Landscaping maintenance intelligent monitoring system based on artificial intelligence
By combining unmanned vehicles with deep learning and image classification technology, the status of garden plants and facilities can be automatically detected, which solves the shortcomings of human judgment in garden greening maintenance and achieves efficient and accurate maintenance and management.
Patent Information
- Application Number
- CN202410355172.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-27
- Publication Date
- 2024-11-26
AI Technical Summary
In existing technologies, plant pruning and replacement, as well as the detection of damage to public facilities in garden and green space maintenance, rely on human judgment and cannot be handled in a timely manner, which affects the aesthetic value of the garden and the comfort of visitors.
Unmanned vehicles are used for park patrols. Combining deep learning and image classification technologies, they can automatically detect the growth status of plants and the damage to public facilities. The unmanned technology allows for patrols during periods of low foot traffic, reducing disturbance to visitors.
It has improved the efficiency and accuracy of garden and green space maintenance, reduced manpower input, enabled timely detection and handling of abnormal targets, and enhanced the garden's aesthetic value and visitor experience.
Smart Images

Figure CN119027905A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of landscaping maintenance monitoring, and more specifically, to an artificial intelligence-based intelligent landscaping maintenance monitoring system. Background Art
[0002] The maintenance of landscaping is also multifaceted. According to the needs of garden design or the growth status of plants, the plants in the garden need to be replaced or trimmed to achieve better results. There are not only plants in the garden, but also public facilities such as public seats, street lights, rockery, etc. The inspection of public facilities in the garden can improve the comfort of tourists when viewing.
[0003] The document with the prior art publication number CN114568265A provides an artificial intelligence-based intelligent monitoring and management system for garden greening maintenance to monitor the garden from the perspective of plant irrigation. With the development of smart gardens and intelligent irrigation technologies, the irrigation method using artificial intelligence has become relatively mature. The document with the prior art publication number CN115393786A provides another artificial intelligence-based intelligent monitoring method and system for garden greening maintenance, which detects plant growth conditions through artificial intelligence, and further detects whether the plants have problems such as water shortage, fertilizer shortage, diseases and insect pests.
[0004] Gardens can not only improve the environment, but also their ornamental value cannot be ignored. Cutting or replacing plants that are not growing well in the garden can improve the ornamental value of the garden. However, judging the growth status of plants has always been subjective, and professionals are required to find them in the garden. When public facilities in the garden are damaged but not handled in time, the comfort of tourists will be greatly reduced. Timely detection of damaged public facilities also requires human inspection in the garden.
[0005] In view of this, we propose an intelligent monitoring system for landscaping maintenance based on artificial intelligence. Summary of the invention
[0006] 1. Technical issues that need to be resolved
[0007] The purpose of this application is to provide an intelligent monitoring system for garden greening maintenance based on artificial intelligence. Plant pruning and replacement, and whether public facilities are damaged are often overlooked in garden greening maintenance monitoring. This application uses unmanned vehicles to patrol in the garden, using unmanned driving technology to ensure safety while reducing manpower input, select patrol time based on traffic flow, reduce disturbance to tourists, and detect plant growth status and whether public facilities are damaged through deep learning.
[0008] 2. Technical solution
[0009] The technical solution of this application provides an intelligent monitoring system for landscaping maintenance based on artificial intelligence, including:
[0010] The patrol module uses an unmanned vehicle to patrol the garden, collect garden information, and provide data support for subsequent analysis.
[0011] A patrol control module selects patrol time according to the flow of people in the garden, selects a time point with less flow of people to patrol, and patrols along a planned route.
[0012] The plant detection module is set on the unmanned vehicle. Through the images of plants in the garden collected by the camera of the unmanned vehicle, image classification is used to detect whether the plants are alive and whether they need to be pruned or replaced.
[0013] The facility detection module is set on the unmanned vehicle. Through the images of the facilities in the garden collected by the camera of the unmanned vehicle, image classification is used to detect whether the facilities in the garden are damaged, such as seats, street lights, etc. in the garden.
[0014] The abnormality positioning module records the shooting location where the abnormality is detected and marks it in the garden plan.
[0015] Through the above technical solution, unmanned vehicles are used to automatically patrol the garden, and time periods with less traffic are automatically selected for garden patrols. Cameras are used to detect the garden conditions, detect whether plants in the garden need to be trimmed or replaced, detect whether public facilities in the garden are damaged, and mark the shooting points where abnormalities are detected on the garden plan, making it easier for staff to find and deal with abnormal targets.
[0016] As an optional solution of this application document, the patrol module includes: an unmanned vehicle, which is equipped with a laser radar, a camera, and a radar. There is one laser radar in front and one in the back, which are set on the roof of the unmanned vehicle. The camera is divided into a camera for environmental perception and a camera for detecting gardens. The radar is a millimeter wave radar. Through the coordination of the laser radar, the camera, and the radar, the unmanned vehicle has a strong environmental perception ability and can avoid obstacles when driving automatically along the specified path.
[0017] Through the above technical solution, unmanned vehicles are used to patrol the garden, and environmental perception methods are used to avoid obstacles. For example, when encountering people or large obstacles on the patrol road, they can bypass and avoid them. Using unmanned vehicles to patrol and monitor the garden can be unaffected by weather and time, improve efficiency, reduce manpower investment, and improve detection accuracy.
[0018] As an optional solution of this application document, the patrol control module includes: a counter, which is set at each entrance and exit of the garden to count the flow of people in the garden; a flow prediction unit, which uses historical flow data to train a recurrent neural network to predict the flow of people in the future. A moment selection unit, when an unmanned vehicle is needed to perform a patrol mission, the unmanned vehicle will wait for a time with less traffic to patrol the garden. A path unit, the path unit will plan the patrol path of the unmanned vehicle and drive along the garden path to ensure comprehensive and complete monitoring.
[0019] Through the above technical solution, patrols are carried out when there is less traffic, which reduces disturbance to tourists who are viewing the gardens, ensures the safety of patrols, reduces the frequency of tourists obstructing patrols, and improves the efficiency of patrols.
[0020] As an optional solution of the present application document, the plant detection module includes: a plant detection training unit, which uses the collected plant images to mark the three categories of need to be replaced, need to be cropped, and need not be operated, and trains the image classification model to obtain an image classification model that can classify the plant status. A plant detection unit, which uses the images collected by the camera collected by the unmanned vehicle to input the trained image classification model to determine whether the plant needs to be cropped or replaced.
[0021] Through the above technical solution, image classification is used to determine whether plants in the garden need to be pruned or replaced, which improves the accuracy and efficiency of plant monitoring, reduces the time for staff to find and determine plants that need to be replaced or pruned, and enhances the maintenance efficiency of plants in the garden.
[0022] As an optional solution of the present application document, the facility detection module includes: a facility detection training unit, which uses the collected images of damaged and undamaged facilities placed in the garden to train an image classification model to obtain an image classification model that can detect whether the public facilities in the garden are damaged. A facility detection unit, which inputs the images of public facilities in the garden collected by the unmanned vehicle into the trained image classification model to detect whether the public facilities in the garden are damaged.
[0023] Through the above technical solution, image classification is used to determine whether the public facilities in the garden are damaged, so that the public facilities in the garden can be controlled in a timely manner. Ensuring the improvement of public facilities can improve the experience of tourists, reduce the investment in human monitoring, and improve the automation and intelligence of garden monitoring and management.
[0024] As an optional solution of the present application document, the abnormal positioning module includes: a positioning unit, which locates the real-time position of the unmanned vehicle through GPS, and a marking unit, which marks the shooting points where the abnormalities are detected in the garden plan, and the garden plan includes the path distribution of the garden.
[0025] Through the above technical solution, the combination of positioning and map marking is used to facilitate the staff to quickly and accurately find the location of abnormal targets, and facilitate the staff to deal with abnormal targets in a timely manner.
[0026] The present application also discloses a detection method of the aforementioned artificial intelligence-based landscaping maintenance intelligent monitoring system, comprising the following steps:
[0027] S1. Garden management personnel arrange patrol tasks for unmanned vehicles;
[0028] S2. The unmanned vehicle autonomously arranges patrol time according to the flow of people in the garden and the predicted flow of people in future time periods.
[0029] S3. The unmanned vehicle detects whether plants in the garden need to be replaced or trimmed, and detects whether public facilities are damaged;
[0030] S4. When an abnormal target is detected, the shooting location is recorded and marked on the garden plan.
[0031] Through the above technical solution, the unmanned vehicle automatically inspects the garden through the combination of unmanned driving technology and deep learning technology, which improves the inspection efficiency, reduces the workload of staff, and improves the maintenance effect of the garden.
[0032] 3. Beneficial effects
[0033] One or more technical solutions provided in the technical solution of this application have at least the following technical effects or advantages:
[0034] The use of unmanned driving technology can reduce the need for manual searches for poorly growing plants and damaged public facilities. Monitoring during low-traffic time periods combined with unmanned driving obstacle avoidance strategies can ensure the safety of monitoring. Training during low-traffic time periods can reduce the disturbance to tourists. Deep learning image classification can be used to determine whether plants need to be replaced or trimmed, and whether public facilities are damaged, thereby reducing the subjectivity of manual judgment and improving the accuracy of judgment. The locations of detected abnormal targets can be saved and marked on the garden plan, making it easier for staff to find targets accurately and quickly and take timely action. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 The overall structure diagram of the landscaping maintenance intelligent monitoring system based on artificial intelligence disclosed in a preferred embodiment of the present application;
[0036] Figure 2 This is a structural diagram of a patrol module of an artificial intelligence-based intelligent monitoring system for landscaping maintenance disclosed in a preferred embodiment of the present application;
[0037] Figure 3 This is a structural diagram of a patrol control module of an artificial intelligence-based intelligent monitoring system for landscaping maintenance disclosed in a preferred embodiment of the present application;
[0038] Figure 4 This is a structural diagram of a plant detection module of an artificial intelligence-based garden greening maintenance intelligent monitoring system disclosed in a preferred embodiment of the present application;
[0039] Figure 5 This is a structural diagram of a facility detection module of an artificial intelligence-based intelligent monitoring system for landscaping maintenance disclosed in a preferred embodiment of the present application. DETAILED DESCRIPTION
[0040] The present application is further described in detail below in conjunction with the accompanying drawings.
[0041] Reference Figure 1 The embodiment of the present application provides an artificial intelligence-based intelligent monitoring system for garden greening maintenance, including: a patrol module, which uses an unmanned vehicle to patrol the garden, collect garden information, and provide data support for subsequent analysis; a patrol control module, which selects the patrol time according to the flow of people in the garden, chooses to patrol at a time with less traffic, and patrols according to a planned path; a plant detection module, which is arranged on the unmanned vehicle, and uses image classification to detect whether the plants are alive and whether they need to be pruned or replaced through the images of plants in the garden collected by the camera of the unmanned vehicle; a facility detection module, which is arranged on the unmanned vehicle, and uses image classification to detect whether the facilities in the garden are damaged through the images of facilities in the garden collected by the camera of the unmanned vehicle, such as seats and street lights in the garden; an abnormal positioning module, which records the shooting location where the abnormality is detected and marks it in the garden plan.
[0042] The unmanned vehicle uses a combination of lidar, cameras, radar, etc. to perceive the environment, identify people or objects on the path, automatically avoid obstacles, use cameras to check whether plants need to be replaced or trimmed, and check whether public facilities in the garden are damaged. When it is detected that plants need to be replaced or trimmed, or public facilities are damaged, the shooting point is recorded and marked on the garden plan to facilitate subsequent processing by staff.
[0043] Reference Figure 1 and Figure 2The embodiment of the present application provides a patrol module of an intelligent monitoring system for landscaping maintenance based on artificial intelligence. The patrol module includes an unmanned vehicle, which includes sensors for environmental perception and object recognition. The unmanned vehicle uses an autonomous driving algorithm and a computing power platform to provide support for obstacle avoidance and behavioral decision-making of the unmanned vehicle. The unmanned vehicle can automatically avoid tourists and obstacles on the path when patrolling, and can provide timely feedback and notify management personnel when the path is impassable. The unmanned vehicle is also equipped with a camera for monitoring the garden to provide data support for subsequent judgments.
[0044] Reference Figure 1 and Figure 3 , the embodiment of the present application provides a patrol control module of an intelligent detection system for garden greening maintenance based on artificial intelligence, the patrol control module includes: a counter, which is set at each entrance and exit of the garden to count the real-time flow of people in the garden; a flow prediction unit, which uses the hourly flow of people from 5:00 to 22:00 in the past three months to divide into a training set and a test set to train a long short-term memory model LSTM, the time step is set to 10, that is, the flow of people in the past 10 hours is used to predict the flow of people in the next hour, the LSTM unit is set to 100, and after training and evaluation, the LSTM model that can predict the hourly flow of people from 5:00 to 22:00 in the next day is obtained. The flow of people in 10 hours from 5:00 to 22:00 is obtained through the counter. The input model predicts the flow of people in the next hour, and the current input data and the flow of people in the next hour output by the model are used as new input data each time to predict the flow of people in the next hour until the data for one day is predicted. The unmanned vehicle compares the predicted flow of people with the set threshold to determine whether to perform a patrol task. Path unit, the garden management personnel can set a path for the unmanned vehicle through the path unit, so that the unmanned vehicle can better monitor the garden.
[0045] Reference Figure 1 and Figure 4 The present application embodiment provides a plant detection module of an artificial intelligence-based garden greening maintenance intelligent detection system, the plant detection module comprising: a plant detection training unit, using the collected images of various plants in the garden to be divided into three categories: plants that need to be replaced, plants that need to be cropped, and plants that do not need to be operated, making a data set and dividing it into a training set and a validation set, training a convolutional neural network model, using ResNet as the architecture of the convolutional neural network, using the softmax function as the classification function, the softmax function is a function that maps any real number to an interval, and the sum is 1, the softmax function formula is y k =exp / ∑ n i=1 exp,a k is the kth input signal in the output layer, y kis the kth output signal, and n is the number of neurons in the output layer. The cross entropy loss function is used as the loss function of the model, and the back propagation optimization model is used. The cross entropy loss function formula is L = -∑ n k=1 t k logy k , where t k is the true label of the kth category, y k is the predicted probability of the kth category, and n is the number of categories. Using ten-fold cross validation, the F1 value is evaluated to adjust the model's hyperparameters, and an image classification model is obtained that can classify different plant images into three categories: plants that need to be replaced, plants that need to be cropped, and plants that do not need to be operated; the plant detection unit inputs the plant images collected by the unmanned vehicle during patrol into the plant detection unit to obtain the classification of plants in the irradiated area.
[0046] Reference Figure 1 and Figure 5 , an embodiment of the present application provides a facility detection module of an artificial intelligence-based garden landscaping maintenance intelligent detection system, the facility detection module comprising: a facility detection training unit, which uses the collected intact and damaged public facility images to mark and divide them into a training set and a validation set to train a convolutional neural network, and also uses ResNet as the architecture of the convolutional neural network, and uses sigmoid as the classification function. The formula of sigmoid is f(x)=1 / (1+ex), and uses binary cross entropy loss to calculate the loss value. The formula of the binary cross entropy loss function is Loss=-∑Nyi·log(p(yi))+(1-yi)·log(1-p(yi)), where y is a binary label 0 or 1, and P is the probability that the output belongs to the y label. When the output value approaches 1, the loss value Loss approaches 0, otherwise it approaches infinity. The model is continuously optimized through the back-propagation algorithm, and the model's hyperparameters are adjusted using ten-fold cross validation and F1 value evaluation to obtain an image classification model that can detect whether public facilities in the garden are damaged. The facility detection unit inputs the public facility images collected by the unmanned vehicle into the trained image classification model to determine whether the public facilities in the image are damaged.
[0047] Reference Figure 1 The embodiment of the present application provides an abnormality positioning module of an intelligent detection system for garden greening maintenance based on artificial intelligence, and the abnormality positioning module includes a positioning unit, which is not only used for the automatic driving of unmanned vehicles, but also records the location information when the plant detection module or the facility detection module detects an abnormality. A marking unit marks the recorded shooting location in the plan view of the garden; the plan view of the garden is a plan view made according to the construction of the garden and the setting of the path.
[0048] The present application also discloses a detection method of the aforementioned artificial intelligence-based landscaping maintenance intelligent monitoring system, comprising the following steps:
[0049] S1. Garden management personnel arrange patrol tasks for unmanned vehicles;
[0050] S2. The unmanned vehicle autonomously arranges patrol time according to the flow of people in the garden and the predicted flow of people in future time periods.
[0051] S3. The unmanned vehicle detects whether plants in the garden need to be replaced or trimmed, and detects whether public facilities are damaged;
[0052] S4. When an abnormal target is detected, the shooting location is recorded and marked on the garden plan.
[0053] In summary, when the artificial intelligence-based intelligent monitoring system for garden landscaping maintenance disclosed in the embodiment of the present application is in use, the garden management personnel set patrol tasks, and the unmanned vehicle determines when to patrol by obtaining the flow of people in the garden. When patrolling, the unmanned vehicle will collect images of plants and public facilities in the garden, and input the images into the deep learning model to determine whether the plants need to be replaced or cropped, and whether the public facilities are damaged. When an abnormal target is detected, the location of the unmanned vehicle when taking the photo is recorded and marked on the garden's floor plan to facilitate subsequent processing by the staff.
Claims
1. An artificial intelligence-based intelligent monitoring system for landscaping maintenance, characterized by: include: Patrol module, which uses unmanned vehicles to patrol the garden, collect garden information, and provide data support for subsequent analysis; A patrol control module, wherein the patrol control module selects patrol time according to the flow of people in the garden, selects a time point with less flow of people to patrol, and patrols along a planned path; A facility detection module is provided on the unmanned vehicle, and uses image classification to detect whether the facilities in the garden are damaged, such as seats and street lights in the garden, based on the images of the facilities in the garden collected by the camera of the unmanned vehicle; The abnormality positioning module records the shooting location where the abnormality is detected and marks it in the garden plan.
2. According to claim 1, an artificial intelligence-based intelligent monitoring system for landscaping maintenance is characterized by: The patrol module includes: an unmanned vehicle, a camera, and a radar; The unmanned vehicle is equipped with a laser radar, one at the front and one at the back, which are arranged on the roof of the unmanned vehicle. The cameras are divided into cameras for environmental perception and cameras for detecting gardens.
3. The intelligent monitoring system for landscaping maintenance based on artificial intelligence according to claim 1 is characterized by: The patrol control module comprises: A counter is provided at each entrance and exit of the garden to count the flow of people in the garden; A crowd flow prediction unit, wherein the crowd flow prediction unit uses historical crowd flow data to train a recurrent neural network to predict the crowd flow in the future; Time selection unit: when the unmanned vehicle is needed to perform patrol tasks, the unmanned vehicle will wait for a time with less traffic to patrol the garden; The path unit plans the patrol path of the unmanned vehicle and drives along the garden path to ensure comprehensive and complete monitoring.
4. The artificial intelligence-based landscaping maintenance intelligent monitoring system according to claim 1 is characterized by: The facility detection module includes: A facility detection training unit, wherein the training unit uses the collected images of damaged and undamaged facilities placed in the garden to train an image classification model to obtain an image classification model that can detect whether public facilities in the garden are damaged; A facility detection unit, wherein the detection unit inputs the images of public facilities in the garden collected by the unmanned vehicle into a trained image classification model to detect whether the public facilities in the garden are damaged.
5. The intelligent monitoring system for landscaping maintenance based on artificial intelligence according to claim 1 is characterized by: The abnormality positioning module includes: a positioning unit, which locates the real-time position of the unmanned vehicle through GPS, and a marking unit, which marks the shooting points where the abnormality is detected in the garden plan, and the garden plan includes the path distribution of the garden.
6. The intelligent monitoring system for landscaping maintenance based on artificial intelligence according to claim 1 is characterized by: Also includes: The plant detection module is set on the unmanned vehicle. Through the images of plants in the garden collected by the camera of the unmanned vehicle, image classification is used to detect whether the plants are alive and whether they need to be pruned or replaced.
7. The artificial intelligence-based intelligent monitoring system for landscaping maintenance according to claim 6 is characterized by: The plant detection module comprises: A plant detection training unit, wherein the training unit uses the collected plant images to mark the three categories of needing to be replaced, needing to be cropped, and needing no operation, and trains an image classification model to obtain an image classification model that can classify plant states; A plant detection unit uses images collected by the camera of the unmanned vehicle to input a trained image classification model to determine whether the plant needs to be cropped or replaced.
8. The detection method of the landscaping maintenance intelligent monitoring system based on artificial intelligence according to claim 6 is characterized by: The following steps are involved: The garden management staff arranges patrol missions for the unmanned vehicles; The unmanned vehicle autonomously arranges patrol times according to the flow of people in the garden and the predicted flow of people in future time periods. Unmanned vehicles can detect whether plants in gardens need to be replaced or pruned, and whether public facilities are damaged; When an abnormal target is detected, the shooting location is recorded and marked on the garden plan.
9. The detection method of the landscaping maintenance intelligent monitoring system based on artificial intelligence according to claim 8 is characterized by: The abnormal targets are plants that need to be replaced or inspected, and damaged public facilities.
Citation Information
Patent Citations
State monitoring and maintaining Internet of Things system for community public facilities
CN110909092A
Road green belt trimming task automatic generation method based on neural network and CIM
CN111951166A
Method and system for individually maintaining plants according to user requirements and storage medium
CN114998875A
Landscaping maintenance intelligent monitoring management system based on artificial intelligence
CN116993038A
Big data management control method based on digital twin three-dimensional scene
CN117196210A
