Property management intelligent scheduling and maintenance method and system

Abstract

The invention discloses a property management intelligent scheduling and maintenance method, and relates to the field of intelligent property management and artificial intelligence application cross technology, and the method comprises the steps: collecting image data of a property management region, and carrying out the preprocessing of the image data based on a deep learning model, and obtaining a standardized property scene image; inputting the standardized property scene image into a pre-trained scene recognition model, extracting a scene feature vector, recognizing facility state information in a property area according to the scene feature vector, constructing a task priority scoring matrix, grading property maintenance tasks, and generating a task scheduling list; and allocating property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through the mobile terminal, and updating a task completion state in real time. According to the invention, a closed-loop management process from automatic problem sensing to task execution monitoring is constructed, intelligentization and refinement of property maintenance are realized, the management cost is reduced, and the service quality and the user satisfaction are improved.

Description

Technical Field

[0001] The present invention relates to the cross - technical field of intelligent property management and artificial intelligence applications, and particularly to an intelligent scheduling and maintenance method and system for property management. Background Art

[0002] Traditional property management models mainly rely on manual inspections, experience - based judgments, and manual scheduling, suffering from problems such as slow response, information blockage, low efficiency, and extensive management. In recent years, with the rapid development of Internet of Things, artificial intelligence, and mobile communication technologies, intelligent property management has gradually emerged. Through means such as sensor networks, video surveillance systems, and mobile terminals, visual, data - based, and mobile management of property areas has been realized. Among them, the facility monitoring technology based on image recognition has gradually been applied to actual property management, providing a technical foundation for realizing intelligent maintenance and scheduling.

[0003] However, most existing property image recognition technologies remain at the stage of simple image detection or target recognition, and fail to fully utilize deep - learning models to deeply extract multi - dimensional information in complex scenarios. At the same time, existing scheduling systems still rely heavily on preset rules and manual experience in task priority determination and personnel matching, lacking a data - driven dynamic decision - making mechanism. In addition, the current facility status recognition and subsequent maintenance scheduling processes are often disjointed, suffering from problems such as lagging data updates, unreasonable task allocation, and low resource utilization efficiency, making it difficult to achieve true closed - loop management. Especially in a property environment with parallel multi - tasks and complex and changeable scenarios, how to improve the accuracy of task recognition and the intelligence of scheduling execution has become a key technical difficulty that urgently needs to be solved in the intelligent property system. In view of the above problems, the present invention provides an intelligent scheduling and maintenance method and system for property management. Summary of the Invention

[0004] In view of the problems that existing property management technologies generally have a shallow understanding of scene information, lack of a priority - grading logic for task scheduling, and an imperfect closed - loop for maintenance task execution, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to achieve intelligent recognition of property scenes based on image data and deep - learning models, construct a task priority mechanism, and improve the execution efficiency and intelligence level of property maintenance tasks through a dynamic scheduling and real - time feedback mechanism.

[0006] To solve the above - mentioned technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a method for intelligent scheduling and maintenance of property management, which includes collecting image data of a property management area, preprocessing the image data based on a deep learning model to obtain a standardized property scene image; inputting the standardized property scene image into a pre-trained scene recognition model, extracting scene feature vectors, and identifying facility status information within the property area according to the scene feature vectors; constructing a task priority scoring matrix based on the facility status information, grading property maintenance tasks, and generating a task scheduling list; allocating property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

[0008] As a preferred solution of the method for intelligent scheduling and maintenance of property management according to the present invention, the steps of allocating property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time include: calling a property maintenance personnel database based on the task type in the task scheduling list, where the property maintenance personnel database includes professional skill tags, skill level scores, historical task completion rates, and current task loads of the maintenance personnel; using an improved Hungarian algorithm to perform an optimal match between maintenance tasks and maintenance personnel to generate a task matching plan; pushing the task matching plan to the corresponding maintenance personnel in real time through a mobile application terminal, and uploading the task execution progress in real time to update the task status in the task scheduling list.

[0009] As a preferred solution of the method for intelligent scheduling and maintenance of property management according to the present invention, the method for generating the task scheduling list is as follows: using the identified facility status information as input, constructing a task priority scoring matrix, and setting weight coefficients, where the task priority scoring matrix includes a failure level index, a facility importance index, and a maintenance timeliness index; multiplying the weight coefficients of the failure level index, the facility importance index, and the maintenance timeliness index to obtain a task comprehensive score; grading property maintenance tasks based on the task comprehensive score, and arranging the maintenance tasks in ascending order to form an initial task sequence; performing cluster analysis on the geographical locations of adjacent tasks in the initial task sequence, combining maintenance tasks with adjacent geographical locations into task groups, and generating a task scheduling list.

[0010] As a preferred solution of the intelligent scheduling and maintenance method for property management according to the present invention, it includes: grading the property maintenance tasks, including: when the comprehensive task score is greater than the first threshold, it is determined that this maintenance task is an urgent task and classified as the first-level priority; when the comprehensive task score is less than or equal to the first threshold and greater than the second threshold, it is determined that this maintenance task is an important task and classified as the second-level priority; when the comprehensive task score is less than or equal to the second threshold, it is determined that this maintenance task is an ordinary task and classified as the third-level priority.

[0011] As a preferred solution of the intelligent scheduling and maintenance method for property management according to the present invention, it includes: the method for identifying the facility status information is to input the standardized property scene image into the scene recognition model, where the scene recognition model adopts the Vision Transformer structure; through the image block processing module of the scene recognition model, the standardized property scene image is divided into a sequence of image blocks, and position encoding information is added to the image blocks; using the multi-head self-attention mechanism of the scene recognition model, calculate the correlation weight matrix between the image blocks, and extract the scene feature vector; based on the preset property facility status discrimination rule library, perform feature matching on the scene feature vector, and the rule library includes a facility intact feature template, a damage feature template, and a serious damage feature template; by calculating the cosine similarity between the scene feature vector and the feature template, determine the specific status level of the property facility and generate the facility status information, where the facility status information includes the facility type, location information, and status level.

[0012] As a preferred solution of the intelligent scheduling and maintenance method for property management according to the present invention, it includes: the method for standardizing the property scene image is to continuously collect image data through the monitoring camera array deployed in the property management area at a preset sampling frequency; perform noise elimination processing on the collected image data, filter the image noise through the Gaussian filtering algorithm, and use the histogram equalization method to adjust the image contrast and improve the image clarity; input the processed image data into the pre-trained deep learning model, where the deep learning model adopts the ResNet-50 network structure and is trained with property scene image samples; for the image data processed by the deep learning model, uniformly adjust the image size through normalization operation and complete the color space standardization of the RGB three channels to obtain the standardized property scene image.

[0013] In a second aspect, an embodiment of the present invention provides a property management intelligent scheduling and maintenance system, which includes: a preprocessing module for collecting image data of a property management area and preprocessing the image data based on a deep learning model to obtain a standardized property scene image; an identification module for inputting the standardized property scene image into a pre-trained scene recognition model, extracting scene feature vectors, and identifying the facility status information within the property area according to the scene feature vectors; a generation module for constructing a task priority scoring matrix based on the facility status information, grading property maintenance tasks, and generating a task scheduling list; and an assignment module for assigning property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

[0014] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program instructions are executed by the processor, the steps of the property management intelligent scheduling and maintenance method as described in the first aspect of the present invention are implemented.

[0015] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program instructions are executed by the processor, the steps of the property management intelligent scheduling and maintenance method as described in the first aspect of the present invention are implemented.

[0016] The beneficial effects of the present invention are as follows: By collecting and preprocessing the image data of the property area to obtain a standardized scene image, and then accurately identifying the facility status information using a pre-trained scene recognition model, the subjectivity and limited coverage of traditional manual inspections are effectively solved; Based on a multi-dimensional task priority scoring matrix, scientific grading and scheduling of maintenance tasks are carried out, and the optimal matching of personnel and tasks is achieved by combining the improved Hungarian algorithm, significantly improving the resource allocation efficiency; This method constructs a closed-loop management process from automatic problem perception to task execution monitoring, realizes the intelligence and refinement of property maintenance, reduces management costs, improves service quality and user satisfaction, and provides a feasible technical path for the digital transformation of the property management industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts. Among them:

[0018] Figure 1 It is a flowchart of the property management intelligent scheduling and maintenance method for Embodiment 1. Detailed Implementation Modes

[0019] To make the above - mentioned objects, features, and advantages of the present invention more obvious and understandable, the following will describe in detail the specific implementation modes of the present invention with reference to the accompanying drawings of the specification.

[0020] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0021] Secondly, the so - called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation mode of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it an individual or selectively exclusive embodiment compared with other embodiments.

[0022] Embodiment 1

[0023] Referring to Figure 1 , which is the first embodiment of the present invention. This embodiment provides an intelligent scheduling and maintenance method for property management, specifically including the following steps as shown in Figure 1 :

[0024] S1: Collect image data of the property management area, and pre - process the image data based on a deep - learning model to obtain a standardized property - scene image;

[0025] S2: Input the standardized property - scene image into a pre - trained scene recognition model, extract scene feature vectors, and identify the facility status information within the property area according to the scene feature vectors;

[0026] S3: Based on the facility status information, construct a task - priority scoring matrix, classify the property maintenance tasks, and generate a task - scheduling list;

[0027] S4: Allocate property maintenance personnel according to the task - scheduling list, push task information to the maintenance personnel through a mobile terminal, and update the task - completion status in real time.

[0028] In the embodiment of the present application, the above - mentioned step S1 includes:

[0029] Specifically, continuously collect image data through a monitoring camera array deployed in the property management area at a preset sampling frequency;

[0030] It should be noted that the monitoring camera array includes fixed - view cameras in aisles, elevators, lobbies, parking lots, and public facility areas; the monitoring cameras continuously collect image data at a sampling frequency of 24 frames per second; the image data collected by the monitoring cameras is aggregated through the data receiving module of the property management data center; the data receiving module classifies and stores the image data according to the timestamp and camera number.

[0031] Furthermore, perform noise elimination processing on the collected image data, filter image noise through the Gaussian filtering algorithm, and use the histogram equalization method to adjust the image contrast and improve image clarity; input the processed image data into a pre - trained deep learning model, where the deep learning model adopts the ResNet - 50 network structure and is trained with property scene image samples.

[0032] Even further, for the image data processed by the deep learning model, uniformly adjust the image size through normalization operations and complete the color space standardization of the RGB three channels to obtain a standardized property scene image.

[0033] In the embodiment of the present application, the above - mentioned step S2 includes:

[0034] Specifically, input the standardized property scene image into a scene recognition model, where the scene recognition model adopts the Vision Transformer structure.

[0035] It should be noted that the scene recognition model includes an encoder layer, a linear projection layer, and a multi - layer perceptron, and each layer in the encoder layer is provided with 8 attention heads.

[0036] Furthermore, through the image chunk processing module of the scene recognition model, divide the standardized property scene image into a sequence of image chunks and add position encoding information to the image chunks; use the multi - head self - attention mechanism of the scene recognition model to calculate the correlation weight matrix between the image chunks and extract the scene feature vector.

[0037] It should be noted that the scene feature vector includes the spatial position and status feature information of the property facilities.

[0038] Even further, based on a preset property facility status discrimination rule library, perform feature matching on the scene feature vector. The rule library includes a facility intact feature template, a damaged feature template, and a severely damaged feature template.

[0039] Specifically, determine the specific status level of the property facilities by calculating the cosine similarity between the scene feature vector and the feature template, and generate facility status information, where the facility status information includes the facility type, location information, and status level.

[0040] Preferably, the specific formula for the cosine similarity between the scene feature vector and the feature template is as follows:

[0041]

[0042] where S(F, T) is the cosine similarity score, F is the scene feature vector, T is the state feature template vector, n is the dimension of the feature vector, w i is the weight coefficient of the i-th dimension feature, F i and T i are the i-th components of the feature vector and the template vector respectively;

[0043] It should be noted that when the cosine similarity score S(F, T) > 0.8, the facility is in good condition, and the status level is recorded as normal in the facility status information; when the cosine similarity score 0.5 ≤ S(F, T) ≤ 0.8, the facility is in a damaged state, and the status level is recorded as damaged in the facility status information; when the cosine similarity score S(F, T) < 0.5, the facility is in a severely damaged state, and the status level is recorded as severely damaged in the facility status information.

[0044] Furthermore, if the deviation between the spatial position information of the scene feature vector and the preset position of the facility exceeds the preset threshold, it is determined that the facility position is abnormal, and the system records the abnormal information and marks it for manual review; if the status level of the same facility drops continuously three times in the detection results, it is determined that the facility status is deteriorating continuously, and the system automatically increases the detection frequency of the facility; if a certain dimension feature value of the scene feature vector exceeds the normal range, it is determined as abnormal data, and the image data is re-collected and analyzed; if the similarity difference between the scene feature vector and multiple templates is less than 0.1, it is determined that the status is ambiguous, and secondary feature extraction and matching verification are performed.

[0045] In the embodiment of the present application, the above step S3 includes:

[0046] Specifically, taking the identified facility status information as input, constructing a task priority scoring matrix, and setting weight coefficients, where the task priority scoring matrix includes a failure level index, a facility importance index, and a maintenance timeliness index;

[0047] It should be noted that the specific formula for the task priority scoring matrix is as follows:

[0048]

[0049] Among them, M is the task priority score, L is the fault level index value, with a value range of [1, 5], K is the facility importance index value, with a value range of [1, 3], H is the maintenance timeliness index value, with a value range of [1, 4], N is the geographical location clustering coefficient, with a value range of [0, 1], λ is the fault level weight coefficient, μ is the facility importance weight coefficient, η is the maintenance timeliness weight coefficient, and λ + μ + η = 1, σ is the geographical location influence coefficient, with a value range of [0, 0.2].

[0050] Preferably, the fault level index is divided into five levels based on the degree of abnormality in the facility status information, and weight coefficients 5, 4, 3, 2, and 1 are assigned respectively; the facility importance index is determined according to the location and usage frequency of the facility. The core facility is assigned a weight coefficient of 1.5, and the general facility is assigned a weight coefficient of 0.1; the maintenance timeliness index is calculated based on the historical data of the facility operation status. For facilities with frequent abnormalities, a weight coefficient of 2.5 is assigned, and for facilities with abnormalities for the first time, a weight coefficient of 0.01 is assigned;

[0051] Furthermore, multiply the weight coefficients of the fault level index, the facility importance index, and the maintenance timeliness index to obtain the task comprehensive score; based on the task comprehensive score, classify the property maintenance tasks and arrange the maintenance tasks in ascending order to form an initial task sequence;

[0052] Preferably, when the task comprehensive score is greater than the first threshold, it is determined that this maintenance task is an urgent task and classified as the first-level priority; when the task comprehensive score is less than or equal to the first threshold and greater than the second threshold, it is determined that this maintenance task is an important task and classified as the second-level priority; when the task comprehensive score is less than or equal to the second threshold, it is determined that this maintenance task is an ordinary task and classified as the third-level priority;

[0053] It should be noted that the first threshold is the lower limit of the score based on the task comprehensive score where the fault level is severely damaged and the facility importance is the core facility; the second threshold is the average value of the score based on the task comprehensive score where the fault level is generally damaged and the facility importance is the general facility; the third threshold is the upper limit of the score based on the task comprehensive score where the fault level is slightly damaged and the maintenance timeliness is low.

[0054] Even further, perform a clustering analysis on the geographical locations of adjacent tasks in the initial task sequence, combine the maintenance tasks with adjacent geographical locations into task groups, and generate a task scheduling list;

[0055] It should be noted that as shown in Table 1, the task scheduling list includes the task number, task priority, facility location, fault type, and estimated maintenance duration.

[0056] Table 1. Task Scheduling List

[0057] Task Number Task Priority Facility Location Fault Type Estimated Repair Duration T001 Level 1 Elevator on the West Side of the Lobby Severely Damaged 90 Minutes T002 Level 1 Area A of the Underground Parking Lot Position Deviation + Damage 60 Minutes T003 Level 2 Section C of the Corridor Damaged 40 Minutes T004 Level 2 Flower Bed in the Public Facility Area Status Ambiguous, Re-inspection Required 30 Minutes T005 Level 2 Left Side of the Lobby Entrance Abnormal Data Re-collection 25 Minutes T006 Level 1 Entrance B of the Elevator Status Deteriorating Continuously 75 Minutes T007 Level 3 Section B of the Corridor Damaged 50 Minutes T008 Level 3 Entrance Area of the Parking Lot First Abnormality 30 Minutes

[0058] In the embodiment of the present application, the above step S4 includes:

[0059] Specifically, based on the task types in the task scheduling list, the property maintenance personnel database is called, where the property maintenance personnel database includes the professional skill tags, skill level scores, historical task completion rates, and current task loads of the maintenance personnel;

[0060] It should be noted that the status of the maintenance personnel includes idle, on task, and resting. The system preferentially selects maintenance personnel in the idle state for task assignment; if there are multiple idle maintenance personnel, the matching degree scores are calculated based on the historical task completion rates, professional ratings, and current locations of the maintenance personnel, and the maintenance personnel with the highest matching degree score are selected to execute the task.

[0061] Furthermore, an improved Hungarian algorithm is used to perform optimal matching between the maintenance tasks and the maintenance personnel to generate a task matching plan;

[0062] It should be noted that the task matching plan includes facility location information, facility status description, task priority identification, estimated completion time, and a list of required tools and materials; if the task priority in the task scheduling list is level one, the system preferentially matches maintenance personnel whose professional skill levels are higher than the set threshold value (such as 80 points) and whose historical completion rates are not less than 90%, and adds an option for two-person collaborative operation to the matching plan to ensure the timely completion of the task; if the task involves multiple skill requirements (such as elevator failure + positioning anomaly), the system determines whether the single-person skill tags cover all the requirements; if the single-person skill tags do not cover all the requirements, the system recommends forming a cross-skill group and matching them to execute the task collaboratively; if the matching degrees of all current idle personnel are lower than the set threshold value (such as 60 points), the backup maintenance personnel recruitment process is triggered to automatically query whether there are high-matching candidates among the personnel in the resting state; if high-matching candidates are found among the personnel in the resting state, a task optional reception notice is sent, and if no high-matching candidates are found among the personnel in the resting state, the administrator is prompted to perform manual intervention and deployment.

[0063] Preferably, if the current task load of a certain maintenance staff is close to the maximum limit (e.g., the task volume ratio > 90%), the matching priority is reduced to avoid overloading; if the matching scores of multiple maintenance staff are the same, the system further judges the distance between their current positions and the task location; if the gap is less than 1 km, the one closer to the task location is preferred to reduce the response time; if the estimated completion time after task matching exceeds 20% of the expected task completion time, the task risk level is prompted to increase, and an alternative matching plan is generated for the administrator to confirm; if a special tool and material list is involved after task matching, the tool and material management module is automatically linked to verify the required tool inventory and allocation status; if the inventory is insufficient, the task is marked as pending materials, and the task scheduling time is adjusted; if the maintenance staff does not confirm receiving the task within 10 minutes after receiving the task push, the task is automatically withdrawn and the matching process is re-executed.

[0064] Furthermore, the task matching plan is pushed to the corresponding maintenance staff in real time through the mobile application terminal, and the task execution progress is uploaded in real time to update the task status in the task scheduling list.

[0065] It should be noted that the mobile terminal is a property management platform installed on the maintenance staff's mobile phone; after the task assignment is completed, the matching plan is pushed to the designated maintenance staff through the property management platform; the property management platform integrates the electronic map function and can plan the optimal working route for the maintenance staff; after the maintenance staff confirms receiving the task through the property management APP, the task status is updated to in progress.

[0066] Specifically, for the part of task status update: after the maintenance staff arrives at the scene, task check-in is carried out through the property management platform, and the system records the task start time; during the maintenance process, the maintenance staff uploads on-site photos and maintenance records through the property management platform to form a task execution track; after the maintenance is completed, the maintenance staff fills in the maintenance result report, including fault cause analysis, treatment methods and used parts, and uploads the on-site photos after maintenance; after the property management system confirms that the maintenance quality is qualified, the task status is updated to completed.

[0067] Furthermore, data statistics are carried out on the completed tasks to calculate indicators such as task response time, maintenance time and maintenance quality score; the statistical data is used for the performance appraisal of maintenance staff and is also stored in the database as historical data for optimizing future task allocation strategies; a maintenance work analysis report is generated regularly, including facility failure frequency statistics, maintenance efficiency analysis and personnel workload distribution.

[0068] Embodiment 2

[0069] This is the second embodiment of the present invention. This embodiment also provides a property management intelligent scheduling and maintenance system, including:

[0070] A preprocessing module for collecting image data of a property management area and preprocessing the image data based on a deep learning model to obtain a standardized property scene image;

[0071] An identification module for inputting the standardized property scene image into a pre-trained scene recognition model, extracting scene feature vectors, and identifying the facility status information within the property area according to the scene feature vectors;

[0072] A generation module for constructing a task priority scoring matrix based on the facility status information, grading property maintenance tasks, and generating a task scheduling list;

[0073] An allocation module for allocating property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

[0074] This embodiment also provides a computer device applicable to the situation of the intelligent scheduling and maintenance method for property management, including a memory and a processor; the memory is used for storing computer-executable instructions, and the processor is used for executing the computer-executable instructions to implement the intelligent scheduling and maintenance method for property management as proposed in the above embodiment.

[0075] It should be noted that the technical solution of this intelligent scheduling and maintenance system for property management belongs to the same concept as the technical solution of the above intelligent scheduling and maintenance method for property management. For the details not described in detail in the technical solution of this intelligent scheduling and maintenance system for property management in this embodiment, reference can be made to the description of the technical solution of the above intelligent scheduling and maintenance method for property management.

[0076] The above-mentioned unit modules can be embedded in the processor in the computer device in hardware form or be independent of the processor, or can be stored in the memory in the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0077] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be achieved through WIFI, operator network, NFC (Near Field Communication) or other technologies. When the computer program is executed by the processor, it realizes a method for evaluating the maximum access capacity of new energy considering the vulnerability of the power grid. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0078] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, the following steps are implemented: collecting image data of a property management area, and preprocessing the image data based on a deep learning model to obtain a standardized property scene image; inputting the standardized property scene image into a pre-trained scene recognition model, extracting scene feature vectors, and identifying the facility status information in the property area according to the scene feature vectors; based on the facility status information, constructing a task priority scoring matrix, grading the property maintenance tasks, and generating a task scheduling list; allocating property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

[0079] In summary, the present invention effectively solves the problems of subjectivity and limited coverage of traditional manual inspections by collecting and preprocessing property area image data to obtain standardized scene images, and then accurately identifying facility status information using a pre-trained scene recognition model; scientifically grading and scheduling maintenance tasks based on a multi-dimensional task priority scoring matrix, and realizing the optimal matching of personnel and tasks by combining an improved Hungarian algorithm, significantly improving the resource allocation efficiency; this method constructs a closed-loop management process from automatic problem perception to task execution monitoring, realizes the intelligentization and refinement of property maintenance, reduces management costs, improves service quality and user satisfaction, and provides a feasible technical path for the digital transformation of the property management industry.

[0080] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An intelligent scheduling and maintenance method for property management, characterized in that: including collecting image data of the property management area and preprocessing the image data based on a deep learning model to obtain a standardized property scene image; inputting the standardized property scene image into a pre-trained scene recognition model, extracting scene feature vectors, and identifying the facility status information within the property area according to the scene feature vectors; constructing a task priority scoring matrix based on the facility status information, grading the property maintenance tasks, and generating a task scheduling list; assigning property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

2. The intelligent dispatching and maintenance method for property management according to claim 1, characterized in that: Assigning property maintenance personnel according to the task scheduling list, pushing task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time, including: calling the property maintenance personnel database based on the task type in the task scheduling list, where the property maintenance personnel database includes the professional skill tags, skill level scores, historical task completion rates, and current task loads of the maintenance personnel; using an improved Hungarian algorithm to perform an optimal match between the maintenance tasks and the maintenance personnel to generate a task matching plan; pushing the task matching plan to the corresponding maintenance personnel in real time through a mobile application terminal, and uploading the task execution progress in real time to update the task status in the task scheduling list.

3. The intelligent scheduling and maintenance method for property management according to claim 2, characterized in that: The method for generating the task scheduling list is as follows: taking the identified facility status information as input, constructing a task priority scoring matrix, and setting weight coefficients, where the task priority scoring matrix includes a fault level index, a facility importance index, and a maintenance timeliness index; multiplying the weight coefficients of the fault level index, the facility importance index, and the maintenance timeliness index to obtain a task comprehensive score; grading the property maintenance tasks based on the task comprehensive score, and arranging the maintenance tasks in ascending order to form an initial task sequence; performing clustering analysis on the geographical locations of adjacent tasks in the initial task sequence, combining the maintenance tasks with adjacent geographical locations into task groups, and generating a task scheduling list.

4. The intelligent scheduling and maintenance method for property management according to claim 3, characterized in that: The grading of the property maintenance tasks includes: when the task comprehensive score is greater than the first threshold, it is determined that this maintenance task is an urgent task and is classified as the first priority level; when the task comprehensive score is less than or equal to the first threshold and greater than the second threshold, it is determined that this maintenance task is an important task and is classified as the second priority level; when the task comprehensive score is less than or equal to the second threshold, it is determined that this maintenance task is a general task and is classified as the third priority level.

5. The intelligent scheduling and maintenance method for property management according to claim 3, characterized in that: The method for identifying the facility status information is as follows: inputting the standardized property scene image into a scene recognition model, where the scene recognition model adopts a Vision Transformer structure; dividing the standardized property scene image into a sequence of image patches through the image patch processing module of the scene recognition model, and adding position encoding information to the image patches; using the multi-head self-attention mechanism of the scene recognition model to calculate the correlation weight matrix between the image patches and extract scene feature vectors; Based on a preset property facility status discrimination rule library, perform feature matching on the scene feature vector. The rule library includes a facility intact feature template, a damage feature template, and a severe damage feature template; Determine the specific status level of the property facility by calculating the cosine similarity between the scene feature vector and the feature template, and generate facility status information, where the facility status information includes the facility type, location information, and status level.

6. The intelligent dispatching and maintenance method for property management according to claim 5, characterized in that: The method for normalizing the property scene image is as follows: Continuously collect image data at a preset sampling frequency through a monitoring camera array deployed in the property management area; Perform noise elimination processing on the collected image data, filter image noise through the Gaussian filtering algorithm, and use the histogram equalization method to adjust the image contrast and improve the image clarity; Input the processed image data into a pre-trained deep learning model, where the deep learning model adopts the ResNet-50 network structure and is trained with property scene image samples; For the image data processed by the deep learning model, uniformly adjust the image size through a normalization operation and complete the color space normalization of the RGB three channels to obtain a normalized property scene image.

7. A property management intelligent scheduling and maintenance system, based on the property management intelligent scheduling and maintenance method according to any one of claims 1 to 6, characterized in that: Including: A preprocessing module for collecting image data in the property management area and preprocessing the image data based on a deep learning model to obtain a normalized property scene image; An identification module for inputting the normalized property scene image into a pre-trained scene recognition model, extracting the scene feature vector, and identifying the facility status information in the property area according to the scene feature vector; A generation module for constructing a task priority scoring matrix based on the facility status information, grading the property maintenance tasks, and generating a task scheduling list; An allocation module for allocating property maintenance personnel according to the task scheduling list, pushing the task information to the maintenance personnel through a mobile terminal, and updating the task completion status in real time.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, the steps of the property management intelligent scheduling and maintenance method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the property management intelligent scheduling and maintenance method according to any one of claims 1 to 6 are implemented.

Images