An intelligent elevator and method for monitoring obstacles

Abstract

This invention discloses an intelligent elevator and method for monitoring obstacles. The elevator is equipped with a controller, a processor, an image acquisition device, and an alarm device. The image acquisition device is used to acquire images of the elevator interior, and the alarm device is used to output alarm information. The controller monitors whether there are moving figures inside the elevator based on the images acquired by the image acquisition device. When a moving figure is detected, the controller controls the image acquisition device to acquire indoor images according to a preset frequency curve. The processor extracts obstacle features from the acquired images, confirms whether an obstacle is present in the acquired images, and feeds back to the controller to control the alarm device to output alarm information and to start or stop the elevator. The obstacle is an object connected to or near the side frame of the elevator door. This device uses real-time monitoring combined with an obstacle recognition model to predict obstacles, enabling timely feedback on elevator obstacles and avoiding situations such as anti-pinch sensor failure or over-processing.

Description

Technical Field

[0001] This invention relates to the field of elevator interior space monitoring technology, and in particular to an elevator and method for intelligently monitoring obstacles. Background Technology

[0002] Elevators are the most common and convenient vertical lifting tool in buildings. To prevent elevator doors from accidentally trapping passengers or objects, anti-pinch sensors are usually installed between the elevator doors and the arrival doors. When an obstacle touches the anti-pinch sensor during the closing process of each door, the door will immediately stop closing and reopen to achieve safety protection.

[0003] However, in existing elevators, due to age or lack of maintenance, the anti-pinch sensor is prone to malfunction and loses its anti-pinch function. Alternatively, when the obstacle (such as a rope or wire) is small in size or thickness, it may only touch the car door and the departure door, but not press against the anti-pinch sensor. In this case, even if the anti-pinch sensor is functioning normally, the doors will continue to close, and the car will continue to move up and down, potentially causing serious accidents such as clothes being torn or pets being strangled.

[0004] To address the aforementioned problems, Japanese Patent (Application No.: JP2016117959) provides a pet protection system for elevators that can cut the cord connecting a pet to the elevator door and quickly protect the pet. The elevator pet protection system is installed on the upper frame and threshold of the elevator entrance / exit. When the elevator car moves while the cord is caught between the car door and the entrance / exit door, the cutting part cuts the cord.

[0005] While the above solutions protect pets' lives to some extent, the cutting mechanism is significantly less effective against hard metal ropes; furthermore, cutting other items can render them unusable, causing unnecessary harm to people and property. Summary of the Invention

[0006] The main objective of this invention is to propose an intelligent elevator and method for monitoring obstacles, aiming to solve the technical problem that when there are obstacles in the elevator doors during closing, the elevator cannot sense them in time or handle them excessively.

[0007] To achieve the above objectives, the present invention proposes an intelligent obstacle monitoring elevator, wherein the elevator is equipped with a controller, a processor, an image acquisition device, and an alarm device, the image acquisition device being used to acquire images of the elevator interior, and the alarm device being used to output alarm information; The controller monitors whether there are moving people inside the elevator based on the images captured by the image acquisition device. When there are moving people inside the elevator, the controller controls the image acquisition device to acquire indoor images according to a preset frequency curve. The processor extracts obstacle features from the acquired images, confirms whether there are obstacles in the acquired images, and feeds back to the controller to control the alarm device to output alarm information. The obstacle is an object connected to or near the side frame of the elevator door. The processor is equipped with an obstacle recognition model that monitors the inside of the elevator. The obstacle recognition model is trained with several image samples marked with obstacles as input. The obstacle recognition model is used to predict the probability of an obstacle in the collected image. If the probability is greater than a set threshold of 0.95, then it is determined that an obstacle exists in the image.

[0008] Optionally, the controller also controls the image acquisition device inside the elevator to acquire images of the elevator interior at a preset frequency; the preset frequency is less than or equal to the lowest frequency in the preset frequency curve; the processor compares the currently acquired image with at least one frame of image acquired near the current acquisition time to determine whether there is a moving person inside the elevator.

[0009] Optionally, the preset frequency curve is the correspondence between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time.

[0010] Optionally, when there are no moving people inside the elevator, the controller reduces the current frequency of the image acquisition device and acquires images of the elevator interior at the reduced frequency.

[0011] Optionally, the image acquisition device is located at the top and / or bottom and / or side of the elevator car.

[0012] Optionally, the obstacle is a connecting object between the person and the side frame of the elevator door.

[0013] Optionally, the image acquisition device includes two cameras and a driving component. The two cameras are positioned near the elevator door, and the driving component drives the two cameras to move simultaneously with the two sides of the door.

[0014] Optionally, the shooting frequency and movement distance of the two cameras are related by f = 20 * s, where s is in meters.

[0015] This invention also proposes a method for intelligently monitoring obstacles, the monitoring method comprising the following steps: Use image acquisition devices inside the elevator to monitor whether there are moving people inside the elevator; When a moving person is detected inside the elevator, images inside the elevator are collected according to a preset frequency curve; the preset frequency curve is the correspondence between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time. Obstacle features are extracted from the acquired images to confirm whether obstacles are present in the acquired images; If an obstacle is found inside the elevator, an alarm message will be output and the elevator will stop moving.

[0016] Optionally, the step of the image acquisition device inside the elevator monitoring whether there is a moving person inside the elevator includes: Images of the elevator interior are captured at a preset frequency, wherein the preset frequency is less than or equal to the lowest frequency in the preset frequency curve. The currently captured image is compared with at least one frame captured around the same time to determine whether there is a moving person inside the elevator.

[0017] Optionally, after the step of comparing the currently acquired image with at least one frame of image acquired near the current acquisition time to determine whether there is a moving person inside the elevator, the method further includes: when there is no moving person inside the elevator, reducing the current frequency of the image acquisition device, and acquiring images of the elevator interior at the reduced frequency.

[0018] In the technical solution of this invention, a controller, processor, image acquisition device, and alarm device are installed in the elevator. By acquiring images of the elevator interior, the system determines whether there are moving figures, extracts obstacle features from the acquired images, confirms the presence of obstacles, and feeds this information back to the controller, which then controls the alarm device to output an alarm message. This device uses real-time monitoring combined with an obstacle recognition model to predict obstacles, enabling timely feedback on elevator obstacles and avoiding situations such as anti-pinch sensor failure or over-processing. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0020] Figure 1 A three-dimensional structural diagram of an embodiment of an elevator with intelligent obstacle monitoring provided by the present invention; Figure 2 for Figure 1 A bottom view of the top floor of the elevator car; Figure 3 for Figure 2 A schematic diagram of the middle section structure; Figure 4 for Figure 1 A cross-sectional view of the top section of the elevator car; Figure 5 This is a flowchart illustrating an embodiment of the intelligent obstacle monitoring method provided by the present invention; Figure 6 The logic diagram of an embodiment of an elevator with intelligent obstacle monitoring provided by the present invention.

[0021] In the diagram: Elevator-100, Image acquisition device-1, Drive unit-11, Camera-12, Rack-13, Gear-14, Top floor of the car-2.

[0022] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of this application.

[0024] To better describe and illustrate the embodiments of this application, reference may be made to one or more accompanying drawings, but the additional details or examples used to describe the drawings should not be considered as limiting the scope of any of the inventive creations of this application, the embodiments or preferred methods described herein.

[0025] In the description of this invention, it should be noted that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and do not indicate that the device referred to must have a specific orientation or operate in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0027] In existing elevators, due to age or lack of maintenance, the anti-pinch sensors are prone to malfunction and lose their anti-pinch function. Japanese Patent (Application No.: JP2016117959) provides a pet protection system for elevators that can cut the cord connecting to a pet when it gets caught in the elevator door, quickly protecting the pet. The system is installed on the upper frame and threshold of the elevator entrance / exit. When the elevator moves while the car is moving with the cord caught in the door, the cutting part cuts the cord. While this solution protects pets to some extent, the cutting part's effectiveness is greatly reduced for hard metal cords; furthermore, cutting other items can damage them, causing unnecessary harm to people and property.

[0028] In view of this, the present invention proposes an elevator with intelligent obstacle monitoring. Figure 1-4 This invention provides an embodiment of an elevator with intelligent obstacle monitoring. Please refer to [link / reference]. Figure 1-4 The elevator 100 is equipped with a controller, a processor, an image acquisition device 1, and an alarm device. The image acquisition device 1 is used to acquire images of the elevator interior, and the alarm device is used to output alarm information. It should be understood that the image acquisition device 1 includes devices such as a camera, preferably an electrically operated zoom camera, and the alarm device includes audible and visual alarms.

[0029] Please see Figure 6 The controller monitors the presence of moving figures inside the elevator based on images acquired by the image acquisition device 1. When a moving figure is detected, the controller directs the image acquisition device 1 to acquire indoor images according to a preset frequency curve. The controller transmits the images to a processor, which has a built-in algorithm model. This processor extracts obstacle features from the acquired images to determine if obstacles are present and feeds this information back to the controller, which then controls the alarm device to output an alarm message. It should be noted that obstacles are objects connected to or near the elevator door side frame, especially objects connecting a person to the elevator door side frame. More specifically, objects connecting a person to the door seam are considered obstacles. Additionally, objects located between the elevator door side frames are also considered obstacles. The processor includes an obstacle recognition model for monitoring the elevator interior. This model is trained using several image samples labeled with obstacles. The model predicts the probability of obstacles in the acquired images; if the probability exceeds a set threshold, an obstacle is confirmed to exist in the image.

[0030] In the technical solution of this invention, the elevator is equipped with a controller, processor, image acquisition device 1, and alarm device. By acquiring images of the elevator interior, the system determines whether a moving person is present, extracts obstacle features from the acquired images, confirms the presence of obstacles, and feeds this information back to the controller, which then controls the alarm device to output an alarm message. This device uses real-time monitoring combined with an obstacle recognition model to predict obstacles, enabling timely feedback on elevator obstacles and avoiding situations such as anti-pinch sensor failure or over-processing.

[0031] To achieve dynamic recognition of people, in one embodiment of the present invention, the controller further controls the image acquisition device 1 inside the elevator to acquire images inside the elevator at a preset frequency; the preset frequency is less than or equal to the lowest frequency in the preset frequency curve; the processor compares the currently acquired image with at least one frame of image acquired near the current acquisition time to determine whether there is a moving person inside the elevator.

[0032] Furthermore, in one embodiment of the present invention, the preset frequency curve is a correspondence between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time. Since obstacles may block or trap the elevator 100 during its closing time, the sampling frequency gradually increases during this short period, resulting in more sampled images of the obstacles, thus making the recognition model's recognition results more accurate.

[0033] To reduce the operating load of the image acquisition device 1, in one embodiment of the present invention, when there are no moving people in the elevator, the controller reduces the current frequency of the image acquisition device 1 and acquires images of the elevator interior at the reduced frequency, thereby reducing its power consumption.

[0034] To predict the presence of obstacles, in this embodiment, the obstacle recognition model is a hierarchical model composed of several convolutional kernels, each consisting of different weights. After the acquired image is input into the obstacle recognition model, it first undergoes a matrix inner product operation with the weights in the first layer of convolutional kernels. The result is then further processed with the next layer of convolutional kernels. The output matrix of the final layer of convolutional kernels is unfolded into a vector, which serves as the input to the fully connected layer. The fully connected layer has two outputs, which are transformed by the Softmax function into two values ​​that sum to 1, such as [0.8, 0.2]. These values ​​represent the probability of an obstacle being present and the probability of an obstacle not being present in the image, respectively. A threshold is set; if the probability of an obstacle being present is greater than this threshold, the image is considered to contain an obstacle. In this embodiment, this threshold is set to 0.95.

[0035] Initially, the weights in the obstacle recognition model are randomized. To obtain the correct probability from the input image, the model's parameters need to be trained. The training process is as follows: Step 1: Initialize the network weights.

[0036] Step 2: The input data is propagated forward through the convolutional layer, downsampling layer, and fully connected layer to obtain the output value.

[0037] Step 3: Calculate the error between the network's output value and the target value, where the target value is a value like [0, 1] or [1, 0].

[0038] Step 4: When the error exceeds our expected value, the error is fed back into the network, and the errors of the fully connected layer, downsampling layer, and convolutional layer are calculated sequentially. The error of each layer can be understood as how much of the total network error the network should bear; when the error is equal to or less than our expected value, training ends.

[0039] Step 5: Update the weights based on the calculated error, and then proceed to step 2.

[0040] By training the parameters in the obstacle recognition model in this way, the obstacle recognition model can obtain the correct output probability value after inputting an image.

[0041] Please see Figure 1 In one embodiment of the present invention, the image acquisition device 1 is disposed on the top and / or bottom and / or side of the elevator car. For a preferred embodiment, please refer to... Figure 1 and Figure 2 The image acquisition device 1 is installed on the top of the elevator car, near the elevator door, to prevent obstruction by other objects. This allows for clearer imaging of the connection between a person and the elevator door side frame, or objects near the elevator door side frame. The image acquisition device 1 includes two cameras 12. To improve the imaging of obstacles, both cameras 12 use telephoto lenses, resulting in clearer magnified images and more accurate obstacle recognition model outputs.

[0042] Specifically, please refer to Figure 3 and Figure 4To ensure accurate obstacle detection during elevator door closing, the image acquisition device 1 further includes a drive unit 11, which is installed above the top layer 2 of the elevator car. Two cameras 12 are installed below the top layer 2. Two racks 13 are slidably installed inside the top layer 2, and each rack 13 is connected to one of the cameras 12. Gears 14 are also rotatably installed inside the top layer 2, meshing with the two racks 13. The drive unit 11 drives the gears 14 to rotate.

[0043] The elevator car's top layer 2 is equipped with a slide rail that slides with the rack 13. When the gear 14 rotates, it drives the two racks 13 to slide, causing the two cameras 12 to move towards or away from each other. Preferably, the drive component 11 is a servo motor, and it is interlocked with the elevator door. When the elevator door opens or closes, the drive component 11 is simultaneously controlled to open or close, so that the two cameras 12 can move synchronously with the elevator doors on both sides, thereby accurately capturing obstacles at the elevator door.

[0044] It should be noted that, due to the different opening and closing speeds of various elevator doors, the shooting frequency and movement distance are set to be positively correlated. In this embodiment, the shooting frequency and movement distance of the two cameras are related as f = 20*s, where s is in meters (m). That is, when the two cameras 12 move synchronously with the elevator door for 0.05m, both cameras take one shot. Of course, the implementation of this invention is not limited to this; the ratio of shooting frequency to movement distance can be larger, resulting in more images being collected and higher accuracy of the processor's output probability value. However, this would affect the speed at which the processor outputs the probability value. Here, f = 20*s is set to ensure both the accuracy of the probability value and timely feedback.

[0045] In one embodiment of the present invention, after confirming that an obstacle appears in the acquired image, the controller controls the alarm device to output alarm information, and also controls the elevator door to stop closing and start opening.

[0046] This invention also proposes a method for intelligently monitoring obstacles, combined with Figure 5 A flowchart illustrating the provided method for intelligent obstacle monitoring, the monitoring method comprising the following steps: Step S10: Use the image acquisition device inside the elevator to monitor whether there are moving people inside the elevator.

[0047] Step S20: When a moving person is detected inside the elevator, images inside the elevator are collected according to a preset frequency curve; the preset frequency curve is the correspondence between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time.

[0048] Step S30: Extract obstacle features from the acquired image to confirm whether there are obstacles in the acquired image.

[0049] In this embodiment, an obstacle recognition model is used to determine whether an obstacle exists in the acquired image. Specifically, the obstacle recognition model is trained using several image samples labeled with obstacles as input; the obstacle recognition model is then used to predict the probability of an obstacle existing in the acquired image; if the probability is greater than a set threshold, then an obstacle is determined to exist in the image.

[0050] Step S40: If an obstacle appears inside the elevator, output an alarm message and stop the elevator from moving.

[0051] In this embodiment, step S10, where the image acquisition device inside the elevator monitors whether there are moving people inside the elevator, includes: Step S11: Acquire images of the elevator interior at a preset frequency, wherein the preset frequency is less than or equal to the lowest frequency in the preset frequency curve.

[0052] Step S12: Compare the currently acquired image with at least one frame of image acquired around the current acquisition time to determine whether there is a moving person inside the elevator.

[0053] Furthermore, after step S12, which compares the currently acquired image with at least one frame acquired around the current acquisition time to determine whether there is a moving person inside the elevator, the method further includes: Step S13: When there are no moving people in the elevator, reduce the current frequency of the image acquisition device and acquire images of the elevator interior at the reduced frequency.

[0054] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. An elevator with intelligent obstacle monitoring, characterized in that, The elevator is equipped with a controller, a processor, an image acquisition device, and an alarm device. The image acquisition device is used to acquire images inside the elevator, and the alarm device is used to output alarm information. The controller monitors whether there are moving people inside the elevator based on the images acquired by the image acquisition device. When there are moving people inside the elevator, the image acquisition device acquires indoor images according to a preset frequency curve. The processor extracts obstacle features from the acquired images, confirms whether there are obstacles in the acquired images, and feeds back to the controller to control the alarm device to output alarm information and start / stop the elevator. The obstacle is an object connected to or near the side frame of the elevator door; the processor is equipped with an obstacle recognition model that monitors the inside of the elevator. The obstacle recognition model is trained with several image samples marked with obstacles as input. The obstacle recognition model is used to predict the collected images to determine the probability that there is an obstacle in the image. Images with a probability greater than a set threshold of 0.95 are selected to determine that there is an obstacle in the image. The preset frequency curve represents the relationship between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time. The image acquisition device includes two cameras and a driving component. The two cameras are positioned near the elevator door, and the driving component drives the two cameras to move synchronously with the two sides of the door.

2. The elevator with intelligent obstacle monitoring as described in claim 1, characterized in that, The controller also controls the image acquisition device inside the elevator to acquire images of the elevator interior at a preset frequency; the preset frequency is less than or equal to the lowest frequency in the preset frequency curve; the processor compares the currently acquired image with at least one frame of image acquired around the current acquisition time to determine whether there is a moving person inside the elevator.

3. The elevator with intelligent obstacle monitoring as described in claim 1, characterized in that, When there are no moving people inside the elevator, the controller reduces the current frequency of the image acquisition device and acquires images of the elevator interior at the reduced frequency.

4. The elevator with intelligent obstacle monitoring as described in claim 1, characterized in that, The image acquisition device is located on the top and / or bottom and / or side of the elevator car.

5. A method for intelligently monitoring obstacles, characterized in that, The method applied to the intelligent obstacle monitoring elevator of claim 1 includes the following steps: The image acquisition device inside the elevator is used to monitor whether there are moving people inside the elevator; When a moving person is detected inside the elevator, images inside the elevator are collected according to a preset frequency curve; the preset frequency curve is the correspondence between image sampling frequency and time, and in the preset frequency curve, the image sampling frequency increases with time. Obstacle features are extracted from the acquired images to confirm whether obstacles are present in the acquired images; If an obstacle is found inside the elevator, an alarm message will be output and the elevator will stop moving.

6. The method for intelligently monitoring obstacles as described in claim 5, characterized in that, The steps of the image acquisition device inside the elevator to monitor whether there are moving people inside the elevator include: Images of the elevator interior are captured at a preset frequency, wherein the preset frequency is less than or equal to the lowest frequency in the preset frequency curve. The currently captured image is compared with at least one frame captured around the same time to determine whether there is a moving person inside the elevator.

7. The method for intelligently monitoring obstacles as described in claim 6, characterized in that, The step of comparing the currently acquired image with at least one frame of image acquired around the current acquisition time to determine whether there is a moving person inside the elevator further includes: when there is no moving person inside the elevator, reducing the current frequency of the image acquisition device and acquiring images of the elevator interior at the reduced frequency.

Images