, and combine the leaf region missing ratio and leaf region area data to evaluate the phenological change evolution characteristics of the marked and divided segments, and determine whether to trigger the phenological change prompt based on the phenological change evolution characteristics.

[0021] The specific implementation is as follows: In step S1, before starting the phenological observation of the poplar tree to be tested, the initial identification of the overall tree structure of the poplar tree to be tested is carried out by the mobile observation device to obtain the branch distribution direction information. The image acquisition unit carried by the mobile observation device collects tree image data covering the trunk and crown area as it moves segment by segment along the trunk height direction. Based on tree image data, the branch regions in the image are identified and processed. Through edge detection and line segment connectivity analysis, morphological feature extraction is performed on the branch line regions that present a linear or forked structure in the tree image data. For each branch line region, the pixel connectivity length in the branch line region is taken as the branch length. If the branch length is greater than the preset length threshold, the branch line region is taken as the target branch. The branch distribution direction information includes the spatial orientation angle of each target branch relative to the central axis of the trunk and the branch height of the target branch in the canopy; Among them, the branch spatial orientation angle is used to reflect the horizontal distribution of the target branch relative to the central axis of the trunk, and the branch height is used to reflect the position of the target branch in the vertical direction of the tree. The target branches are sorted from low to high according to their numerical height. The sorting results are then processed into layers according to the preset height layering intervals. Target branches located in the same preset height layering interval are merged into the same division area, forming multiple division areas in the vertical direction. For each target branch within a defined region, directional segmentation is performed based on its corresponding branch spatial orientation angle. Specifically, the target branches within the same division area are sorted according to their spatial orientation angle, and the target branches whose spatial orientation angle is within the preset angle division interval are grouped into the same division segment.

[0022] The preset basic observation time is used to limit the time window for leaf observation of each segment after the segment division of the poplar tree to be tested is completed; During the basic observation period, the image acquisition unit on the mobile observation equipment detects the tree region images of each segment. The image regions in the tree region images that are connected to the target branches and present a sheet-like structure are identified and processed, and the sheet-like structure image regions are used as leaf image data.

[0023] It should be noted that the mobile observation device is an intelligent observation device that can move along the height of the poplar trunk to be measured. It includes an image acquisition unit and a moving actuator for changing the device's position. The preset height stratification interval can be set according to the overall height of the poplar to be measured, the density of branch distribution, or the required observation accuracy. The preset angle segmentation interval is used to limit the orientation scale of the target branches in the horizontal direction, and can be set according to the crown spread, the dispersion of branch orientation, or the coverage of the image acquisition angle. The preset basic observation time can be set according to the phenological change cycle characteristics of the poplar to be measured.

[0024] In step S2, for each segment, the leaf image data is grayscaled to obtain a leaf brightness image. The pixel brightness values ​​in the leaf brightness image are statistically analyzed to obtain leaf brightness information, which includes the average brightness value of the leaf brightness image. The average brightness values ​​corresponding to the leaf brightness images obtained within the basic observation time for the same segment are arranged in the order of acquisition time to form the leaf brightness time series for that segment. In the leaf brightness time series, the difference between the average brightness values ​​at adjacent acquisition times is used to obtain the brightness change between adjacent times; The average value of each adjacent brightness change is used to obtain the trend of blade brightness change; The greater the trend of leaf brightness change, the greater the magnitude of the change in leaf brightness over time, and the more significant the change in leaf condition. The smaller the trend of leaf brightness change, the more stable the leaf condition. The phenological analysis period corresponding to the poplar variety to be tested is retrieved from the phenological parameter database. The phenological analysis period refers to the set of time intervals pre-configured for different phenological stages of the poplar variety to be tested within a complete observation year. Each phenological stage corresponds to a start time and an end time, which is used to limit the possible range of phenological stages in the time dimension. Match the current time point of the basic observation time with the phenological analysis period of different phenological stages. If the current time point corresponding to the basic observation time falls into the phenological analysis period, then mark the phenological analysis period. Access the phenological parameter database to obtain the benchmark interval of brightness change for the marked phenological analysis period. The benchmark interval of brightness change for the stage is a pre-set reference interval that reflects the trend of leaf brightness change within the phenological stage. The brightness variation trend of the divided sections is compared with the benchmark interval of brightness variation in each stage to analyze the matching degree of brightness trends. The width of the brightness change range is obtained by subtracting the upper and lower limits of the reference range for the phase brightness change. The deviation outside the interval is obtained by comparing the trend of leaf brightness change with the upper and lower limits of the benchmark interval for brightness change in each stage. Specifically, if the trend of leaf brightness change is greater than the upper limit of the benchmark interval for stage brightness change, the deviation outside the interval is obtained by subtracting the trend of leaf brightness change from the upper limit of the benchmark interval for stage brightness change. If the trend of leaf brightness change falls within the benchmark range of stage brightness change, then the deviation outside the range is set to 0. If the trend of leaf brightness change is less than the lower limit of the benchmark interval for stage brightness change, the deviation outside the interval is obtained by subtracting the lower limit of the benchmark interval for stage brightness change from the trend of leaf brightness change. The ratio of the deviation outside the interval to the width of the brightness change interval is used as the brightness deviation coefficient; The smaller the brightness deviation coefficient, the closer the blade brightness change trend is to the stage brightness change reference range; The brightness deviation coefficient is compared with the preset brightness deviation threshold to analyze the candidate phenological states of the divided sections; If the brightness deviation coefficient is less than the preset brightness deviation threshold, then the phenological candidate state of the segment is determined to be the priority candidate state. Conversely, if the phenological candidate states for dividing the area are determined to be candidate states to be reviewed; When the phenological candidate state is a priority candidate state, it means that the segmented area has brightness change evolution characteristics that conform to the target phenological stage at the current time point, triggering the execution of enhanced observation or continuous detection processing on the segmented area.

[0025] It should be noted that grayscale processing refers to the process of converting leaf image data from color images to single-channel brightness images; the phenological parameter database is a database used to store time interval information and corresponding brightness change benchmark intervals for different tree species at different phenological stages; the preset brightness deviation threshold can be set according to the brightness deviation distribution of the stable period of phenological stages in historical samples.

[0026] In step S3, when the phenological candidate state is a priority candidate state, the corresponding division segment is marked, and the marked division segment is further observed and scheduled. The height of the marked sections in the vertical direction of the poplar tree to be measured is obtained by moving the observation equipment. The section height is the difference between the minimum and maximum height values ​​of the target branch in the direction of the trunk height. After standardizing the segment height, a segment height coefficient is obtained. The larger the segment height coefficient, the greater the distribution span of the target branches in the vertical direction of the tree, and the higher the spatial coverage complexity.

[0027] Access the observation parameter configuration table to retrieve the observation frequency reference, and calculate the observation frequency by combining the observation frequency reference with the section height coefficient: ,in, This is the section height coefficient. To preset the observation scheduling factor, As the observation frequency reference, For observation frequency; Based on the observation frequency, the corresponding image acquisition time interval is determined, the acquisition cycle is preset, and the marked segments are continuously acquired and detected according to the image acquisition time interval to obtain leaf image data. The leaf image data is processed by color space conversion, and the color feature range of the leaf in the color space is preset. Pixels that fall within the preset color feature range are marked.

[0028] It should be explained that the standardization processing methods include, but are not limited to, standard linear transformation based on interval scaling, Z-Score standardization based on statistics, or normalization based on nonlinear mapping functions. The application methods of standardization processing will not be elaborated here. The observation parameter configuration table is a data configuration table used to store and manage the parameters related to observation scheduling. The preset observation scheduling factor can be set according to the image acquisition capability of the mobile observation equipment, the equipment movement speed, or the sensitivity of the phenological stage to the time resolution. The preset acquisition cycle can be set according to the phenological change rate and the duration of the phenological stage of the poplar tree to be tested. The color space conversion processing is the process of converting the leaf image data from the original color space to a target color space that is more conducive to distinguishing the leaves from the background. The preset color feature range is used to limit the typical color value range of the leaves in the color space, and can be set according to the statistical results of the color samples of the poplar leaves under different phenological stages.

[0029] The ratio of the number of marked pixels to the total number of pixels in the leaf image data is used as the leaf area data. The larger the value, the larger the visible coverage area of ​​the leaf within the marked segment. Perform connected component analysis on each marked pixel, merge spatially continuous marked pixels into several connected regions, and use the connected regions as leaf regions; The connected regions corresponding to the leaf image data are mapped to binary region images, where the pixels in the leaf region are denoted as foreground pixels and the pixels in the non-leaf region are denoted as background pixels. For each foreground pixel, it is detected whether a foreground pixel and a background pixel exist simultaneously within a preset neighborhood. When a foreground pixel is adjacent to at least one background pixel within its neighborhood, the foreground pixel is determined to be a boundary pixel. Based on the spatial coordinate relationship of boundary pixels in the image, the boundary pixels are connected and sorted point by point according to the spatial adjacency order between pixels to form a pixel sequence extending along the outer edge of the leaf region. The boundary curve formed by the pixel sequence serves as the blade profile boundary data.

[0030] During continuous observation, the leaf area data and leaf outline boundary data obtained at different observation times for the marked and divided sections are stored in time series to achieve traceable recording of the leaf coverage and morphological changes within the same divided section.

[0031] It should be noted that connected component analysis is a process of grouping spatially continuous labeled pixels into several connected pixel sets based on their spatial adjacency. The preset neighborhood range can be set according to the spatial resolution of the leaf image, the width of the leaf edge transition, or the required contour extraction accuracy.

[0032] In step S4, contour closure detection and contour filling processing are performed on the blade contour boundary data to generate the complete area of ​​the blade target, which represents the area that the blade should cover under the current observation view without curling, obstruction or defects, and is defined by the blade contour boundary data; The number of pixels that were not identified as leaf areas within the complete area of ​​the leaf target is counted, and the ratio of this number to the total number of pixels within the complete area of ​​the leaf target is calculated to obtain the leaf area missing ratio. The larger the value, the more obvious the decrease in the integrity of the leaf structure. The missing proportions of each leaf region are sorted in chronological order. The difference between the missing proportions of adjacent leaf regions is used to obtain the change in missing proportion. The average of the change in missing proportion is taken as the average change in missing proportion. The area data of each leaf region are sorted in chronological order. The difference between the area data of adjacent leaf regions is used to obtain the area change range. The average of the area change ranges is taken as the average area change range. After standardizing the mean area variation and mean missing data variation respectively, we obtained the mean area variation coefficient and the mean missing data variation coefficient. The evolution characteristics of phenological changes were calculated by combining the average area change coefficient and the average missing change coefficient. ,in, The average area variation coefficient, The mean missing value variation coefficient. and For preset weighting coefficients, Characteristics of phenological changes and evolution; The evolution characteristics of phenological changes are compared with preset phenological change evolution thresholds to determine whether a phenological change alert has been triggered. If the evolution characteristics of phenological changes exceed the preset phenological change evolution threshold, then a phenological change prompt is triggered. Conversely, if the phenological change warning is not triggered, it will be determined that the warning will not be issued.

[0033] The trigger phenological change prompt is used to indicate that the marked division section has entered the phenological stage evolution state, so as to trigger the adjustment of the observation scheduling of the marked division section, the labeling of phenological stages, and the recording and archiving of observation results.

[0034] It should be noted that the preset weighting coefficient can be configured according to historical phenological samples or observation sensitivity requirements; the preset phenological change evolution threshold can be set according to the typical change amplitude of different phenological stages, so as to achieve adjustable control of the triggering conditions for phenological change prompts.

[0035] This step enables a joint assessment based on changes in leaf morphology integrity and coverage, allowing phenological change judgments to be independent of single image features, thereby improving the stability and engineering reliability of phenological stage identification results.

[0036] Finally, it should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0037] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0038] In this document, the singular forms “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that terms such as “comprising / including” or “having” specify the presence of the stated features, integrals, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, integrals, steps, operations, components, parts, or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0039] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0040] The above description of the disclosed embodiments will enable those skilled in the art to make or use various modifications to these embodiments. It will be readily apparent to those skilled in the art that the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A mobile intelligent observation method for accurately acquiring tree phenological stages, characterized in that: Includes the following steps: Step S1: Identify the branch distribution direction information of the poplar tree to be tested, divide the poplar tree to be tested into sections based on the branch distribution direction information, set the basic observation time, and collect leaf image data of each section by using a mobile observation device during the basic observation time. Step S2: Extract leaf brightness information based on leaf image data and evaluate the leaf brightness change trend. Retrieve the phenological analysis period of the poplar tree to be tested, and analyze the phenological candidate status of the segment based on the leaf brightness change trend. Step S3: Use phenological candidate states to filter and mark the division of segments, collect the segment height of the marked division segments and generate the observation frequency, and perform continuous image acquisition and detection on the marked division segments based on the observation frequency to obtain leaf area data and leaf outline boundary data. Step S4: Use the leaf outline boundary data to generate the leaf region missing ratio, and combine the leaf region missing ratio and leaf region area data to evaluate the phenological change evolution characteristics of the marked and divided segments, and determine whether to trigger the phenological change prompt based on the phenological change evolution characteristics.

2. The mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 1, characterized in that: In step S1, the initial identification of the overall tree structure of the poplar tree under test is performed by moving the observation device to obtain information on the branch distribution direction; Tree image data covering the trunk and canopy area is acquired through the image acquisition unit; Perform morphological feature extraction on the branch line regions in tree image data that present a linear or bifurcated structure; The pixel connectivity length in the branch line region is taken as the branch length. If the branch length is greater than the preset length threshold, the branch line region is taken as the target branch. The branch distribution direction information includes the spatial orientation angle of each target branch relative to the central axis of the trunk and the branch height of the target branch in the canopy.

3. The mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 2, characterized in that: In step S1, the target branches are sorted from low to high according to their numerical height. The sorting results are then processed into layers according to a preset height layering interval, and the target branches located in the same preset height layering interval are merged into the same division area. For each target branch within a defined region, directional segmentation is performed based on its corresponding branch spatial orientation angle to obtain the segmented area. The basic observation time is preset, and the tree area images of each segment are detected by the image acquisition unit; The image regions in the tree region image that are connected to the target branch and present a sheet-like structure are identified and processed, and the sheet-like structure image regions are used as leaf image data.

4. The mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 3, characterized in that: In step S2, the leaf image data is converted to grayscale to obtain a leaf brightness image. The pixel brightness values ​​in the leaf brightness image are statistically analyzed to obtain leaf brightness information, which includes the average brightness value of the leaf brightness image. The average brightness values ​​corresponding to the brightness images of each leaf obtained within the basic observation time in the same segment are arranged in the order of acquisition time, and the difference between the average brightness values ​​at adjacent acquisition times is obtained to obtain the change in brightness between adjacent times. The average value of each adjacent brightness change is used to obtain the trend of blade brightness change; The phenological analysis period corresponding to the poplar variety to be tested was retrieved from the phenological parameter database.

5. A mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 4, characterized in that: In step S2, if the current time point corresponding to the basic observation time falls within the phenological analysis period, then the phenological analysis period is marked. Access the phenological parameter database to obtain the baseline interval of brightness change for the corresponding period of phenological analysis; The brightness variation trend of the divided sections is compared with the benchmark interval of brightness variation in each stage, and the brightness deviation coefficient is calculated. If the brightness deviation coefficient is less than the preset brightness deviation threshold, then the phenological candidate state of the segment is determined to be the priority candidate state. Conversely, if the phenological candidate states for the segmented area are not found to be true, they are considered as candidate states to be reviewed.

6. The mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 1, characterized in that: In step S3, when the phenological candidate state is a priority candidate state, the corresponding division segment is marked; The height of the marked sections in the vertical direction of the poplar tree to be measured is obtained by moving the observation equipment, and the section height coefficient is obtained after standardizing the section height. Access the observation parameter configuration table to retrieve the observation frequency benchmark, and calculate the observation frequency by combining the observation frequency benchmark with the section height coefficient. Based on the observation frequency, the corresponding image acquisition time interval is determined, the acquisition cycle is preset, and the marked segments are continuously acquired and detected according to the image acquisition time interval to obtain leaf image data.

7. A mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 6, characterized in that: In step S3, the leaf image data is subjected to color space conversion processing, and the color feature range of the leaf in the color space is preset. Pixels that fall within the preset color feature range are marked. The ratio of the number of marked pixels to the total number of pixels in the leaf image data is used as the leaf area data. Perform connectivity analysis on each marked pixel to obtain the leaf region, and filter boundary pixels based on the leaf region; By utilizing the spatial coordinate relationship of boundary pixels in the image, the boundary pixels are connected and sorted point by point to form a pixel sequence extending along the outer edge of the leaf region. The boundary curve formed by the pixel sequence serves as the blade profile boundary data.

8. A mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 1, characterized in that: In step S4, contour closure detection and contour filling processing are performed on the blade contour boundary data to generate the complete target area of ​​the blade. The number of pixels within the complete area of ​​the leaf target that were not identified as leaf areas was counted, and the ratio of this number of pixels to the total number of pixels within the complete area of ​​the leaf target was calculated to obtain the leaf area missing ratio. The proportion of missing parts in each leaf region was sorted in chronological order, and the average change in missing parts was calculated based on the sorting results. The area data of each leaf region are sorted in chronological order, and the average area change is calculated based on the sorting results.

9. A mobile intelligent observation method for accurately acquiring tree phenological periods according to claim 8, characterized in that: In step S4, the average area change rate and the average missing value change rate are standardized to obtain the average area change coefficient and the average missing value change coefficient. The evolution characteristics of phenological changes were calculated by combining the average area change coefficient and the average missing change coefficient. If the evolution characteristics of phenological changes exceed the preset phenological change evolution threshold, then a phenological change prompt is triggered. Conversely, if the phenological change warning is not triggered, it will be determined that the warning will not be issued.

Images