Greening plant withering detection method and system based on dynamic patrol

Abstract

This invention discloses a method and system for detecting withered green plants based on dynamic inspection. The method includes: acquiring video streams and GPS data collected by a mobile vehicle; continuously observing, compensating for occlusion, and tracking multiple targets in the green areas of the video stream to obtain an image sequence of the green areas; performing comprehensive analysis of the image sequence of the green areas in terms of color, posture, and context to identify and mark color and posture abnormalities of potentially withered vegetation, and comparing the state of individual plants with that of surrounding vegetation to obtain suspected withered targets; wherein, the posture includes the degree of drooping and curling of the vegetation; and analyzing changes in color and posture based on the GPS data through historical data to confirm the withering status of the suspected withered targets. By implementing the method of this invention, the ability to automatically and dynamically monitor the health status of urban green vegetation, automatically identify withered vegetation, and quickly issue early warning information can be achieved.

Description

Technical Field

[0001] This invention relates to computer vision, and more specifically to a method and system for detecting wilting of green plants based on dynamic inspection. Background Technology

[0002] In urban environments, the green belts along roadsides and the vegetation in central medians play an irreplaceable role in enhancing the aesthetic appeal and ecological quality of cities. However, due to factors such as pest and disease infestation, drought, improper maintenance, or natural aging, these vegetation plants may wither or even die. Traditionally, the discovery of such problems has relied primarily on manual on-site inspections, a method with several limitations: low inspection efficiency, limited coverage, inconsistent judgment criteria, and a long cycle from problem discovery to implementation.

[0003] Faced with the above challenges, existing technological solutions have failed to effectively meet practical needs, particularly in terms of timeliness and accuracy. Specifically, manual inspections cannot achieve efficient and comprehensive monitoring of large areas; different inspectors identify withered plants based on personal experience and subjective judgment, which may lead to misjudgments or omissions; at the same time, the entire process involves multiple stages, and the time delay between the discovery of anomalies and the implementation of corresponding measures is significant, affecting the effectiveness and timeliness of the response.

[0004] Therefore, it is necessary to design a new method to enable automated and dynamic monitoring of the health status of urban green vegetation, automatic identification of withered vegetation, and rapid issuance of early warning information. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for detecting the wilting of green plants based on dynamic inspection.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting wilting of green plants based on dynamic inspection, comprising:

[0007] Acquire video streams and GPS data collected by mobile vehicles;

[0008] The green areas in the video stream are continuously observed, occlusion compensated, and multi-target tracked to obtain a sequence of green area images;

[0009] A comprehensive analysis of the image sequence of the green area is performed on color, pose, and context dimensions to identify and mark areas of color and pose abnormalities in potential withered vegetation, and to compare the state of individual plants with that of surrounding vegetation to obtain suspected withered targets; wherein, the pose includes the degree of drooping and curling of the vegetation.

[0010] Based on the GPS data, the changes in color and posture are analyzed through historical data to confirm the vegetation withering status of the suspected withered target.

[0011] The further technical solution is as follows: The continuous observation, occlusion compensation, and multi-target tracking of the green areas in the video stream to obtain a sequence of green area images includes:

[0012] Identify and track the green areas in each frame of the video stream;

[0013] Detect whether the green areas in the image frame are occluded;

[0014] When the green area in an image frame is occluded, compensate for the occluded part or exclude image frames whose occlusion level meets the requirements to obtain the occlusion processing result;

[0015] The same plant is located in consecutive frames of the occlusion processing result and assigned an identifier to obtain a sequence of green area images.

[0016] The further technical solution is as follows: when the green area in an image frame is occluded, the occluded portion is compensated or image frames with occlusion levels meeting the requirements are excluded to obtain the occlusion processing result, including:

[0017] When the green area in an image frame is occluded, the vegetation outline of the occluded part is predicted and completed based on the information of the preceding and following image frames, or image frames with the required degree of occlusion are excluded, so as to obtain the occlusion processing result.

[0018] The further technical solution is as follows: A comprehensive analysis of the image sequence of the green area, considering color, pose, and context, is performed to identify and mark areas of abnormal color and pose of potentially withered vegetation, and the condition of individual plants is compared with that of surrounding vegetation to obtain suspected withered targets, including:

[0019] Calculate the RGB value of each pixel in the image sequence of the green area and apply the VARI formula to identify healthy vegetation and withered vegetation, and mark areas with abnormal colors;

[0020] Texture and key point detection are performed on the color anomaly areas to calculate the degree of vegetation drooping and curling, and the posture anomaly areas are marked.

[0021] Identify the status of individual plants and adjacent vegetation, and compare and judge whether the greening plants are high-confidence individual anomalies or regional characteristics;

[0022] When the green plant is a high-confidence individual anomaly and has areas of abnormal color and posture, the green plant is identified as a suspected withering target.

[0023] The further technical solution is as follows: The process of detecting texture and key points in the color-abnormal areas, calculating the degree of vegetation drooping and curling, and marking posture-abnormal areas includes:

[0024] The color anomaly areas were analyzed in depth, and texture analysis and key point detection techniques were used to identify vegetation edges and tips.

[0025] By quantifying the angle of key points at the tips relative to the direction of gravity, the degree of drooping and curling of vegetation is measured, and areas of abnormal posture are marked.

[0026] The further technical solution is as follows: the identification of individual plants and the status of adjacent vegetation, and the comparison and judgment of whether the greening plants are high-confidence individual anomalies or regional characteristics, includes:

[0027] Instance segmentation technology is used to identify the status of individual vegetation plants and their neighboring vegetation.

[0028] When the similarity between the green plant and the surrounding healthy vegetation does not meet the requirements, the green plant is identified as a high-confidence individual anomaly; when the similarity between the green plant and the surrounding healthy vegetation meets the requirements, the green plant is identified as a regional feature.

[0029] The further technical solution is as follows: the state similarity includes color similarity and droop index similarity.

[0030] The further technical solution is as follows: The confirmation of vegetation withering of the suspected withering target based on the GPS data combined with multiple anomaly indicators and through historical data analysis includes:

[0031] The data packets of the suspected withered targets are matched with the corresponding location patrol data in the historical database to compare the multidimensional health status at different time points; wherein, the data packets include GPS data, corresponding green area images, and analytical evidence; the multidimensional health status includes color status and posture status;

[0032] If the multidimensional health status of the suspected withered target shows a continuous deterioration in the regional health status and does not conform to the normal seasonal change pattern, then the suspected withered target will be confirmed as a withered target.

[0033] Its further technical solution is as follows: After confirming the vegetation withering situation of the suspected withering target based on the GPS data combined with multiple anomaly indicators and through historical data analysis, it includes:

[0034] An early warning work order is generated and pushed to the terminal. The early warning work order includes GPS data, on-site screenshots, multi-dimensional analysis evidence, and confirmation time.

[0035] This invention also provides a greening plant wilting detection system based on dynamic inspection, comprising:

[0036] The acquisition unit is used to acquire video streams and GPS data collected by the mobile vehicle.

[0037] An image sequence generation unit is used to continuously observe, compensate for occlusion, and track multiple targets in the green area of ​​the video stream to obtain an image sequence of the green area.

[0038] The suspected target identification unit is used to perform a comprehensive analysis of the image sequence of the green area in terms of color, posture, and context, identify and mark areas of color and posture abnormalities of potential withered vegetation, and compare the state of a single plant with the surrounding vegetation to obtain suspected withered targets; wherein, the posture includes the degree of drooping and curling of the vegetation.

[0039] The withering condition determination unit is used to analyze changes in color and posture based on the GPS data and historical data to confirm the withering condition of the vegetation of the suspected withering target.

[0040] The advantages of this invention compared to existing technologies are as follows: By integrating video streams and GPS data collected by mobile vehicles, the system can perform automated and dynamic health monitoring of urban green areas. It utilizes continuous observation, occlusion compensation, and multi-target tracking technologies to process the video streams, generating image sequences of green areas that include a time dimension. Then, it comprehensively analyzes the color, posture, and contextual information of the vegetation in these image sequences to identify and mark potential withering areas, and compares the state changes of individual plants with surrounding vegetation to confirm suspected withering targets. Combined with GPS data and its historical records, the system can not only accurately locate and assess vegetation withering but also quickly issue early warning information, thereby achieving real-time monitoring and management of urban green vegetation health. This process fully demonstrates the system's intelligent and refined characteristics, providing a strong guarantee for improving urban ecological quality and landscape aesthetics.

[0041] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 A flowchart illustrating the method for detecting wilting of green plants based on dynamic inspection provided in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of image frames of a video stream provided in an embodiment of the present invention;

[0045] Figure 3 This is a schematic diagram of a green area image sequence provided in an embodiment of the present invention;

[0046] Figure 4 A schematic diagram of a suspected withered target provided in an embodiment of the present invention;

[0047] Figure 5 This is a schematic diagram illustrating vegetation withering as provided in an embodiment of the present invention.

[0048] Figure 6 A schematic block diagram of a greening plant wilting detection system based on dynamic inspection provided in an embodiment of the present invention;

[0049] Figure 7 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0050] 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, not all, of the embodiments of the present invention. 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.

[0051] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0052] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0053] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0054] Please see Figure 1 , Figure 1This is a flowchart illustrating the dynamic patrol-based method for detecting wilting vegetation in green spaces, as provided in this embodiment of the invention. This method is applied to a server. It acquires video streams and GPS data collected by mobile vehicles, continuously observes, compensates for occlusion, and tracks multiple targets within the green areas of the video stream to generate image sequences. Color, posture (including drooping and curling degrees), and contextual dimensions are used to comprehensively analyze and identify areas with abnormal color and posture of potentially wilting vegetation. The status of individual plants is compared with surrounding vegetation to determine suspected wilting targets. Subsequently, combined with GPS data, historical data analysis is used to confirm the health status trends of these targets. If continuous deterioration is found and it does not conform to normal seasonal change patterns, it is ultimately confirmed as a wilting target. Finally, the system automatically generates an early warning work order containing GPS coordinates, on-site screenshots, multi-dimensional analysis evidence, and confirmation time, and pushes it to relevant maintenance departments, achieving automated and dynamic monitoring of the health status of urban green vegetation and the ability to quickly issue early warning information. This method ensures timely and accurate identification of wilting vegetation and the implementation of corresponding measures, improving the maintenance efficiency and management level of urban green spaces.

[0055] Figure 1 This is a flowchart illustrating the method for detecting wilting of green plants based on dynamic inspection, provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S140.

[0056] S110: Acquire video streams and GPS data collected by the mobile vehicle.

[0057] In this embodiment, as Figure 2 As shown, a video stream refers to a continuous sequence of images dynamically captured by high-definition cameras mounted on mobile vehicles (such as municipal patrol vehicles, buses, or sanitation vehicles) during normal vehicle operation, covering both sides of the road and the central green belt. These images are recorded in video format, providing real-time visual information about the green vegetation. The video stream includes not only static images but also information that changes over time, enabling the system to continuously observe the same plant from multiple angles and identify and track specific green areas by analyzing these consecutive frames, maintaining tracking accuracy even in the presence of occlusion.

[0058] GPS data refers to the precise geographic coordinate information provided by the vehicle-mounted Global Positioning System (GPS) module. Every mobile vehicle used for data collection is equipped with a GPS device, which records the vehicle's location and movement trajectory in real time. GPS data is crucial for monitoring green vegetation because it is synchronized with the video stream, providing a geographic location tag for each frame in the video. This means the system can accurately determine the actual geographic location corresponding to a specific frame or portion of the video stream, making the monitoring of green vegetation health not only continuous in time but also precisely spatially located. Furthermore, GPS data helps the system match corresponding location inspection data in the historical database. By comparing the vegetation status at the same location at different times, it can confirm whether vegetation withering conforms to seasonal change patterns or other anomalies. Combining video stream and GPS data, the entire system can achieve automated and dynamic monitoring of the health status of urban green vegetation and issue timely early warning information.

[0059] S120. Continuously observe, compensate for occlusion and track multiple targets in the green area in the video stream to obtain a sequence of green area images.

[0060] In this embodiment, the green area image sequence refers to a series of continuous images of green vegetation (such as street trees and shrubs) obtained by processing the video stream captured by the moving vehicle. These images not only include visual information about the green vegetation, but also ensure that specific plants in each frame can be accurately identified and tracked through a multi-target tracking algorithm, maintaining continuity and consistency even in the presence of occlusion. The green area image sequence forms the basis for subsequent health analysis, providing the system with the necessary data support to accurately assess the health status of the vegetation.

[0061] In one embodiment, step S120 described above may include steps S121 to S124.

[0062] S121. Identify and track the green areas in each frame of the video stream.

[0063] This step aims to identify green areas from the video stream and continuously track the plants within these areas. By utilizing computer vision technology, the system can automatically detect all green areas appearing in the video frames and use a multi-object tracking algorithm to locate each plant, assigning them a unique temporary ID. This ensures that subsequent analysis is based on a continuous sequence of images of the same plant, rather than isolated images.

[0064] S122. Detect whether the green areas in the image frame are occluded.

[0065] In this step, the system examines each frame of the image to determine if any green areas are temporarily obscured by pedestrians, vehicles, or other objects. This step is crucial because occlusion can affect the accurate assessment of vegetation condition.

[0066] S123. When the green area in an image frame is occluded, compensate for the occluded part or exclude image frames whose occlusion level meets the requirements to obtain the occlusion processing result.

[0067] In this embodiment, the occlusion processing result refers to the processed image frame. For the occluded portion, the system predicts and completes the occluded vegetation outline based on information from preceding and following frames, or directly excludes "inferior frames" that are so severely occluded that they cannot provide useful information. This strategy ensures that only clear and useful image frames are used in subsequent analysis processes, thereby improving the accuracy of the analysis.

[0068] When the green area in an image frame is occluded, the vegetation outline of the occluded part is predicted and completed based on the information of the preceding and following image frames, or image frames with the required degree of occlusion are excluded, so as to obtain the occlusion processing result.

[0069] Specifically, when the green area in an image frame is occluded, the system will predict and complete the vegetation outline of the occluded part based on the information of the previous and next frames, or choose to ignore the frame if the occlusion is too severe, and only retain those "high-quality frames" with low occlusion and that can provide effective information.

[0070] S124. Locate the same plant in consecutive frames of the occlusion processing result, assign an identifier to it, and obtain a sequence of green area images.

[0071] Finally, after occlusion processing, the system relocates each previously marked plant in consecutive image frames and assigns them a unique identifier. This ensures that throughout the monitoring process, even in occluded image frames, the system can still perform refined health analysis based on consecutive image sequences of the same plant. This process lays the foundation for subsequent color analysis, morphological analysis, and contextual analysis, enabling the system to more accurately determine the health status of the vegetation.

[0072] Please see Figure 3The system utilizes video streams generated by moving vehicles to continuously observe the same plant from multiple angles. When pedestrians or vehicles temporarily obscure the green belt in the footage, the system uses the preceding and following frame information—that is, the time-series context—to predict and complete the mask of the obscured portion, or automatically discards severely obscured "inferior frames," retaining only "high-quality frames" with clear green areas, thus identifying and delineating all "green areas." Simultaneously, a multi-target tracking algorithm is used to lock onto the same tree (assigning it a unique temporary ID) within dozens of consecutive frames of the moving vehicle footage, ensuring that subsequent analysis is based on a continuous image sequence of that tree, rather than a single isolated image.

[0073] S130. Perform a comprehensive analysis of the image sequence of the green area in terms of color, posture, and context, identify and mark areas with abnormal color and posture of potential withered vegetation, and compare the state of individual plants with that of surrounding vegetation to obtain suspected withered targets; wherein, the posture includes the degree of drooping and curling of the vegetation.

[0074] In this embodiment, as Figure 4 As shown, suspected withered targets refer to green vegetation identified by the system as potentially having health problems after comprehensive analysis of color, posture, and context. These targets typically exhibit abnormal color (such as yellowish-brown), unhealthy morphology (such as drooping or curling), and significant differences compared to surrounding vegetation. To ensure accuracy and reduce false alarm rates, the system relies not only on the results of a single detection but also on historical data for temporal status confirmation to ultimately determine whether to generate an alert work order.

[0075] In one embodiment, step S130 described above may include steps S131 to S134.

[0076] S131. Calculate the RGB value of each pixel in the green area image sequence and apply the VARI formula to identify healthy vegetation and withered vegetation, and mark the color abnormality area.

[0077] In this embodiment, color aberration regions refer to those calculated using the Visible Atmospheric Impedance Index (VARI) formula, which is as follows: Here, IG, IR, and IB represent the normalized intensity values ​​of a pixel in the green, red, and blue channels, respectively. The calculated color values ​​are for areas where the values ​​are below a preset threshold (e.g., 0.4). Healthy vegetation typically appears green with a high VARI value, while withered vegetation may appear yellow or brown with a low VARI value. This analysis method can effectively identify areas of abnormal color, which may indicate withered vegetation.

[0078] S132. Perform texture and key point detection on the color abnormality area, calculate the degree of vegetation drooping and curling, and mark the posture abnormality area.

[0079] In this embodiment, the abnormal posture area refers to the vegetation parts that have significant drooping or curling features, as identified by texture analysis and key point detection technology.

[0080] In one embodiment, step S132 described above may include steps S1321 to S1322.

[0081] S1321. Conduct in-depth analysis of the color anomaly area and use texture analysis and key point detection technology to identify vegetation edges and tips.

[0082] More detailed texture analysis and key point detection are performed on the identified color anomaly areas to accurately determine the edges and tips of the vegetation. This information is crucial for subsequent calculations of drooping and curling.

[0083] Texture analysis identifies and classifies different types of surface features by analyzing the spatial distribution patterns between pixels in an image. In vegetation analysis, texture can reflect information about the health status of vegetation.

[0084] Texture features can be extracted by constructing a statistical matrix through calculating the grayscale value changes between adjacent pixels. For vegetation, healthy leaves usually have a relatively uniform texture, while withered or diseased leaves may exhibit irregular or rough textures.

[0085] A "pattern" for a pixel can also be formed by comparing each pixel with other pixels in its neighborhood and generating a binary number based on the result. This helps distinguish between healthy and unhealthy vegetation areas.

[0086] Keypoints are locations in an image that have significant features, such as corners, edges, or vertices of specific shapes. In vegetation analysis, keypoints are typically located at the edges of leaves, the ends of branches, and other similar locations.

[0087] S1322. By quantifying the angle of the key points at the tip relative to the direction of gravity, the degree of drooping and curling of the vegetation is measured, and areas of abnormal posture are marked.

[0088] By quantifying the angle of the key points at the tips relative to the direction of gravity, the degree of drooping and curling of the vegetation is measured, thereby determining the health status of the vegetation.

[0089] The system uses geometric principles to calculate the relative angles between key points. For each monitored plant, the system first determines the direction of its main stem as a reference line, and then calculates the angles between the key points at the tips and the reference line.

[0090] Transform the 2D coordinate system in the image into a coordinate system parallel to the ground, making the direction of gravity the Y-axis. This allows for direct calculation of angles using standard geometric formulas. For each key point at the tip, construct a vector pointing from the end of the trunk to that key point. Then, calculate the angle between this vector and the vertically downward direction (i.e., the direction of gravity) using the dot product or cross product formula. If the angle is close to 90 degrees, it indicates upright vegetation; if the angle is smaller, it indicates drooping vegetation.

[0091] The drooping index and the curling index are two indicators used to quantify the degree of abnormal vegetation posture.

[0092] Sag index: The average angle between the terminal key point and the direction of gravity, calculated using the method described above. A lower sag index indicates that the vegetation as a whole exhibits a sag trend.

[0093] Curl Index: In addition to considering the direction of the terminal key points, it is also necessary to examine their degree of curvature. This can be achieved by analyzing the curvature of the curve formed between the key points. A larger curl index indicates that the vegetation exhibits significant curling.

[0094] When the drooping index and curling index of a vegetation area both exceed a set threshold, that area is considered an abnormal posture zone. This multi-dimensional assessment method helps improve the accuracy and reliability of the system and reduce the false alarm rate.

[0095] Through the detailed analysis process described above, the system can not only accurately locate potentially problematic vegetation but also provide scientific evidence to support its health status assessment. This is of great significance for the refined management of urban greening.

[0096] S133. Identify the status of individual plants and adjacent vegetation, and compare and judge whether the greening plants are high-confidence individual abnormalities or regional characteristics.

[0097] In one embodiment, step S133 described above may include steps S1331 to S1332.

[0098] S1331. Use instance segmentation technology to identify the status of a single vegetation plant and its neighboring vegetation.

[0099] In this embodiment, instance segmentation is a computer vision technique that can not only classify each pixel like semantic segmentation, but also provide independent identifiers for different objects belonging to the same category (e.g., different vegetation individuals). This is particularly useful for analyzing vegetation in complex scenes.

[0100] By analyzing the color distribution of each separated plant, its health status can be preliminarily determined. Healthy vegetation is usually dark green, while withered or diseased vegetation may show abnormal hues such as yellow or brown.

[0101] Combining the previously mentioned key point detection technology, the angle of the terminal key point of each plant relative to the direction of gravity is calculated to obtain the droop index. A higher droop index may indicate a health problem in the plant.

[0102] S1332. When the similarity between the green plant and the surrounding healthy vegetation does not meet the requirements, the green plant is identified as a high-confidence individual anomaly; when the similarity between the green plant and the surrounding healthy vegetation meets the requirements, the green plant is identified as a regional characteristic.

[0103] The state similarity includes color similarity and droop index similarity.

[0104] Compare the color histograms of the target vegetation with those of other surrounding vegetation. If the differences between the two are significant (e.g., the target vegetation is noticeably yellowish), the color similarity is considered low.

[0105] Compare the droop index of the target vegetation with that of the surrounding vegetation. If the droop index of the target vegetation is much higher than that of the surrounding vegetation, it indicates that there may be growth problems.

[0106] When the color similarity and drooping index of a green plant do not meet the standards of the surrounding healthy vegetation, it indicates that the plant may be in an unhealthy state and is therefore marked as a high-confidence individual anomaly. In this case, the system will issue an early warning signal to alert management personnel and prompt them to take appropriate measures.

[0107] Conversely, if the color and droop index of the target vegetation match those of the surrounding vegetation, even if these characteristics deviate from the ideal "healthy" standard (such as autumn maple leaf color change), they are considered normal seasonal variations or regional features and will not trigger anomaly alerts. This is to avoid misjudging natural phenomena as diseases or environmental stress.

[0108] In parks or street greening, this system can help quickly identify individual trees that are declining due to pests, diseases or other factors, allowing for timely intervention and treatment to prevent the problem from spreading.

[0109] In summary, by using instance segmentation technology and multi-dimensional state similarity assessment, we can not only improve the accuracy of identifying the health status of individual vegetation plants, but also understand the behavioral patterns of vegetation groups on a larger scale, thereby better serving the actual needs of ecological and environmental protection.

[0110] Specifically, such as Figure 4 As shown, by using instance segmentation technology, the individual vegetation (A) belonging to the "abnormal posture area" and its adjacent vegetation (B, C, D) are identified in the green area.

[0111] Individual lesions: If A is yellowish-brown and drooping, while B, C, and D are green and upright, the system determines A to be a "high-confidence individual abnormality".

[0112] Regional consistency: If the states of A, B, C, and D are similar (for example, in autumn, A, B, C, and D are all red / yellow maple trees), the system determines that this is a "regional feature" and will reduce the weight of its withering warning, thereby avoiding seasonal false alarms.

[0113] S134. When the green plant is a high-confidence individual abnormality and has abnormal color and abnormal posture areas, the green plant is determined to be a suspected withering target.

[0114] This means that only when a plant simultaneously meets the criteria of abnormal color, abnormal posture, and a significant difference in condition from surrounding vegetation will it be ultimately identified by the system as a suspected withering target. This multi-dimensional analysis method not only improves detection accuracy but also effectively reduces false alarm rates, ensuring efficient resource utilization and timely response.

[0115] S140. Based on the GPS data, analyze the changes in color and posture through historical data to confirm the vegetation withering status of the suspected withered target.

[0116] In this embodiment, vegetation withering refers to the phenomenon of confirming whether a specific vegetation area has experienced non-seasonal deterioration and ultimately determining its withering state by analyzing multi-dimensional health status (including color, posture, and contextual information) and its changing trends over time.

[0117] In one embodiment, step S140 described above may include steps S141 to S142.

[0118] S141. Use the data packet of the suspected withered target to match the corresponding location inspection data in the historical database, and compare the multidimensional health status at different time points; wherein, the data packet includes GPS data, corresponding green area images, and analysis evidence; the multidimensional health status includes color status and posture status.

[0119] In this embodiment, the data packet contains GPS coordinate information, the corresponding green area image (mask), and color and morphological analysis evidence obtained in the previous steps.

[0120] Using GPS coordinates as a key index, we searched historical databases for patrol records of the same location over a past period. These records also included multi-dimensional health assessment results, such as the Color Health Index (VARI) and Sagging Index.

[0121] Compare the current image with historical VARI values ​​to observe if there is a continuous downward trend. If the greenness of a vegetation area decreases significantly, it may indicate that its health condition is deteriorating.

[0122] By comparing the changes in the directional vectors of key points at the plant's tips over a continuous time period, we can analyze whether the vegetation exhibits increased drooping or curling. Healthy plants typically remain upright, while diseased plants will show obvious morphological changes.

[0123] S142. When the multidimensional health status of the suspected withered target shows a continuous deterioration in the regional health status and does not conform to the normal seasonal change pattern, the suspected withered target is confirmed as a withered target.

[0124] In this embodiment, if the color and posture indicators of a specific vegetation area show signs of deterioration (e.g., the VARI value gradually decreases and the drooping index increases) during several consecutive inspections, this indicates that there may be a real health problem in the area.

[0125] Given that some plants naturally change color or appearance during specific seasons (such as maple leaves turning red in autumn), the system needs to verify whether the current changes are normal seasonal fluctuations based on known plant physiological models. If not, this further supports the hypothesis that the vegetation is in an abnormal state.

[0126] Once it is determined that the health status of a certain vegetation area has indeed deteriorated non-seasonally, and this deterioration trend spans multiple monitoring periods, the system officially marks it as a "wilt target".

[0127] For each vegetation identified as withered, the system automatically generates a detailed early warning work order, including the location of the vegetation (GPS coordinates), a screenshot of the site, multi-dimensional health analysis results (color / morphology / context), and the specific time of confirmation. This work order is then sent to the relevant urban greening management department so that appropriate maintenance measures can be taken in a timely manner.

[0128] Daily patrols using cameras mounted on municipal vehicles simultaneously collect video streams and GPS data, ensuring broad data coverage and real-time performance. Advanced AI algorithms process the video streams to identify and track each tree in the greenbelt, accurately capturing its health status even in obstructed conditions. Combining color, morphology, and contextual analysis, the health status of each plant is comprehensively assessed, improving detection accuracy. By integrating data from multiple patrols with geographic location information, the true health status of vegetation is accurately determined, avoiding false alarms.

[0129] This method not only improves the management efficiency of urban greening, but also effectively reduces labor costs and enhances the ability to maintain the urban ecological environment and landscape quality.

[0130] In this embodiment, as Figure 5 As shown, GPS coordinates and visual odometry technology are used to accurately match the location with historical inspection data in the database (e.g., the previous month's inspection record). The multidimensional health status (color, morphology, etc.) of the location at different time points is compared. If the health status of the area (e.g., low VARI, high drooping index) continues to deteriorate in 2-3 consecutive inspections, and the database shows that this does not conform to the normal seasonal physiological model of this tree species (e.g., it should not yellow or droop in spring and summer), the system ultimately identifies it as a "wilt target." If this change is consistent with the seasonal patterns in historical data, it is marked as a "seasonal change," and no warning is generated.

[0131] Furthermore, in another embodiment, the above method further includes:

[0132] An early warning work order is generated and pushed to the terminal. The early warning work order includes GPS data, on-site screenshots, multi-dimensional analysis evidence, and confirmation time.

[0133] In this embodiment, when the system identifies vegetation in a specific area as a "withering target" through multi-dimensional analysis (color, shape, context) and comparison with historical data, it will automatically generate an early warning work order. This work order contains the following key information:

[0134] GPS data: Records the specific geographical location of vegetation, making it easier for maintenance personnel to quickly and accurately locate problem areas.

[0135] On-site screenshots: Provide high-definition images or video frames to show the current state of the vegetation, helping maintenance personnel to intuitively understand the severity of the problem.

[0136] Multidimensional analysis of evidence:

[0137] Color analysis: Displays the color health status of vegetation, quantifying its health level through indices such as VARI.

[0138] Morphological analysis: describes abnormal vegetation postures, such as drooping index and curling index, to further confirm its health status.

[0139] Contextual analysis: By comparing the condition of the surrounding vegetation, it can be determined whether the disease is an individual lesion or a regional characteristic, thereby improving diagnostic accuracy.

[0140] Confirmation time: The point in time when the system determines that the vegetation is in a withered state, which helps to track the development of the problem.

[0141] Once an alert work order is generated, the system will automatically push it to the management terminals or mobile devices of the relevant maintenance departments. These terminals can be desktop computers, tablets, or smartphones, depending on the maintenance team's workflow and technical configuration. Push methods include, but are not limited to:

[0142] Email notification: Send the alert ticket directly to the designated email address.

[0143] In-app messaging: If the maintenance department uses a dedicated app to manage daily tasks, work orders can be displayed directly within the app.

[0144] SMS alerts: In case of emergencies, brief information and links can be sent via SMS to guide users to view the full work order details.

[0145] Once an early warning work order is received, maintenance personnel can take swift action based on the provided information, including but not limited to on-site inspections, pest and disease control, watering and fertilization, or other necessary maintenance measures. Furthermore, after the action is completed, the system can be updated through a feedback mechanism, forming a closed-loop management system and further enhancing the intelligent management level of urban greening.

[0146] This embodiment of the method, by introducing advanced AI visual perception technology and a dynamic patrol data acquisition system, significantly improves the accuracy of urban green vegetation detection and reduces the false alarm rate, while greatly increasing patrol efficiency and coverage. This system not only achieves refined and intelligent greening management but also effectively reduces operating costs and manpower input, thereby enhancing the urban ecological environment and landscape quality, and providing strong support for building a more livable and beautiful urban environment. This innovative solution, like a Korean drama, pays meticulous attention to every detail, ensuring that every tree is cared for, making the city's greenery vibrant and thriving.

[0147] The aforementioned method for detecting withered vegetation based on dynamic patrols integrates video streams collected by mobile vehicles and GPS data. This allows for automated and dynamic health monitoring of urban green areas. The system processes the video stream using continuous observation, occlusion compensation, and multi-target tracking technologies to generate a sequence of images of the green areas, including a temporal dimension. Then, it comprehensively analyzes the color, posture, and contextual information of the vegetation in these image sequences to identify and mark potential withering areas. By comparing the state changes of individual plants with surrounding vegetation, suspected withering targets are confirmed. Combined with GPS data and its historical records, the system can not only accurately locate and assess vegetation withering but also quickly issue early warning information, thereby achieving real-time monitoring and management of urban green vegetation health. This process fully demonstrates the system's intelligent and refined characteristics, providing a strong guarantee for improving urban ecological quality and landscape aesthetics.

[0148] Figure 6This is a schematic block diagram of a green plant wilting detection system 300 based on dynamic inspection, provided in an embodiment of the present invention. Figure 6 As shown, corresponding to the above-described method for detecting wilting of green plants based on dynamic inspection, the present invention also provides a system 300 for detecting wilting of green plants based on dynamic inspection. This system 300 includes a unit for executing the above-described method for detecting wilting of green plants based on dynamic inspection, and the system can be configured in a server. Specifically, please refer to... Figure 6 The green plant withering detection system 300 based on dynamic inspection includes an acquisition unit 301, an image sequence generation unit 302, a suspected target identification unit 303, and a withering status identification unit 304.

[0149] The acquisition unit 301 is used to acquire video streams and GPS data collected by mobile vehicles; the image sequence generation unit 302 is used to continuously observe, compensate for occlusion, and track multiple targets in the green areas of the video stream to obtain an image sequence of the green areas; the suspected target determination unit 303 is used to perform comprehensive analysis of the image sequence of the green areas in terms of color, posture, and context, identify and mark areas with abnormal color and posture of potential withered vegetation, and compare the state of a single plant with the surrounding vegetation to obtain suspected withered targets; wherein, the posture includes the degree of drooping and curling of the vegetation; the withering condition determination unit 304 is used to analyze the changes in color and posture state based on the GPS data and historical data to confirm the withering condition of the vegetation of the suspected withered targets.

[0150] In one embodiment, the image sequence generation unit 302 includes:

[0151] The system includes a tracking subunit for identifying and tracking the green areas in each frame of the video stream; a detection subunit for detecting whether the green areas in an image frame are occluded; a compensation subunit for compensating for the occluded portion or excluding image frames with the required degree of occlusion when the green areas in an image frame are occluded, in order to obtain an occlusion processing result; and a positioning subunit for locating the same plant in consecutive frames of the occlusion processing result and assigning it an identifier, in order to obtain a sequence of green area images.

[0152] In one embodiment, the compensation subunit is used to predict and complete the vegetation outline of the occluded part based on the information of the preceding and following image frames when the green area in the image frame is occluded, or to exclude image frames whose occlusion degree meets the requirements, so as to obtain the occlusion processing result.

[0153] In one embodiment, the suspected target determination unit 303 includes:

[0154] A color calculation subunit is used to calculate the RGB value of each pixel in the image sequence of the green area and apply the VARI formula to identify healthy and withered vegetation, and mark color aberration areas; a posture calculation subunit is used to perform texture and key point detection on the color aberration areas, calculate the degree of vegetation drooping and curling, and mark posture aberration areas; an individual identification subunit is used to identify the status of individual plants and adjacent vegetation, and compare and determine whether the green plant is a high-confidence individual aberration or a regional feature; a determination subunit is used to determine the green plant as a suspected withered target when the green plant is a high-confidence individual aberration and there are color aberration areas and posture aberration areas.

[0155] In one embodiment, the attitude calculation subunit includes:

[0156] The detection module is used to perform in-depth analysis of the color abnormality area, and to identify the vegetation edges and tips using texture analysis and key point detection technology; the measurement module is used to measure the degree of drooping and curling of the vegetation by quantifying the angle of the key points of the tips relative to the direction of gravity, and to mark the abnormal posture area.

[0157] In one embodiment, the individual identification subunit includes:

[0158] The segmentation module is used to identify the state of a single plant and its neighboring vegetation using instance segmentation technology; the identification module is used to identify the plant as a high-confidence individual anomaly when the similarity between the state of the green plant and the surrounding healthy vegetation does not meet the requirements; and to identify the green plant as a regional feature when the similarity between the state of the green plant and the surrounding healthy vegetation meets the requirements.

[0159] In one embodiment, the wilting condition determination unit 304 includes:

[0160] The comparison subunit is used to match the data packets of the suspected withered target with corresponding location patrol data in the historical database to compare the multidimensional health status at different time points; wherein, the data packets include GPS data, corresponding green area images, and analytical evidence; the multidimensional health status includes color status and posture status; the target confirmation subunit is used to confirm the suspected withered target as a withered target when the multidimensional health status of the suspected withered target shows that the regional health status is continuously deteriorating and does not conform to the normal seasonal change pattern.

[0161] In one embodiment, the system further includes:

[0162] The generation unit is used to generate an early warning work order and push it to the terminal. The early warning work order includes GPS data, on-site screenshots, multi-dimensional analysis evidence, and confirmation time.

[0163] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned green plant wilting detection system 300 based on dynamic inspection and each unit can be referred to the corresponding description in the aforementioned method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0164] The aforementioned green plant wilting detection system 300 based on dynamic inspection can be implemented as a computer program, which can be used in various ways, such as... Figure 7 It runs on the computer device shown.

[0165] Please see Figure 7 , Figure 7 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0166] See Figure 7 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0167] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a method for detecting wilting of green plants based on dynamic inspection.

[0168] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0169] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a method for detecting the withering of green plants based on dynamic inspection.

[0170] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0171] The processor 502 is used to run the computer program 5032 stored in the memory to implement all the steps of the dynamic inspection-based green plant wilting detection method.

[0172] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0173] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0174] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all the steps of the dynamic inspection-based green plant wilting detection method.

[0175] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0176] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0177] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0178] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0179] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0180] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for detecting wilting of green plants based on dynamic inspection, characterized in that, include: Acquire video streams and GPS data collected by mobile vehicles; The green areas in the video stream are continuously observed, occlusion compensated, and multi-target tracked to obtain a sequence of green area images; wherein, when the green area in an image frame is occluded, the vegetation outline of the occluded part is predicted and completed based on the information of the preceding and following image frames, or image frames with the required degree of occlusion are excluded. A comprehensive analysis of the image sequence of the green area is performed on color, pose, and context dimensions to identify and mark areas of color and pose abnormalities in potential withered vegetation, and to compare the state of individual plants with that of surrounding vegetation to obtain suspected withered targets; wherein, the pose includes the degree of drooping and curling of the vegetation. Based on the GPS data, the changes in color and posture are analyzed through historical data to confirm the vegetation withering status of the suspected withered target; The process involves a comprehensive analysis of the image sequence of the green area, considering color, pose, and context, to identify and mark areas of color and pose abnormalities in potentially withered vegetation. Individual plants are then compared with the surrounding vegetation to identify suspected withered targets. Calculate the RGB value of each pixel in the image sequence of the green area and apply the VARI formula to identify healthy vegetation and withered vegetation, and mark areas with abnormal colors; Texture and key point detection are performed on the color anomaly areas to calculate the degree of vegetation drooping and curling, and the posture anomaly areas are marked. The system identifies the state of individual plants and adjacent vegetation, comparing and determining whether each plant is a high-confidence individual anomaly or a regional feature. Specifically, instance segmentation technology is used to identify the state of individual plants and their adjacent vegetation. When the similarity between a plant and its surrounding healthy vegetation does not meet the requirements, the plant is identified as a high-confidence individual anomaly; when the similarity between a plant and its surrounding healthy vegetation meets the requirements, the plant is identified as a regional feature. When the green plant is a high-confidence individual anomaly and has areas of abnormal color and posture, the green plant is identified as a suspected withering target.

2. The method for detecting wilting of green plants based on dynamic inspection according to claim 1, characterized in that, The step of continuously observing, compensating for occlusion, and tracking multiple targets in the green areas of the video stream to obtain a sequence of green area images includes: Identify and track the green areas in each frame of the video stream; Detect whether the green areas in the image frame are occluded; When the green area in an image frame is occluded, compensate for the occluded part or exclude image frames whose occlusion level meets the requirements to obtain the occlusion processing result; The same plant is located in consecutive frames of the occlusion processing result and assigned an identifier to obtain a sequence of green area images.

3. The method for detecting wilting of green plants based on dynamic inspection according to claim 1, characterized in that, The process of performing texture and keypoint detection on the color anomaly areas, calculating the degree of vegetation drooping and curling, and marking the posture anomaly areas includes: The color anomaly areas were analyzed in depth, and texture analysis and key point detection techniques were used to identify vegetation edges and tips. By quantifying the angle of key points at the tips relative to the direction of gravity, the degree of drooping and curling of vegetation is measured, and areas of abnormal posture are marked.

4. The method for detecting wilting of green plants based on dynamic inspection according to claim 1, characterized in that, The state similarity includes color similarity and droop index similarity.

5. The method for detecting wilting of green plants based on dynamic inspection according to claim 1, characterized in that, The process of confirming the vegetation withering status of the suspected withering target based on the GPS data, combined with multiple anomaly indicators and historical data analysis, includes: The data packets of the suspected withered targets are matched with the corresponding location patrol data in the historical database to compare the multidimensional health status at different time points; wherein, the data packets include GPS data, corresponding green area images, and analytical evidence; the multidimensional health status includes color status and posture status; If the multidimensional health status of the suspected withered target shows a continuous deterioration in the regional health status and does not conform to the normal seasonal change pattern, then the suspected withered target will be confirmed as a withered target.

6. The method for detecting wilting of green plants based on dynamic inspection according to claim 1, characterized in that, After confirming the vegetation withering status of the suspected withering target based on the GPS data, combined with multiple anomaly indicators and through historical data analysis, the process includes: An early warning work order is generated and pushed to the terminal. The early warning work order includes GPS data, on-site screenshots, multi-dimensional analysis evidence, and confirmation time.

7. A greening plant wilting detection system based on dynamic inspection, characterized in that, The system uses the dynamic inspection-based method for detecting wilting of green plants as described in any one of claims 1 to 6, including: The acquisition unit is used to acquire video streams and GPS data collected by the mobile vehicle. An image sequence generation unit is used to continuously observe, compensate for occlusion, and track multiple targets in the green area of ​​the video stream to obtain an image sequence of the green area. The suspected target identification unit is used to perform a comprehensive analysis of the image sequence of the green area in terms of color, posture, and context, identify and mark areas of color and posture abnormalities of potential withered vegetation, and compare the state of a single plant with the surrounding vegetation to obtain suspected withered targets; wherein, the posture includes the degree of drooping and curling of the vegetation. The withering condition determination unit is used to analyze changes in color and posture based on the GPS data and historical data to confirm the withering condition of the vegetation of the suspected withering target.

Images