An Automatic Method for Calculating Vegetation Coverage in Municipal Greenbelts Based on Image Extraction

Abstract

This invention discloses an automatic method for calculating vegetation coverage in municipal green belts based on image extraction, belonging to the field of intelligent environmental assessment technology. The method includes the following steps: establishing a rainfall time recording sequence during the automatic vegetation coverage calculation process; synchronously recording ambient light intensity changes during image acquisition; and marking high reflectivity risk states after rain at corresponding time nodes in the rainfall time recording sequence. This invention achieves dynamic identification and hierarchical management of post-rain reflectivity risk by establishing a rainfall time recording sequence and combining it with light intensity change information. Through the constraint of a reliable vegetation reference area and the adjustment of brightness rhythm lines, reflective areas are gradually returned to the normal recognition range, avoiding misjudgments and sudden drops in coverage after rain, ensuring that the calculation results are continuous, stable, and accurately reflect the greening status.

Description

Technical Field

[0001] This invention relates to the field of intelligent environmental assessment technology, specifically to an automatic method for calculating the vegetation coverage of municipal green belts based on image extraction. Background Technology

[0002] Image extraction refers to using images captured in the field or acquired through remote sensing as data sources. With the support of computer vision technology, it analyzes the pixel information in the images to extract effective features that can characterize the state of vegetation, such as color distribution, brightness differences, texture continuity, and boundary morphology, thereby distinguishing green vegetation areas from non-vegetated areas such as roads, bare soil, and buildings. Automatic calculation of vegetation coverage in municipal green belts refers to, after completing the identification and extraction of vegetation areas, relying on the spatial information analysis capabilities of computer vision, according to a predetermined spatial range or management unit, statistically analyzing the area of ​​identified vegetation pixels, and combining image scale or spatial calibration relationships to automatically calculate the proportion of vegetation area to the overall area of ​​the green belt. This objectively reflects the vegetation coverage level of municipal green belts, enabling quantitative assessment and dynamic monitoring of the current state of greening, and reducing subjective errors and workload caused by manual visual inspection or on-site measurement.

[0003] The existing technology has the following shortcomings: In existing technologies, automatic calculation of vegetation coverage in municipal green belts based on image extraction typically relies on the color distribution characteristics, brightness range characteristics, and texture continuity characteristics of vegetation areas within the image. However, shortly after rainfall, the leaves of green belt vegetation are generally covered with a film of water. This film easily produces significant specular reflection under natural light or ambient lighting, causing some vegetation areas to exhibit abnormally high brightness responses at the moment of image acquisition. Because existing technologies lack effective mechanisms for identifying and distinguishing the reflective state of leaves after rainfall, the image extraction stage often misidentifies these highly reflective areas as road reflections, water surface reflections, or other non-vegetation highly reflective objects, thus incorrectly removing actual vegetation pixels. This misjudgment usually occurs over large, continuous areas, leading to a sudden and significant drop in single coverage calculation results. This causes the system to quickly arrive at a coverage change conclusion that is clearly inconsistent with the actual greening situation, and further misjudged as vegetation degradation in municipal greening management applications, triggering unnecessary abnormal warnings or management responses.

[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] The purpose of this invention is to provide an automatic method for calculating the vegetation coverage of municipal green belts based on image extraction, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an automatic calculation method for vegetation coverage rate of municipal green belts based on image extraction, comprising the following steps: During the automatic calculation of vegetation coverage, a rainfall time record sequence is established, and the ambient light intensity change information is recorded synchronously during image acquisition. The high reflectivity risk status after rain is marked at the corresponding time nodes of the rainfall time record sequence. Based on the high reflectivity risk after rain, image data for the corresponding time period is retrieved, the duration of abnormal brightness increase in the image is statistically analyzed, and continuous image regions with reflective characteristics are extracted to form the range of reflective influence area. Based on the re-comparison of the original vegetation identification results with the range of the reflected light impact area, the spatial continuity change of vegetation within the reflected light impact area is analyzed, the vegetation areas that were incorrectly excluded are marked, and a reliable vegetation reference area is generated. Using a reliable vegetation reference area as a constraint, the brightness judgment range used in the vegetation identification process is adjusted so that the reflective area gradually returns to the control of the reliable vegetation reference area in the time dimension, forming a brightness adjustment rhythm line. The brightness impact on the reflective area is adjusted in segments according to the brightness adjustment rhythm line. By reducing the discrimination weight of the reflective area in a short period of time and gradually restoring the vegetation display characteristics, the continuous and stable calculation of vegetation coverage after rain is completed.

[0007] Preferably, the steps for marking a high reflectivity risk state after rain are as follows: During the automatic calculation of vegetation coverage, synchronous records of time information and environmental status information are obtained. A time series is established using the timestamp information of the image acquisition device, and the time series is correlated with the rainfall monitoring records on the time axis. After completing the synchronized recording of time series and illumination information, a rainfall time record sequence is established based on the start time, end time and duration of the rainfall event, and the sequence includes illumination intensity, air humidity and image acquisition status information; After the rainfall time record sequence is established, the information on changes in light intensity is continuously analyzed, and the areas of sudden changes in light intensity are compared with the rainfall time record sequence. The high reflectivity risk status after the rain ends is marked shortly after the rain ends. After marking the high reflectivity risk status after rain, the rainfall time record sequence and light intensity change information are integrated to form a time index dataset, which provides time dimension input for vegetation identification and supports subsequent brightness adjustment.

[0008] Preferably, the formation process of the reflective influence area is as follows: After establishing the post-rain high reflectivity risk status, based on the risk time marker information in the rainfall time record sequence, image data for the corresponding time period is retrieved from the image acquisition dataset, and the light intensity record value is associated with and saved with environmental parameters; After retrieving the image data for the corresponding time period, the brightness change amplitude of adjacent image frames on the time axis is continuously statistically analyzed to determine the duration range of abnormal brightness increase and establish the abnormal brightness duration range. After obtaining the duration of the brightness anomaly, the distribution area of ​​bright pixels is analyzed based on the image frame set. The spatial continuity of the bright area in adjacent time frames is judged, and continuous image areas with reflective features are extracted. After extracting continuous image regions, spatially fuse partially overlapping regions within the same time period to form a reflective influence area range that maintains a correspondence in both time and space dimensions.

[0009] Preferably, during the formation of the reflective influence area, the boundary information of each reflective candidate area is correlated with the duration of brightness anomaly, so that each sub-region within the reflective influence area has a clear brightness duration attribute, thereby realizing the spatial quantification of the degree of reflective interference in the post-rain environment, and providing a time and space joint constraint basis for subsequent vegetation identification and brightness adjustment.

[0010] The preferred steps for generating a reliable vegetation reference area are as follows: After the area affected by reflected light is defined, the original vegetation identification results are extracted and located again. The areas marked as non-vegetation are spatially superimposed with the area affected by reflected light, and combined with time information to form a re-identification dataset. After obtaining the re-identification dataset, the changes in the spatial continuity of vegetation within the area affected by reflection were analyzed. By comparing the differences in vegetation morphology before rainfall, during reflection, and after rainfall, the spatial breaks and regional gaps caused by reflection were identified. After identifying potential areas for false exclusion, these areas are marked and classified, and time nodes, reflectivity information and vegetation boundary data are bound and fused to generate connected regions. After marking is completed, the correctly identified vegetation areas and the incorrectly excluded areas are merged, and time layers are performed according to the duration of reflection to generate a reliable vegetation reference area with continuity in space and time.

[0011] Preferably, when generating a reliable vegetation reference area, the temporal attributes and spatial boundary information of the reflected light influence area are synchronously mapped to the merged vegetation area, and the temporal level of the reference area is adjusted according to the order of the duration of the reflected light, so that the vegetation reference area remains continuous in spatial distribution and reflects the gradual changes of vegetation recovery in time series, thereby enhancing the temporal stability of the vegetation coverage measurement results.

[0012] Preferably, the steps for generating the brightness adjustment rhythm line are as follows: After generating a reliable vegetation reference region, the spatial extent and temporal attributes of the reliable vegetation reference region are synchronously mapped to the original image sequence, so that the overlapping parts are given priority in the recognition process, and a brightness adjustment reference frame is established. After establishing a brightness adjustment reference framework, the brightness distribution characteristics in the reflective area are analyzed over time and compared with the average brightness and illumination recovery trend of the reference area to dynamically adjust the brightness judgment range. After adjusting the brightness judgment range, the brightness change tracking of continuous time frames is used to gradually bring the reflective area back to the control of the reliable vegetation reference area in the time dimension, and to achieve fusion with the reference area in the spatial dimension. After the reflection-affected area is controlled, the brightness adjustment range and recovery rate at each time point are integrated with the brightness benchmark of the reference area to form a brightness adjustment rhythm line describing the brightness change pattern.

[0013] Preferably, during the formation of the brightness adjustment rhythm line, the brightness recovery rate of the reflective area in the time dimension is correlated with the brightness stability trend of the reliable vegetation reference area, so that the brightness adjustment rhythm line reflects both the brightness change amplitude and the spatial regression state, thereby achieving continuous constraint and dynamic control of the brightness of the reflective area in the subsequent identification process.

[0014] Preferably, the steps for segmenting and adjusting the reflective area according to the brightness adjustment rhythm line, reducing the discrimination weight of the reflective area, and gradually restoring the vegetation display characteristics are as follows: After establishing the brightness adjustment rhythm line, the brightness change trend in the area affected by reflection is segmented over time, and a time frame for brightness adjustment is established based on the brightness recovery speed and the duration of reflection. After establishing the time frame, the discrimination weight of the reflective influence area is systematically reduced according to the brightness characteristics of each stage in the brightness adjustment rhythm line, while maintaining the spatial correspondence with the credible vegetation reference area. After the weight of the reflective area is reduced, the vegetation display characteristics of the reflective area are gradually restored according to the brightness adjustment rhythm line, so that the reflective area can regain the ability to distinguish normal vegetation area. After completing the segmented adjustment and weight restoration, the brightness adjustment results in each time segment are integrated to generate a time-continuous coverage calculation sequence, and a continuous and stable measurement result of the post-rain vegetation coverage is formed.

[0015] Preferably, when gradually restoring the vegetation display characteristics of the reflective area according to the brightness adjustment rhythm line, the brightness restoration rate is synchronously compared with the brightness distribution of the reference vegetation area, so that the brightness regression process of the reflective area is consistent with the change of ambient light in time, thereby ensuring the temporal continuity and spatial stability of the coverage calculation.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention establishes a rainfall time record sequence during image acquisition and combines it with information on changes in ambient light intensity. It marks high-reflection-risk states after rain, achieving proactive identification and time-layered management of reflective effects. This allows the image recognition process to adaptively adjust the recognition range for different climatic conditions. This method effectively avoids the problem of false exclusion of vegetation areas due to water film reflection, ensuring that the calculated results remain stable and consistent with the actual greening status for a short period after rain. This fundamentally improves the reliability and continuity of vegetation coverage calculation under changing environmental conditions.

[0017] This invention introduces a reliable vegetation reference area as a dynamic constraint during brightness adjustment, allowing the reflective area to gradually return to its normal recognition range over time, and achieving segmented adjustment based on the brightness adjustment rhythm. Through this continuous control mechanism, vegetation display characteristics can gradually return to their natural state during the light recovery phase, resulting in a smooth transition in the measurement results over time. This avoids sudden drops in coverage caused by short-term reflections, ensuring stable output and accurate reflection of greening monitoring results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a flowchart of the method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction, as described in this invention. Detailed Implementation

[0020] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0021] This invention provides, for example Figure 1 The automatic calculation method for vegetation coverage of municipal green belts based on image extraction, as shown, includes the following steps: During the automatic calculation of vegetation coverage, a rainfall time record sequence is established, and the ambient light intensity change information is recorded synchronously during image acquisition. The high reflectivity risk status after rain is marked at the corresponding time nodes of the rainfall time record sequence. In the automatic vegetation cover calculation process, to accurately identify the impact of post-rainfall environmental conditions on the brightness characteristics of vegetation images during the image extraction stage, it is necessary to establish a time-series recording that reflects the rainfall process and its relationship with changes in illumination conditions. This is so that high-reflectance-risk states after rain can be identified and marked in subsequent image data processing, thus providing a fundamental support for the continuity and accuracy of the vegetation cover calculation results. The specific steps are as follows: The automatic vegetation cover calculation process involves synchronously recording time and environmental status information. Specifically, at the start of each image acquisition, a basic time series is established using the timestamp information of the image acquisition device, and this time series is correlated with local rainfall monitoring records along the timeline. In this way, all image data can be linked to the rainfall process in the time dimension, allowing each frame to be clearly located to a specific time period before, during, or after rainfall. Simultaneously, ambient light intensity changes are recorded synchronously during each image acquisition. This light information includes natural light levels, shadow distribution ratios, and trends in light source angle changes, characterizing the overall state of the external lighting environment at the time of acquisition. By synchronously recording light intensity changes with the time series, brightness anomalies caused by sudden changes in light intensity can be identified in subsequent steps, thus providing basic data for identifying post-rain reflection risks.

[0022] After completing the synchronized recording of the time sequence and illumination information, a complete rainfall time sequence is established based on the start time, end time, and duration of rainfall events in the time sequence. This sequence includes not only the period of rainfall but also a buffer period after the rainfall ends, reflecting the post-rain environmental light recovery and changes in leaf condition. Specifically, when establishing the rainfall time sequence, the time sequence is divided at fixed intervals, and each time node includes information such as rainfall status, light intensity, air humidity, and image acquisition status at that moment. This hierarchical recording method forms a continuous timeline data structure to represent the dynamic changes in rainfall and post-rain conditions. This rainfall time sequence will serve as a reference for subsequent labeling of high reflectivity risk states after rain, providing a temporal input for identifying the reflectivity risk level of image data at different time periods.

[0023] After establishing the rainfall time-series, continuous analysis is performed on the light intensity changes during image acquisition. Abrupt light intensity regions within consecutive time periods are compared with the rainfall time-series to determine the rate of change in light levels and the brightness recovery trend shortly after rainfall. This cross-comparison of time and light information identifies the time window where environmental light conditions are unstable after rainfall. Within this time window, incompletely evaporated water films may still exist on the surface of vegetation leaves, leading to strong specular reflection. Therefore, markers need to be added to corresponding time nodes in the rainfall time-series to indicate high reflectivity risks during these periods. The marker establishment process must be synchronized with the light intensity change trend. When the light intensity increases at a rate exceeding a preset threshold within a short period, the system automatically adds a high reflectivity risk marker to that time node, thus transforming the ordinary rainfall record sequence into an extended time series with reflectivity risk identification capabilities. In this way, rainfall and light intensity change information can be integrated into a unified temporal dimension system, providing a risk perception basis for subsequent image extraction and recognition steps.

[0024] After marking the post-rain high reflectivity risk state, the rainfall time record sequence and light intensity change information are integrated to form a time-indexed dataset that can be directly used for vegetation identification. This dataset not only contains rainfall status information and light intensity parameters for each time point, but also includes corresponding reflectivity risk level labels to guide the subsequent retrieval and selection of image data. During the automatic vegetation coverage calculation process, when the system needs to analyze image data for a certain period, it can first retrieve the corresponding time point through the time-indexed dataset and determine the image data processing strategy based on its reflectivity risk level. If the current time point is marked as a post-rain high reflectivity risk state, the system will apply targeted brightness weight adjustments and regional feature constraints to the images of that period in subsequent image recognition steps to reduce misjudgments caused by water film reflection. This process not only ensures that the image recognition stage can dynamically adapt to environmental conditions based on rainfall history and light intensity change information, but also makes the entire calculation process self-interpretable and continuous in the time dimension. In this way, the establishment of rainfall time record sequences and the marking of high reflectivity risk status after rain achieve a complete closed loop from time information collection to light change analysis and risk status identification. This enables the automatic vegetation coverage calculation process to adaptively adjust to the environmental conditions after rainfall, fundamentally reducing the phenomenon of erroneous vegetation removal caused by post-rain reflection and improving the stability and reliability of the calculation results.

[0025] Based on the high reflectivity risk after rain, image data for the corresponding time period is retrieved, the duration of abnormal brightness increase in the image is statistically analyzed, and continuous image regions with reflective characteristics are extracted to form the range of reflective influence area. To accurately address the abnormal brightness of vegetation areas caused by changes in environmental conditions after rain, targeted analysis of image data for the corresponding time period is necessary under high reflectivity risk conditions after rain. By statistically analyzing the duration of brightness changes and extracting the continuity of spatial features, continuous image regions with reflective characteristics are identified, and the extent of the reflective influence area is defined. This provides a spatial constraint for subsequent vegetation identification and brightness adjustment. The specific steps are as follows: After establishing the post-rain high reflectivity risk status, image data for the corresponding time period is retrieved from the image acquisition dataset based on the risk time marker information in the rainfall time record sequence. To ensure that the retrieved image data is consistent with the environmental conditions, precise matching is required based on the time node of each marker in the time record sequence, and image frames within a defined range before and after that time node are uniformly classified into the same post-rain high reflectivity risk period. Simultaneously, the light intensity records of these image frames are associated and saved with the external environmental parameters at the time of acquisition, ensuring that each image frame has traceable lighting conditions and risk level attributes. In this way, subsequent steps can compare the trend of image brightness changes chronologically and make a comprehensive judgment based on changes in lighting conditions, thus providing a continuous data source for the statistical analysis of the duration of abnormal brightness.

[0026] After retrieving the image data for the corresponding time period, the brightness changes of all image frames within that period are continuously statistically analyzed. Specifically, the retrieved image data is arranged sequentially on the time axis, and the duration range of abnormal brightness increases is determined by comparing the amplitude and direction of the overall brightness changes between adjacent time frames. When the average brightness value of multiple consecutive frames maintains an upward trend compared to the baseline brightness level of the image before rainfall, and the increase exceeds the normal fluctuation range of ambient light, it can be determined that there is a continuous brightness anomaly during that time period. This phenomenon is usually directly related to the specular reflection characteristics of the water film on the surface of leaves after rain and has obvious temporal continuity. Therefore, at this stage, each time period in which the brightness remains high needs to be recorded as a brightness anomaly duration range, and it is associated with and stored with the corresponding time node, illumination parameters, and image frame number. By establishing the brightness anomaly duration range, not only can the duration of the reflective phenomenon be revealed, but it can also provide a time dimension filtering basis for subsequent spatial extraction steps.

[0027] After obtaining the duration of abnormal brightness, the distribution of pixels with high-brightness characteristics in each image frame is analyzed based on the corresponding set of image frames. To ensure that the extracted regions accurately reflect the spatial characteristics of reflective effects, the spatial continuity of high-brightness regions in each frame is determined by combining the brightness statistics from the previous stage. Specifically, by comparing the distribution and morphological changes of high-brightness regions in adjacent time frames, groups of regions that are temporally continuous and spatially adjacent are identified, and these regions are integrated into continuous image regions with consistent reflective characteristics. Since the reflection of water film on leaf surfaces after rain is usually concentrated within a specific geometric direction and has obvious regional connectivity, the combined determination of temporal and spatial continuity can effectively distinguish between sporadic illumination reflection and continuous water film reflection. In this process, each group of continuous image regions is recorded as a candidate region for reflectivity, along with information on the corresponding time period, brightness increase, and light intensity change, providing spatial data support for the subsequent determination of the reflectivity influence area.

[0028] After extracting continuous image regions with reflective features, all candidate reflective regions need to be aggregated to form an overall reflective influence area. Specifically, multiple candidate reflective regions that are close to or partially overlap each other within the same time period are spatially fused to generate a unified reflective influence area description. This reflective influence area not only spatially covers all image regions with continuous reflective features but also temporally corresponds to high-reflection-risk periods in the rainfall time record sequence, thus achieving a dual binding of time and space. To ensure that the reflective influence area accurately reflects the degree of interference of the post-rain environment on image recognition, the boundary information of this area also needs to be correlated with the duration of brightness anomalies, so that each sub-region within the reflective influence area contains a clear brightness duration attribute. In this way, the resulting reflective influence area can serve as a reference for a specific spatial range in subsequent vegetation identification, brightness adjustment, and coverage calculation processes, thereby limiting the propagation of misjudgments and reducing recognition bias caused by water film reflection. This process realizes a complete technical link from identifying high reflectivity risk conditions after rain to defining reflective spatial areas, enabling automatic vegetation coverage calculation to identify and isolate reflective interference factors in a short time after rainfall, providing a solid data foundation and spatial constraints for subsequent reliable vegetation extraction and brightness adjustment steps.

[0029] Based on the re-comparison of the original vegetation identification results with the range of the reflected light impact area, the spatial continuity change of vegetation within the reflected light impact area is analyzed, the vegetation areas that were incorrectly excluded are marked, and a reliable vegetation reference area is generated. Once the reflective area is defined, it's necessary to re-compare the original vegetation identification results within this area to identify and restore vegetation areas that were mistakenly excluded due to post-rain reflection. By analyzing the changes in the spatial continuity of vegetation within the reflective area, incorrectly excluded vegetation pixel groups can be accurately marked, thereby generating a reliable vegetation reference area. This provides a stable spatial basis for subsequent brightness adjustments and coverage smoothing calculations. The specific steps are as follows: After defining the area affected by reflective light, the original vegetation identification results are re-extracted and located to ensure that all data involved in the comparison correspond one-to-one with the reflective light affected area in the spatial dimension. During this process, areas marked as "non-vegetation" in the original identification results are spatially overlaid with the reflective light affected area, and the overlaid portion serves as a preliminary set of potential false exclusion areas. The technical purpose of this operation is to determine the specific image locations affected by reflective light interference through spatial overlap, allowing subsequent analysis to focus on areas with interference characteristics without affecting the vegetation areas in the normal identification results. Simultaneously, this stage also requires combining the temporal information accompanying the reflective light affected area to pair consecutive image frames within the same time period to ensure that the spatial comparison results remain consistent in the temporal dimension. In this way, a re-identification dataset is formed with time as the sequence and the reflective light spatial range as the core, providing accurate input conditions for subsequent spatial continuity analysis.

[0030] After obtaining the re-identification dataset, a systematic analysis was conducted on the spatial continuity changes of vegetation within the reflected light impact area. Specifically, by comparing the morphological differences of vegetation areas before rainfall, during reflection, and in the post-rain recovery phase, spatial breaks, boundary contraction, and area omissions caused by reflection were identified. Since vegetation typically exhibits continuous distribution characteristics in its natural state, while reflection interference often causes sudden discontinuities in local areas, analyzing the trend of spatial continuity changes allows for a direct assessment of which areas are incorrectly excluded due to reflection. This process requires comprehensively considering the correspondence between the boundary contours of the reflected light impact area and the boundaries of the original vegetation areas. When an area is found to have a natural extension trend with the original vegetation boundary in space, but is interrupted or missing during the reflection period, this area can be identified as a potential area of ​​incorrectly excluded vegetation. Through this spatial morphological comparison method, the specific location of incorrectly excluded areas can be quickly located without re-identifying the entire image, providing a spatial basis for subsequent labeling.

[0031] After identifying potential mis-excluded areas, these areas are marked and classified to establish a complete set of incorrectly excluded vegetation regions. Specifically, each identified potential mis-excluded area is bound to its corresponding time point, reflective intensity information, and original vegetation region boundary data, forming a traceable marking record. To ensure the spatial integrity of the marking, adjacent and continuous mis-excluded areas need to be merged to generate connected regions with a certain area and morphological consistency. The technical significance of this step lies in integrating discrete groups of mis-excluded pixels into vegetation region units with consistent attributes through region fusion and feature binding, enabling these regions to be identified and referenced as independent spatial entities in the subsequent reliable reference generation process. Simultaneously, the spatial correspondence between the reflective influence area and these mis-excluded areas must be maintained during the marking process, ensuring that the marking results reflect the causal relationship between reflective influence and vegetation identification deviation, providing structured input for generating reliable vegetation reference regions.

[0032] After marking the incorrectly excluded vegetation areas, a reliable vegetation reference area is generated by combining the range of the reflected light-affected area with the original vegetation identification results. Specifically, the correctly identified vegetation areas from the original identification results are merged with the marked incorrectly excluded vegetation areas to form a new vegetation spatial distribution map. To ensure that this reference area is representative of the real environment, the merged area is also time-stratified according to the temporal attributes of the reflected light-affected area, based on the chronological order of the reflected light duration. This ensures that the reference area not only spatially covers the complete vegetation distribution range but also temporally reflects the continuity of vegetation recovery. In this way, the reliable vegetation reference area corrects for post-rain reflected light interference in both spatial and temporal dimensions, providing a stable benchmark for brightness adjustment in subsequent steps. The generated reliable vegetation reference area can be considered a spatial compensation result for the reflected light effect, constraining the vegetation identification boundary in the subsequent brightness adjustment stage and preventing identification deviations caused by local brightness anomalies. This process achieves a complete closed loop from identifying the range of reflective influence, analyzing spatial continuity, marking incorrectly excluded areas to generating reference areas, enabling the automatic calculation of vegetation cover to still have environmental adaptability and calculation stability under complex climatic conditions.

[0033] Using a reliable vegetation reference area as a constraint, the brightness judgment range used in the vegetation identification process is adjusted so that the reflective area gradually returns to the control of the reliable vegetation reference area in the time dimension, forming a brightness adjustment rhythm line. To enable vegetation identification to dynamically adapt to abnormal brightness changes under post-rain conditions, a reliable vegetation reference area needs to be used as a constraint to adjust the brightness judgment range used in the vegetation identification process. This allows the reflective areas to gradually return to the control of the reliable vegetation reference area over time, forming a continuous brightness adjustment rhythm. Through this process, vegetation identification can achieve a smooth transition from an abnormal state to a stable state over time, thus providing a reliable brightness control basis for subsequent coverage calculations. The specific steps are as follows: After generating a reliable vegetation reference region, this region serves as the spatial constraint basis for vegetation identification, establishing a reference framework for brightness adjustment. Specifically, the spatial range and temporal attributes of the reliable vegetation reference region are synchronously mapped to the original image sequence, ensuring that the parts overlapping with the reference region in each frame are given priority during identification. In this way, vegetation identification no longer relies solely on the brightness features of a single frame, but dynamically corrects the brightness of the corresponding position in the current frame based on the historical spatial distribution of the reference region. At this stage, the boundary information of the reference region needs to be matched with the boundary information of the reflective area, establishing a one-to-one spatial correspondence. This matching process provides a clear reference target for subsequent brightness adjustments, limiting the reasonable range of brightness judgment through the reliable vegetation reference region, thereby preventing bright pixels in the reflective area from being misidentified as non-vegetation areas during identification.

[0034] After establishing a spatial matching relationship constrained by a reliable vegetation reference area, the brightness judgment range used in vegetation identification is dynamically adjusted. Specifically, the brightness distribution characteristics within the reflective area need to be analyzed over time, comparing them with the average brightness, brightness fluctuation amplitude, and light recovery trend of the reference area. Based on this, the brightness judgment boundary used to distinguish between vegetation and non-vegetation in vegetation identification is gradually corrected, causing it to gradually converge to the brightness range of the reference area over time within the reflective area. This adjustment process is continuous in time; that is, as time passes after rain, the brightness judgment range gradually tightens from a relatively wide tolerance range to the standard range under normal identification conditions, thus achieving adaptive control of the identification process in response to changes in light intensity. In this way, the vegetation identification process is no longer constrained by short-term abnormal brightness fluctuations, but rather, through a time-adjustment mechanism, the adjustment of the brightness judgment range is synchronized with the natural recovery process of environmental brightness.

[0035] After dynamically adjusting the brightness judgment range, the reflective area needs to be gradually brought back to the control range of the reliable vegetation reference area over time. The core of this step lies in tracking brightness changes across continuous time frames, allowing the brightness characteristics of the reflective area to gradually approach the normal brightness distribution of the reference area at each moment. Specifically, by analyzing the attenuation trend of bright pixels in the reflective area, the relationship between its brightness recovery rate and the stable brightness state of the reference area is identified, and the brightness weight allocation during the recognition process is adjusted accordingly. As the brightness of the reflective area gradually decreases and approaches the brightness level of the reference area, the recognition process will gradually reintegrate this area into the vegetation recognition range. Simultaneously, the brightness recovery information in the time dimension needs to be consistent with the changes in the reference area boundary in the spatial dimension, ensuring that the regression of the reflective area not only approaches the reference state in brightness values ​​but also merges with the reliable vegetation reference area in spatial range. Through this time-driven brightness regression process, the system can automatically restore recognition accuracy shortly after rain, ensuring that the coverage calculation maintains a continuous and stable trend over time.

[0036] After the reflective area gradually returns to the control of the reliable vegetation reference area, a brightness adjustment rhythm line needs to be formed throughout the entire time series to dynamically guide the subsequent identification process. Specifically, the brightness adjustment amplitude, recovery rate, and brightness benchmark of the reference area at each time point are integrated to form a time curve describing the brightness change pattern; this curve is the brightness adjustment rhythm line. The brightness adjustment rhythm line reflects the complete evolution process of the reflective area from an abnormally bright state to a stable brightness state and can serve as the basis for brightness control in the vegetation identification stage. During the automatic vegetation coverage calculation process, each new image acquisition will automatically select the corresponding brightness adjustment strategy based on the current time point's position in the rhythm line, ensuring that the identification process always remains consistent with environmental changes. In this way, vegetation identification no longer relies solely on static threshold judgment but is driven by the time series, continuously adjusting the brightness weight according to the trend of the rhythm line, so that the reflective area is always confined to the control of the reference area in the time dimension. This rhythm line not only records the entire process of brightness adjustment, but also provides a quantitative basis for subsequent brightness adjustment steps, thus making the entire recognition and measurement process continuous in time, traceable in space, and adaptive in brightness control.

[0037] The brightness impact on the reflective area is adjusted in segments according to the brightness adjustment rhythm line. By reducing the discrimination weight of the reflective area in a short period of time and gradually restoring the vegetation display characteristics, the continuous and stable calculation of vegetation coverage after rain is completed. To prevent post-rain brightness anomalies caused by water film reflection from interfering with vegetation identification results, it is necessary to adjust the brightness changes in the reflective areas segmentally according to a brightness adjustment rhythm line. This involves temporarily reducing the discrimination weight of the reflective areas and gradually restoring the vegetation's display characteristics as light conditions recover, thus achieving continuous and stable calculation of post-rain vegetation coverage. This process not only achieves a dynamic transition from brightness anomalies to stable identification but also ensures the smoothness of the calculation results in the temporal dimension and the consistency in the spatial dimension. The specific steps are as follows: After establishing a brightness adjustment rhythm line, the brightness change trend within the reflective area is segmented over time to create a time framework for brightness adjustment. Specifically, the brightness adjustment rhythm line is divided according to the brightness recovery rate and the duration of reflectivity, decomposing the time axis into multiple consecutive brightness adjustment stages, each corresponding to a specific brightness recovery level. In this way, the reflective area is no longer considered a single state in the time dimension but is divided into multiple brightness change stages, allowing subsequent brightness adjustments to employ differentiated adjustment strategies for different recovery stages. In this stage, the brightness change rate, reflectivity attenuation trend, and brightness recovery value of the reference vegetation area recorded in the brightness adjustment rhythm line need to be correlated, ensuring that the brightness adjustment interval of each time segment is synchronized with the recovery level of the reference area. Through this time segmentation, the brightness adjustment of the reflective area acquires a temporal structure, enabling subsequent weight adjustments and vegetation feature recovery to be seamlessly integrated over time.

[0038] After establishing the time segments, the discrimination weight of the reflective area is systematically reduced based on the brightness characteristics of each stage in the brightness adjustment rhythm line. Specifically, the participation weight of the reflective area in vegetation identification is gradually reduced according to the brightness recovery level corresponding to the time segment, ensuring that its impact on the overall identification result during periods of abnormal brightness is controlled within an acceptable range. This weight reduction process is not a one-time operation but is executed sequentially according to the time segments, keeping the reflective area under control throughout the entire brightness recovery cycle. Simultaneously, it is necessary to ensure that the reflective area maintains spatial correspondence with the reliable vegetation reference area while reducing the weight, preventing discontinuity in the identification range during the weight reduction process. In this way, areas with high brightness but potential vegetation features are not completely excluded but are retained as potential vegetation candidate areas under reduced weight, and are reintroduced into the vegetation display range after subsequent brightness recovery. This process effectively prevents the problem of instantaneous false rejection caused by post-rain reflection, laying the foundation for the gradual recovery of subsequent vegetation display features.

[0039] After the weight of the reflective area is reduced, the vegetation display characteristics of the reflective area need to be gradually restored according to the brightness adjustment rhythm line over time and as ambient light stabilizes. Specifically, by tracking the decreasing trend of brightness changes in each time segment, when the brightness level approaches the normal range of the credible vegetation reference area, the weight of this area is gradually increased, enabling it to regain the same discrimination ability as normal vegetation areas during the identification process. During this process, the speed and magnitude of the weight increase must be consistent with the trend of the rhythm line to ensure a smooth and continuous brightness adjustment process. When the brightness of the reflective area gradually returns to the brightness distribution range of natural vegetation, the area will be re-identified by the system as a valid vegetation area, thus restoring its statistical status in the overall coverage calculation. Through this gradual recovery mechanism, the identification state of the reflective area can naturally transition with changes in light intensity, achieving a smooth transition from an abnormally bright state to a stable vegetation state, ensuring that the calculation results are continuous and without abrupt changes over time.

[0040] After completing the segmented adjustment and weight restoration of the reflective area, the brightness adjustment results across all time segments need to be integrated to form a complete coverage calculation result. In practice, the changes in vegetation identification weights and corresponding brightness adjustment amplitudes within each stage are uniformly summarized to generate a time-continuous coverage calculation sequence. This calculation sequence not only reflects the cumulative changes in vegetation identification results across time periods but also embodies the overall rhythm of brightness restoration and vegetation display feature regression over time. To ensure the stability of the calculation results, the coverage calculation sequence needs to be aligned with the brightness adjustment rhythm line, ensuring that the temporal trend of coverage changes is synchronized with the brightness restoration process. Once the brightness and weight adjustments for all time periods are completed, the final coverage result will appear as a smooth transition curve from the initial post-rain moment to the light restoration stage. This curve accurately reflects the actual coverage changes of vegetation in the post-rain environment, avoiding the problem of sudden coverage drops caused by short-term brightness anomalies. Through this process, the post-rain vegetation coverage calculation not only maintains continuous stability in results but also achieves dynamic adaptation to changes in environmental brightness, giving the calculation results high temporal consistency and high environmental robustness.

[0041] This invention establishes a rainfall time record sequence during image acquisition and combines it with information on changes in ambient light intensity. It marks high-reflection-risk states after rain, achieving proactive identification and time-layered management of reflective effects. This allows the image recognition process to adaptively adjust the recognition range for different climatic conditions. This method effectively avoids the problem of false exclusion of vegetation areas due to water film reflection, ensuring that the calculated results remain stable and consistent with the actual greening status for a short period after rain. This fundamentally improves the reliability and continuity of vegetation coverage calculation under changing environmental conditions.

[0042] This invention introduces a reliable vegetation reference area as a dynamic constraint during brightness adjustment, allowing the reflective area to gradually return to its normal recognition range over time, and achieving segmented adjustment based on the brightness adjustment rhythm. Through this continuous control mechanism, vegetation display characteristics can gradually return to their natural state during the light recovery phase, resulting in a smooth transition in the measurement results over time. This avoids sudden drops in coverage caused by short-term reflections, ensuring stable output and accurate reflection of greening monitoring results.

[0043] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An automatic method for calculating vegetation coverage in municipal green belts based on image extraction, characterized in that, Includes the following steps: During the automatic calculation of vegetation coverage, a rainfall time record sequence is established, and the ambient light intensity change information is recorded synchronously during image acquisition. The high reflectivity risk status after rain is marked at the corresponding time nodes of the rainfall time record sequence. The steps for marking a high reflectivity risk status after rain are as follows: During the automatic calculation of vegetation coverage, synchronous records of time information and environmental status information are obtained. A time series is established using the timestamp information of the image acquisition device, and the time series is correlated with the rainfall monitoring records on the time axis. After completing the synchronized recording of time series and illumination information, a rainfall time record sequence is established based on the start time, end time and duration of the rainfall event, and the sequence includes illumination intensity, air humidity and image acquisition status information; After the rainfall time record sequence is established, the information on changes in light intensity is continuously analyzed, and the areas of sudden changes in light intensity are compared with the rainfall time record sequence. The high reflectivity risk status after the rain ends is marked shortly after the rain ends. After marking the high reflectivity risk status after rain, the rainfall time record sequence and light intensity change information are integrated to form a time index dataset, which provides time dimension input for vegetation identification and supports subsequent brightness adjustment; Based on the high reflectivity risk after rain, image data for the corresponding time period is retrieved, the duration of abnormal brightness increase in the image is statistically analyzed, and continuous image regions with reflective characteristics are extracted to form the range of reflective influence area. Based on the re-comparison of the original vegetation identification results with the range of the reflected light impact area, the spatial continuity change of vegetation within the reflected light impact area is analyzed, the vegetation areas that were incorrectly excluded are marked, and a reliable vegetation reference area is generated. The steps for generating a reliable vegetation reference region are as follows: After the area affected by reflected light is defined, the original vegetation identification results are extracted and located again. The areas marked as non-vegetation are spatially superimposed with the area affected by reflected light, and combined with time information to form a re-identification dataset. After obtaining the re-identification dataset, the changes in the spatial continuity of vegetation within the area affected by reflection were analyzed. By comparing the differences in vegetation morphology before rainfall, during reflection, and after rainfall, the spatial breaks and regional gaps caused by reflection were identified. After identifying potential areas for false exclusion, these areas are marked and classified, and time nodes, reflectivity information and vegetation boundary data are bound and fused to generate connected regions. After marking is completed, the normally identified vegetation areas and the incorrectly excluded areas are merged, and the time layer is performed according to the duration of reflection to generate a reliable vegetation reference area with continuity in space and time. Using a reliable vegetation reference area as a constraint, the brightness judgment range used in the vegetation identification process is adjusted so that the reflective area gradually returns to the control of the reliable vegetation reference area in the time dimension, forming a brightness adjustment rhythm line. The steps for generating the brightness adjustment rhythm line are as follows: After generating a reliable vegetation reference region, the spatial extent and temporal attributes of the reliable vegetation reference region are synchronously mapped to the original image sequence, so that the overlapping parts are given priority in the recognition process, and a brightness adjustment reference frame is established. After establishing a brightness adjustment reference framework, the brightness distribution characteristics in the reflective area are analyzed over time and compared with the average brightness and illumination recovery trend of the reference area to dynamically adjust the brightness judgment range. After adjusting the brightness judgment range, the brightness change tracking of continuous time frames is used to gradually bring the reflective area back to the control of the reliable vegetation reference area in the time dimension, and to achieve fusion with the reference area in the spatial dimension. After the reflection-affected area is controlled, the brightness adjustment range and recovery rate at each time point are integrated with the brightness benchmark of the reference area to form a brightness adjustment rhythm line that describes the brightness change pattern. The brightness of the reflective area is adjusted in segments according to the brightness adjustment rhythm line. The discrimination weight of the reflective area is reduced in a short period of time and the vegetation display characteristics are gradually restored. The steps for adjusting the reflective area in segments according to the brightness adjustment rhythm line, reducing the discrimination weight of the reflective area, and gradually restoring the vegetation display characteristics are as follows: After establishing the brightness adjustment rhythm line, the brightness change trend in the area affected by reflection is segmented over time, and a time frame for brightness adjustment is established based on the brightness recovery speed and the duration of reflection. After establishing the time frame, the discrimination weight of the reflective influence area is systematically reduced according to the brightness characteristics of each stage in the brightness adjustment rhythm line, while maintaining the spatial correspondence with the credible vegetation reference area. After the weight of the reflective area is reduced, the vegetation display characteristics of the reflective area are gradually restored according to the brightness adjustment rhythm line, so that the reflective area can regain the ability to distinguish normal vegetation area. After completing the segmented adjustment and weight restoration, the brightness adjustment results in each time segment are integrated to generate a time-continuous coverage calculation sequence, and a continuous and stable measurement result of the post-rain vegetation coverage is formed.

2. The method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction according to claim 1, characterized in that, The formation process of the reflective area is as follows: After establishing the post-rain high reflectivity risk status, based on the risk time marker information in the rainfall time record sequence, image data for the corresponding time period is retrieved from the image acquisition dataset, and the light intensity record value is associated with and saved with environmental parameters; After retrieving the image data for the corresponding time period, the brightness change amplitude of adjacent image frames on the time axis is continuously statistically analyzed to determine the duration range of abnormal brightness increase and establish the abnormal brightness duration range. After obtaining the duration of the brightness anomaly, the distribution area of ​​bright pixels is analyzed based on the image frame set. The spatial continuity of the bright area in adjacent time frames is judged, and continuous image areas with reflective features are extracted. After extracting continuous image regions, spatially fuse partially overlapping regions within the same time period to form a reflective influence area range that maintains a correspondence in both time and space dimensions.

3. The method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction according to claim 2, characterized in that, During the formation of the reflective influence area, the boundary information of each reflective candidate area is correlated with the duration of brightness anomaly, so that each sub-region within the reflective influence area has a clear brightness duration attribute.

4. The method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction according to claim 1, characterized in that, When generating a reliable vegetation reference area, the temporal attributes and spatial boundary information of the reflected light influence area are synchronously mapped to the merged vegetation area. The temporal hierarchy of the reference area is adjusted according to the order of the duration of the reflected light, so that the vegetation reference area remains continuous in spatial distribution and reflects the gradual changes of vegetation restoration in time series.

5. The method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction according to claim 1, characterized in that, In the process of forming the brightness adjustment rhythm line, the brightness recovery rate of the reflective area in the time dimension is correlated with the brightness stability trend of the reliable vegetation reference area, so that the brightness adjustment rhythm line reflects both the brightness change amplitude and the spatial regression state.

6. The method for automatically calculating the vegetation coverage rate of municipal green belts based on image extraction according to claim 1, characterized in that, When gradually restoring the vegetation display characteristics of the reflective area according to the brightness adjustment rhythm line, the brightness recovery rate is compared synchronously with the brightness distribution of the reference vegetation area to ensure that the brightness recovery process of the reflective area is consistent with the changes in ambient light in time, thereby ensuring the temporal continuity and spatial stability of the coverage calculation.

Images