An elevator intelligent control system and method based on artificial intelligence

Abstract

The application discloses an elevator intelligent control system and method based on artificial intelligence, relates to the technical field of intelligent control, and solves the technical problem that in the prior art, the control of an elevator depends on physical case operation, it is difficult to identify the environment in the elevator car, elevator failure may be caused due to improper operation of a passenger, and the normal operation of the elevator is affected; the application obtains an elevator image by collecting the image in the elevator through an image collection device; real-time log data of elevator operation is acquired; the elevator image is analyzed and identified to obtain an identification result; the operation mode of the elevator is determined based on the identification result; the elevator is controlled according to the operation mode of the elevator; the failure of the elevator is detected according to the real-time log data; the elevator is repaired based on the detection result, and the trapped personnel are rescued; the elevator can intelligently adjust the operation state according to the actual situation, and the safety and rationality of operation are improved.

Description

Technical Field

[0001] This application belongs to the field of intelligent control and relates to elevator intelligent control technology, specifically an elevator intelligent control system and method based on artificial intelligence. Background Technology

[0002] An elevator is a mechatronic device that uses an electric motor as its power source and transmits power through steel cables, hydraulic systems, or rack and pinion gears to efficiently and safely transport people or goods within a shaft with a vertical or inclined angle of less than 15 degrees. Its core function is to overcome spatial limitations and enhance the functionality and efficiency of buildings. By analyzing passenger flow, floor distribution, and time periods, elevator operation strategies are dynamically adjusted. Combined with IoT sensors and AI technology, passenger demand can be predicted in advance, and elevators can be proactively dispatched to target floors, reducing unnecessary stops. The intelligent system can automatically identify emergencies, triggering backup power, automatic leveling, or linkage with fire suppression systems, while simultaneously providing voice prompts to reassure passengers and shorten rescue time. Intelligent control of elevators can significantly improve operational efficiency, safety, user experience, and energy savings.

[0003] In existing technologies, intelligent control of elevators generally involves controlling the elevator's operation based on the number of people going up and down the stairs to achieve energy-saving effects. However, existing technologies rely on physical buttons for elevator control, making it difficult to identify the environment inside the elevator car. This can lead to elevator malfunctions due to improper operation by passengers, affecting the normal operation of the elevator. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems existing in the prior art; to this end, this application proposes an intelligent elevator control system and method based on artificial intelligence to solve the technical problem that the existing technology relies on the operation of physical keys to control the elevator, which makes it difficult to identify the environment inside the elevator car, and may cause elevator malfunctions due to improper operation by passengers, thus affecting the normal operation of the elevator.

[0005] To achieve the above objectives, the first aspect of this application provides an intelligent elevator control system based on artificial intelligence, including: an image acquisition module, an analysis and control module, and a fault rescue module; The image acquisition module is used to acquire images inside the elevator through an image acquisition device to obtain elevator images; and to acquire real-time log data of elevator operation. The analysis and control module is used to analyze and recognize elevator images to obtain recognition results; determine the elevator operating mode based on the recognition results; and control the elevator according to the elevator operating mode. The fault rescue module is used to detect elevator faults based on real-time log data; repair the elevator based on the detection results; and rescue trapped personnel.

[0006] Preferably, the analysis and recognition of the elevator image to obtain the recognition result includes: The process involves: retrieving elevator images; preprocessing the elevator images; using an edge detection algorithm to annotate the edges of the preprocessed elevator images to obtain the processed images; and integrating the processed images within a set time period into an image analysis sequence according to the acquisition time. The preprocessing includes grayscale conversion, image enhancement, and illumination adjustment. The image recognition model is invoked; the image analysis sequence is input into the image recognition model to obtain the corresponding recognition label; the recognition label is set to a positive integer; the image recognition model is built based on an artificial intelligence model.

[0007] Preferably, the image recognition model is built based on an artificial intelligence model, including: Obtain the standard dataset; the standard dataset includes standard input data consistent with the content attributes of the image analysis sequence; and standard output data consistent with the content attributes of the recognition labels; Select model frameworks and deep learning algorithms from an artificial intelligence library; construct models based on deep learning algorithms and model frameworks to obtain deep learning models; The standard dataset is divided into a training set, a validation set, and a test set according to a set ratio; the deep learning model is trained using the training set; the internal parameters of the deep learning model are adjusted using the validation set; and the trained deep learning model is tested using the test set to obtain test metrics. Obtain the indicator threshold; compare the test indicator with the indicator threshold; if all test indicators are greater than the indicator threshold, then mark the deep learning model as an image recognition model; otherwise, rebuild and retrain the image recognition model.

[0008] It should be noted that the test metrics include: accuracy, recall, F1 score, and stability; the thresholds for the metrics and the division ratio of the standard dataset are set by professional technicians according to the actual situation; generally, the division ratio of the standard dataset is set to 8:1:1; when it is necessary to rebuild and retrain the image recognition model, the deep learning model is rebuilt and retrained or the division ratio of the standard dataset is changed until the test metrics meet the standards.

[0009] Preferably, determining the elevator operating mode based on the recognition result includes: Retrieve the identification tags; Obtain the tag runtime library; The elevator operation mode is obtained by matching the identification tags with the tag runtime library; the elevator operation mode includes: personnel identification mode and personnel riding mode.

[0010] Preferably, controlling the elevator according to the elevator operating mode includes: When the elevator is in human recognition mode and the recognition tag type is electric vehicle entering the elevator, an electric vehicle alarm voice will be generated, and the elevator image will be continuously recognized until no electric vehicle is recognized in the elevator image. When the elevator is in passenger mode and the identification tag type is abnormal behavior, it is determined whether the number of people is greater than the number of registered floors; if yes, a voice alarm signal is issued and the abnormal floor is deregistered; if no, the behavior of obstructing the elevator door is analyzed and a voice reminder signal is generated. When the elevator is in personnel identification mode and the identification tag type is "solo passenger", the control is based on the type of solo passenger. When the elevator is in the personnel identification model and the identification tag type is "personnel stuck", the elevator door will be opened and passengers will be reminded to leave the car or press the floor button.

[0011] Preferably, the control based on the type of individual passenger includes: Retrieve the corresponding elevator images; use image recognition technology to identify the types of people in the elevator images; the types of people include: children and women; When the passenger is a child, the elevator door will remain open, and a corresponding voice prompt will be generated. When the passenger is a woman, the elevator will proceed directly to the target floor, and the video inside the elevator will be recorded until the passenger leaves the elevator.

[0012] Preferably, the step of detecting elevator faults based on real-time log data includes: Acquire real-time log data of elevator operation; the real-time log data includes: sensor data, operation status log, and error code log; Obtain a preset standard range; compare the real-time log data with the standard range, and generate a fault alarm signal when the real-time log data exceeds the standard range.

[0013] Preferably, the elevator maintenance based on the test results includes: Retrieve real-time log data that exceeds the standard range and mark it as abnormal data; obtain the corresponding standard data; calculate the difference between the abnormal data and the standard data to obtain the deviation data; Obtain the deviation handling library; match the deviation data with the deviation handling library to obtain the corresponding fault handling measures, and send the fault handling measures to the corresponding technical personnel.

[0014] Preferably, the rescue of trapped personnel includes: Acquire elevator images when an elevator malfunctions; if people are trapped when an elevator malfunctions, generate an auxiliary voice prompt to prompt the trapped people to press the alarm button, and send the entrapment alarm information to the corresponding rescue personnel; When rescuers confirm that they have received the trapped alarm information, they will display the rescue information in real time on the smart screen. The rescue information includes: location coordinates, direction of movement and estimated arrival time.

[0015] The second aspect of this application provides an intelligent elevator control method based on artificial intelligence, comprising: Elevator images are obtained by capturing images inside the elevator using an image acquisition device; Obtain real-time log data of elevator operation; The elevator images were analyzed and identified to obtain the recognition results; Determine the elevator operating mode based on the identification results; Control the elevator according to its operating mode; Fault detection of elevators is performed based on real-time log data; Based on the test results, the elevator was repaired and the trapped people were rescued.

[0016] A third aspect of this application provides a computer-readable storage medium storing instructions for performing the method as described in the second aspect implementation.

[0017] Compared with the prior art, the beneficial effects of this application are: 1. This application effectively improves the quality of elevator images through preprocessing operations such as grayscale conversion, image enhancement, and illumination adjustment. Grayscale conversion reduces data volume while retaining key information; image enhancement highlights image details; illumination adjustment solves image blurring or insufficient contrast caused by lighting issues; edge detection algorithms are used to annotate the edges of the preprocessed images, clearly defining the contours of objects in the image, which helps to more accurately identify various objects and personnel behaviors inside the elevator, improving recognition accuracy; processed images within a set time period are integrated into an image analysis sequence according to the acquisition time, giving the image data temporal continuity, facilitating subsequent analysis of changes in the elevator's interior over time; an image recognition model is constructed based on an artificial intelligence model, utilizing the powerful feature extraction and pattern recognition capabilities of deep learning algorithms to automatically learn and recognize various features from images, achieving rapid and accurate recognition of elevator images; recognition tags are retrieved and matched with a tag runtime library to obtain elevator operating modes, such as personnel recognition mode and personnel riding mode. This operating mode matching method based on recognition results can accurately determine the operating mode according to the actual situation inside the elevator.

[0018] 2. This application acquires real-time log data of elevator operation, covering sensor data, operation status logs, and error code logs, among other information. It reflects the elevator's operating status from multiple perspectives, ensuring no detail that could potentially affect elevator safety is overlooked. By comparing real-time log data with a preset standard range, a fault alarm signal is immediately generated when the data exceeds the standard range. This allows for timely detection of potential elevator faults, preventing escalation and effectively reducing the risk of elevator accidents. Real-time log data exceeding the standard range is retrieved and marked as abnormal data, while the corresponding standard data is obtained, and the deviation between the two is calculated. This accurately quantifies the degree and characteristics of the fault, helping technicians to more accurately pinpoint the cause. A deviation processing library is acquired, and the deviation data is matched with it to obtain corresponding fault handling measures. The deviation processing library integrates a wealth of fault handling experience and professional knowledge, quickly providing technicians with scientific and reasonable handling suggestions, improving the efficiency and accuracy of fault handling, and reducing improper handling due to human error. When an elevator malfunction leads to personnel being trapped, an auxiliary voice prompt is generated to encourage the trapped personnel to press the alarm button, and the entrapment alarm information is sent to the corresponding rescue personnel. Auxiliary voice prompts can promptly soothe the emotions of trapped individuals, preventing them from acting inappropriately due to panic. Simultaneously, rapid and accurate alarm information transmission ensures that rescuers can quickly learn of the situation and initiate rescue procedures. Real-time display of rescuer information, including location coordinates, direction of movement, and estimated arrival time, is available on a smart screen. This real-time information sharing allows trapped individuals to clearly understand the rescue progress, increasing their confidence in being rescued. It also facilitates monitoring and coordination of the rescue process by management personnel, improving the transparency and efficiency of the rescue operation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram showing the connection of the system modules in this application; Figure 2 This is a schematic diagram illustrating the working steps of the analysis and control module in this application; Figure 3 This is a schematic diagram illustrating the working steps of the fault recovery module in this application; Figure 4 This is a schematic diagram of the method steps in this application. Detailed Implementation

[0021] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0022] Please see Figure 1 The first aspect of this application provides an intelligent elevator control system based on artificial intelligence, including: an image acquisition module, an analysis and control module, and a fault rescue module; The image acquisition module is used to acquire images inside the elevator through an image acquisition device to obtain elevator images; and to acquire real-time log data of elevator operation. The analysis and control module is used to analyze and recognize elevator images to obtain recognition results; determine the elevator operating mode based on the recognition results; and control the elevator according to the elevator operating mode. The fault rescue module is used to detect elevator faults based on real-time log data; repair the elevator based on the detection results; and rescue trapped personnel.

[0023] Based on the above system, images inside the elevator are captured by the image acquisition device, providing a clear view of the elevator's interior. Simultaneously, real-time elevator operation log data is acquired, recording various parameters and events during elevator operation. Real-time image and log data acquisition promptly reflects the elevator's current status, enabling management personnel or the system to quickly understand its operational situation. This provides a rich and comprehensive data foundation for subsequent analysis. Analysis and recognition of the acquired elevator images automatically determine various conditions inside the elevator, identifying the elevator's operating mode based on these results. This allows the elevator to intelligently adjust its operating status according to actual conditions, improving operational safety and rationality. Real-time log data is used to detect elevator malfunctions, promptly identifying potential or existing faults during operation, improving the efficiency and accuracy of fault detection. The intelligent control mode and rapid fault handling capabilities reduce the impact of elevator malfunctions on passengers, providing a safer, more comfortable, and convenient riding environment, enhancing user satisfaction and trust in the elevator system.

[0024] Please see Figure 2 The specific steps for analyzing the working principle of the control module are as follows: S201. Retrieve elevator images; preprocess the elevator images; use an edge detection algorithm to annotate the edges of the preprocessed elevator images to obtain processed images; integrate the processed images within a set time period into an image analysis sequence according to the acquisition time.

[0025] The preprocessing includes grayscale conversion, image enhancement, and illumination adjustment.

[0026] Example: An elevator camera captured 20 images over a continuous 30-second period during the morning rush hour. Preprocessing stage: Grayscale conversion: Converting an RGB image to a grayscale image reduces computational load; Image enhancement: Histogram equalization is used to improve contrast, making the outline of the electric vehicle clearer; Lighting adjustment: Corrects exposure issues in backlit scenes using adaptive gamma correction.

[0027] The Canny edge detection algorithm is used to annotate object edges, generating a sequence of processed images with edge annotations, which are then integrated into an image analysis sequence according to timestamps.

[0028] S202. Retrieve the image recognition model; input the image analysis sequence into the image recognition model to obtain the corresponding recognition label.

[0029] The identification labels are set to positive integers; the image recognition model is built based on an artificial intelligence model.

[0030] In one possible implementation, the image recognition model is built upon an artificial intelligence model, including: Obtain the standard dataset; the standard dataset includes standard input data consistent with the content attributes of the image analysis sequence; and standard output data consistent with the content attributes of the recognition labels; Select model frameworks and deep learning algorithms from an artificial intelligence library; construct models based on deep learning algorithms and model frameworks to obtain deep learning models; The standard dataset is divided into a training set, a validation set, and a test set according to a set ratio; the deep learning model is trained using the training set; the internal parameters of the deep learning model are adjusted using the validation set; and the trained deep learning model is tested using the test set to obtain test metrics. Obtain the indicator threshold; compare the test indicator with the indicator threshold; if all test indicators are greater than the indicator threshold, then mark the deep learning model as an image recognition model; otherwise, rebuild and retrain the image recognition model.

[0031] It should be noted that the test metrics include: accuracy, recall, F1 score, and stability; the thresholds for the metrics and the division ratio of the standard dataset are set by professional technicians according to the actual situation; generally, the division ratio of the standard dataset is set to 8:1:1; when it is necessary to rebuild and retrain the image recognition model, the deep learning model is rebuilt and retrained or the division ratio of the standard dataset is changed until the test metrics meet the standards.

[0032] Example: The image recognition model built on ResNet-50 was validated through the following process: Standard dataset: Contains 100,000 elevator scene images (80,000 for training, 10,000 for validation, and 10,000 for testing), labeled with tags such as electric vehicles, children, and abnormal behavior; Training configuration: Adam optimizer, learning rate 0.001, batch size 32; Test metrics: accuracy 98.2%, recall 97.5%, F1 score 97.8%, and stability after 200 consecutive tests 99.1%, all exceeding the thresholds (95% / 95% / 95% / 98%), indicating the model passed validation.

[0033] S203. Retrieve the identification tag; obtain the tag runtime library; match the identification tag with the tag runtime library to obtain the elevator operation mode.

[0034] The elevator operation modes include: personnel identification mode and personnel riding mode.

[0035] Example: When the identification tag is 1 and 2, the system enters the personnel identification mode; when the identification tag is 3 and 4, the system enters the personnel riding mode.

[0036] S204. When the elevator is in personnel recognition mode and the recognition tag type is electric vehicle entering the elevator, an electric vehicle alarm voice will be generated, and the elevator image will be continuously recognized until no electric vehicle is recognized in the elevator image.

[0037] S205. When the elevator is in passenger mode and the identification tag type is abnormal behavior, determine whether the number of passengers is greater than the number of registered floors; if yes, issue a voice alarm signal and cancel the registration of the abnormal floor; if no, analyze the behavior of obstructing the elevator door and generate a voice reminder signal.

[0038] S206. When the elevator is in personnel identification mode and the identification tag type is "single passenger", control is performed according to the type of person taking the elevator alone.

[0039] In one possible implementation, control is based on the type of person using the elevator individually, including: Retrieve the corresponding elevator images; use image recognition technology to identify the types of people in the elevator images; the types of people include: children and women; When the passenger is a child, the elevator door will remain open, and a corresponding voice prompt will be generated. When the passenger is a woman, the elevator will proceed directly to the target floor, and the video inside the elevator will be recorded until the passenger leaves the elevator.

[0040] S207. When the elevator is in the personnel identification model and the identification tag type is personnel staying, the elevator door is opened and passengers are reminded to leave the car or press the floor button.

[0041] Example: The system detects an electric vehicle entering the elevator, immediately plays the voice message "Electric vehicles are prohibited from entering the elevator," and continues to monitor until the electric vehicle leaves (takes 8 seconds). When there are 3 passengers and the registered floor is the 4th floor, all floor registrations will be cancelled and a message saying "Re-register floor" will be played. When someone blocks the elevator door for more than half a minute, a message saying "Please do not block the elevator door" will be played. When a child (10 years old) is identified as an individual, the elevator door remains open and a message saying "Please have a parent accompany the child" is played, while the monitoring center is notified. When an adult woman is identified as riding the elevator alone, the elevator will automatically go directly to the target floor (such as the 28th floor) and start recording the entire process (storing it to a dedicated encrypted hard drive). The system detected that three passengers had lingered in the elevator car for more than one minute. The system automatically opened the door and gave a voice prompt: "Please press the floor button or leave the elevator car."

[0042] Based on the above steps, preprocessing operations such as grayscale conversion, image enhancement, and illumination adjustment can effectively improve the quality of elevator images. Grayscale conversion reduces data volume while retaining key information; image enhancement highlights image details; illumination adjustment solves image blurring or insufficient contrast caused by lighting issues; edge detection algorithms are used to annotate the edges of the preprocessed images, clearly defining the contours of objects in the image, which helps to more accurately identify various objects and human behaviors inside the elevator, improving recognition accuracy; processed images within a set time period are integrated into an image analysis sequence according to the acquisition time, giving the image data temporal continuity, facilitating subsequent analysis of changes in the elevator's interior over time; an image recognition model is built based on an artificial intelligence model, utilizing the powerful feature extraction and pattern recognition capabilities of deep learning algorithms to automatically learn and recognize various features from images, achieving rapid and accurate recognition of elevator images; recognition tags are retrieved and matched with a tag runtime library to obtain elevator operating modes, such as personnel recognition mode and personnel riding mode. This operating mode matching method based on recognition results can accurately determine the operating mode according to the actual situation inside the elevator.

[0043] Please see Figure 3 The working principle and specific steps of the fault rescue module are as follows: S301. Obtain real-time log data of elevator operation; obtain a preset standard range; compare the real-time log data with the standard range, and generate a fault alarm signal when the real-time log data exceeds the standard range.

[0044] The real-time log data includes: sensor data, operational status logs, and error code logs.

[0045] S302. Retrieve real-time log data that exceeds the standard range and mark it as abnormal data; obtain the corresponding standard data; calculate the difference between the abnormal data and the standard data to obtain the deviation data.

[0046] S303. Obtain the deviation processing library; match the deviation data with the deviation processing library to obtain the corresponding fault handling measures, and send the fault handling measures to the corresponding technical personnel.

[0047] S304. Obtain elevator images when an elevator malfunctions; if people are trapped when an elevator malfunctions, generate an auxiliary voice prompt to the trapped people to press the alarm button, and send the trapped alarm information to the corresponding rescue personnel.

[0048] S305. When rescuers confirm that they have received the trapped alarm information, they will display the rescue information in real time on the smart screen.

[0049] The rescue information includes: location coordinates, direction of movement, and estimated arrival time.

[0050] Example: Real-time acquisition of door switch sensor data (normal range 0.5-2.0 seconds). In one instance, the door closing time was detected to be 3.2 seconds (exceeding the threshold). The deviation from the standard value (1.2 seconds) was calculated to be +2.0 seconds, and the "door machine failure" entry in the deviation processing library was matched.

[0051] If three people are detected inside the elevator car during a malfunction (confirmed by image recognition), the system will automatically trigger a entrapment alarm: voice guidance: "Please press the red alarm button," and at the same time send an alarm message to the rescue center containing the location of the elevator car (elevator A of building 3) and the estimated arrival time (5 minutes); Once the rescue is confirmed, the smart screen displays the rescuers' real-time location (via GPS), movement trajectory (from north to south), and estimated arrival time (updated to 4 minutes and 30 seconds).

[0052] Based on the above steps, real-time log data of elevator operation is acquired, covering sensor data, operation status logs, and error code logs, among other information. This reflects the elevator's operating status from multiple perspectives, ensuring no detail that could potentially affect elevator safety is overlooked. By comparing real-time log data with preset standard ranges, a fault alarm signal is immediately generated when the data exceeds the standard range. This allows for timely detection of potential elevator faults, preventing escalation and effectively reducing the risk of elevator accidents. Real-time log data exceeding the standard range is retrieved and marked as abnormal data, while the corresponding standard data is obtained, and the deviation between the two is calculated. This accurately quantifies the degree and characteristics of the fault, helping technicians to more accurately pinpoint the cause. A deviation processing library is acquired, and the deviation data is matched with it to obtain corresponding fault handling measures. The deviation processing library integrates a large amount of fault handling experience and professional knowledge, quickly providing technicians with scientific and reasonable handling suggestions, improving the efficiency and accuracy of fault handling, and reducing improper handling caused by human judgment errors. When an elevator malfunction leads to people being trapped, an auxiliary voice prompt is generated to prompt the trapped person to press the alarm button, and the entrapment alarm information is sent to the corresponding rescue personnel. Auxiliary voice prompts can promptly soothe the emotions of trapped individuals, preventing them from acting inappropriately due to panic. Simultaneously, rapid and accurate alarm information transmission ensures that rescuers can quickly learn of the situation and initiate rescue procedures. Real-time display of rescuer information, including location coordinates, direction of movement, and estimated arrival time, is available on a smart screen. This real-time information sharing allows trapped individuals to clearly understand the rescue progress, increasing their confidence in being rescued. It also facilitates monitoring and coordination of the rescue process by management personnel, improving the transparency and efficiency of the rescue operation.

[0053] Please see Figure 4 The second aspect of this application provides an intelligent elevator control method based on artificial intelligence, comprising: Elevator images are obtained by capturing images inside the elevator using an image acquisition device; Obtain real-time log data of elevator operation; The elevator images were analyzed and identified to obtain the recognition results; Determine the elevator operating mode based on the identification results; Control the elevator according to its operating mode; Fault detection of elevators is performed based on real-time log data; Based on the test results, the elevator was repaired and the trapped people were rescued.

[0054] A third aspect of this application provides a computer-readable storage medium storing instructions for performing the method as described in the second aspect implementation.

[0055] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0056] The working principle of this application is as follows: This application acquires images inside the elevator using an image acquisition device to obtain elevator images; obtains real-time log data of elevator operation; analyzes and identifies the elevator images to obtain identification results; determines the elevator operation mode based on the identification results; controls the elevator according to the elevator operation mode; detects elevator faults based on real-time log data; repairs the elevator based on the detection results; and rescues trapped personnel.

[0057] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. An intelligent elevator control system based on artificial intelligence, characterized in that, include: Image acquisition module, analysis and control module, and fault rescue module; The image acquisition module is used to acquire images inside the elevator through an image acquisition device to obtain elevator images; and to acquire real-time log data of elevator operation. The analysis and control module is used to analyze and recognize elevator images to obtain recognition results; Determine the elevator operating mode based on the identification results; Control the elevator according to its operating mode; The fault rescue module is used to detect elevator faults based on real-time log data; Based on the test results, the elevator was repaired and the trapped people were rescued.

2. The elevator intelligent control system based on artificial intelligence according to claim 1, characterized in that, The analysis and recognition of the elevator images to obtain the recognition results include: The process involves: retrieving elevator images; preprocessing the elevator images; using an edge detection algorithm to annotate the edges of the preprocessed elevator images to obtain the processed images; and integrating the processed images within a set time period into an image analysis sequence according to the acquisition time. The preprocessing includes grayscale conversion, image enhancement, and illumination adjustment. The image recognition model is invoked; the image analysis sequence is input into the image recognition model to obtain the corresponding recognition label; the recognition label is set to a positive integer; the image recognition model is built based on an artificial intelligence model.

3. The elevator intelligent control system based on artificial intelligence according to claim 2, characterized in that, The image recognition model is built based on an artificial intelligence model and includes: Obtain the standard dataset; the standard dataset includes standard input data consistent with the content attributes of the image analysis sequence; and standard output data consistent with the content attributes of the recognition labels; Select model frameworks and deep learning algorithms from an artificial intelligence library; construct models based on deep learning algorithms and model frameworks to obtain deep learning models; The standard dataset is divided into a training set, a validation set, and a test set according to a set ratio; the deep learning model is trained using the training set; the internal parameters of the deep learning model are adjusted using the validation set; and the trained deep learning model is tested using the test set to obtain test metrics. Obtain the indicator threshold; compare the test indicator with the indicator threshold; if all test indicators are greater than the indicator threshold, then mark the deep learning model as an image recognition model; otherwise, rebuild and retrain the image recognition model.

4. The elevator intelligent control system based on artificial intelligence according to claim 1, characterized in that, The process of determining the elevator operating mode based on the recognition results includes: Retrieve the identification tags; Obtain the tag runtime library; The elevator operation mode is obtained by matching the identification tags with the tag runtime library; the elevator operation mode includes: personnel identification mode and personnel riding mode.

5. The elevator intelligent control system based on artificial intelligence according to claim 4, characterized in that, The method of controlling the elevator according to the elevator operating mode includes: When the elevator is in human recognition mode and the recognition tag type is electric vehicle entering the elevator, an electric vehicle alarm voice will be generated, and the elevator image will be continuously recognized until no electric vehicle is recognized in the elevator image. When the elevator is in passenger mode and the identification tag type is abnormal behavior, it is determined whether the number of people is greater than the number of registered floors; if yes, a voice alarm signal is issued and the abnormal floor is deregistered; if no, the behavior of obstructing the elevator door is analyzed and a voice reminder signal is generated. When the elevator is in personnel identification mode and the identification tag type is "solo passenger", the control is based on the type of solo passenger. When the elevator is in the personnel identification model and the identification tag type is "personnel stuck", the elevator door will be opened and passengers will be reminded to leave the car or press the floor button.

6. The elevator intelligent control system based on artificial intelligence according to claim 1, characterized in that, The control based on the type of person using the elevator alone includes: Retrieve the corresponding elevator images; use image recognition technology to identify the types of people in the elevator images; the types of people include: children and women; When the passenger is a child, the elevator door will remain open, and a corresponding voice prompt will be generated. When the passenger is a woman, the elevator will proceed directly to the target floor, and the video inside the elevator will be recorded until the passenger leaves the elevator.

7. The elevator intelligent control system based on artificial intelligence according to claim 1, characterized in that, The method of detecting elevator malfunctions based on real-time log data includes: Acquire real-time log data of elevator operation; the real-time log data includes: sensor data, operation status log, and error code log; Obtain a preset standard range; compare the real-time log data with the standard range, and generate a fault alarm signal when the real-time log data exceeds the standard range.

8. The elevator intelligent control system based on artificial intelligence according to claim 7, characterized in that, The elevator maintenance based on the test results includes: Retrieve real-time log data that exceeds the standard range and mark it as abnormal data; obtain the corresponding standard data; calculate the difference between the abnormal data and the standard data to obtain the deviation data; Obtain the deviation handling library; match the deviation data with the deviation handling library to obtain the corresponding fault handling measures, and send the fault handling measures to the corresponding technical personnel.

9. The elevator intelligent control system based on artificial intelligence according to claim 1, characterized in that, The rescue operation for the trapped personnel includes: Acquire elevator images when an elevator malfunctions; if people are trapped when an elevator malfunctions, generate an auxiliary voice prompt to prompt the trapped people to press the alarm button, and send the entrapment alarm information to the corresponding rescue personnel; When rescuers confirm that they have received the trapped alarm information, they will display the rescue information in real time on the smart screen. The rescue information includes: location coordinates, direction of movement and estimated arrival time.

10. An artificial intelligence-based intelligent elevator control method, applied to an artificial intelligence-based intelligent elevator control system as described in any one of claims 1-9, characterized in that, include: Elevator images are obtained by capturing images inside the elevator using an image acquisition device; Obtain real-time log data of elevator operation; The elevator images were analyzed and identified to obtain the recognition results; Determine the elevator operating mode based on the identification results; Control the elevator according to its operating mode; Fault detection of elevators is performed based on real-time log data; Based on the test results, the elevator was repaired and the trapped people were rescued.

Images