Method and system for determining a filling level of an elevator car
By using a pre-trained convolutional neural network model to determine the elevator car filling level, the inaccuracy and environmental sensitivity of traditional methods are solved, achieving more accurate and flexible elevator car filling level detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- KONE OYJ
- Filing Date
- 2023-11-08
- Publication Date
- 2026-06-05
AI Technical Summary
In existing technologies, methods for determining the elevator car filling level based on weight and image analysis are inaccurate, causing unnecessary stops of the elevator car at landings. Furthermore, traditional image analysis methods are sensitive to environmental changes and require calibration.
A pre-trained filling level classification model based on convolutional neural networks is used to acquire elevator car image data through an imaging device. The filling level is determined by a computing unit without calibration and can adapt to different environmental changes.
It improves the accuracy and adaptability of determining the elevator car filling level, reduces unnecessary stops at floors, and is suitable for various elevator car types.
Smart Images
Figure CN122161772A_ABST
Abstract
Description
Technical Field
[0001] This invention generally relates to the technical field of elevator systems. In particular, this invention relates to the filling level of an elevator car. Background Technology
[0002] The elevator car fill level can be used for various operations of the elevator system. For example, the elevator car fill level can be used for elevator call allocation. For instance, if the elevator fill level indicates that the elevator car is full, no further floor calls are allocated to said elevator car. In other words, an elevator car determined to be full based on the determined fill level bypasses multiple floors between the departure floor and the destination floor, regardless of whether floor calls are generated from the bypassed floors.
[0003] The filling level of an elevator car can be determined, for example, by weight-based determination. In weight-based determination, the filling level of the elevator car can be determined based on weight data provided by the elevator car's weighing system. However, weight-based determination may be inaccurate due to non-personnel objects (such as luggage, wheelchairs, and / or cargo) within the elevator car. Therefore, weight-based determination may lead to an incorrect determination of the elevator car's filling level, which in turn may result in unnecessary stops of the elevator car at landings, even though there is no space within the elevator car for new passengers.
[0004] Alternatively, the elevator car's fill level can be determined using image analysis-based determination. In image analysis-based determination, the elevator car's fill level can be determined based on image data provided by an optical imaging device. Image analysis-based determination requires calibration, which is typically done using one or more baseline images provided by the optical imaging device. These one or more baseline images represent an empty elevator car, i.e., the interior of the elevator car is empty. In image analysis-based determination, when assuming the elevator car is filled with passengers and / or non-personnel objects, the elevator car's fill level can be determined, for example, based on the image differences between one or more baseline images and at least one image of the elevator car's interior captured by the optical imaging device. The accuracy of image analysis-based determination can depend on one or more baseline images. For example, if the elevator car floor changes (e.g., due to new carpet, new flooring material, carpet wear, flooring material wear, or floor fading, etc.), the accuracy of the image analysis-based determination solution may decrease, and the determination results will be affected.
[0005] Therefore, there is a need to further develop solutions for determining the filling level of elevator cars. Summary of the Invention
[0006] The following is a simplified summary of the invention to provide a basic understanding of some aspects of various embodiments of the invention. This summary is not a broad overview of the invention. It is neither intended to identify key or essential elements of the invention nor to depict its scope. The following summary presents only some concepts of the invention in a simplified form, serving as a prelude to a more detailed description of exemplary embodiments of the invention.
[0007] The purpose of this invention is to provide a method, a system, a computer program, and a computer-readable medium for determining elevator car fill level data; and a method, a computer program, and a computer-readable medium for training a fill level classification model for determining elevator car fill level data. Another object of this invention is that the method, fill level determination system, elevator system, computer program, and computer-readable medium can provide an image analysis-based solution for determining elevator car fill level without calibration.
[0008] The object of the present invention is achieved by the method, fill level determination system, elevator system, computer program and computer-readable medium as defined by the individual claims.
[0009] According to a first aspect, a method for determining fill level data of an elevator car is provided, wherein the method includes: acquiring image data of the interior of the elevator car; determining fill level data of the elevator car from the acquired image data by applying a pre-trained fill level classification model; and providing the determined fill level data of the elevator car to an elevator control system for further elevator use.
[0010] Determining the fill level data of the elevator car may include determining, from the acquired image data, a fill level belonging to multiple predetermined fill level levels by applying a pre-trained fill level classification model, which corresponds to the fill level of the elevator car in the acquired image data.
[0011] The determined elevator car fill level data may include data indicating the determined fill level grade.
[0012] Further elevator use may include elevator call assignment and / or information presentation.
[0013] Pre-trained filled-level classification models can be based on convolutional neural networks (CNNs).
[0014] According to a second aspect, a fill level determination system is provided for determining fill level data of an elevator car, wherein the fill level determination system includes: an imaging device disposed inside the elevator car and configured to generate image data of the elevator car interior; and a computing unit configured to: obtain the image data of the elevator car interior from the imaging device; determine the fill level data of the elevator car from the obtained image data by applying a pre-trained fill level classification model; and provide the determined fill level data of the elevator car to an elevator control system for further elevator use.
[0015] Determining the fill level data of the elevator car may include: the computing unit may be configured to determine, from the obtained image data, a fill level belonging to a plurality of predetermined fill level levels by applying a pre-trained fill level classification model, which corresponds to the fill level of the elevator car in the obtained image data.
[0016] The determined elevator car fill level data may include data indicating the determined fill level grade.
[0017] Further elevator use may include elevator call assignment and / or information presentation.
[0018] Pre-trained filled-level classification models can be based on convolutional neural networks (CNNs).
[0019] According to a third aspect, an elevator system is provided, wherein the elevator system includes: at least one elevator car arranged to travel along a corresponding elevator shaft, an elevator control system configured to control the operation of the elevator system; and a fill level determination system as described above.
[0020] According to a fourth aspect, a computer program is provided, wherein the computer program includes instructions that, when executed by a computer, cause the computer to perform the method as described above.
[0021] According to a fifth aspect, a computer-readable medium is provided, wherein the computer-readable medium includes instructions that, when executed by a computer, cause the computer to perform the method as described above.
[0022] According to a sixth aspect, a method is provided for training a fill level classification model for determining fill level data of an elevator car, wherein the method includes: obtaining training input data comprising a plurality of images including the interior of at least one elevator car; classifying the plurality of images of the interior of at least one elevator car into a plurality of predetermined fill level levels; and training a fill level classification model for determining fill level data of an elevator car by using the training input data classified into the plurality of predetermined fill levels.
[0023] According to a seventh aspect, a computer program is provided, wherein the computer program includes instructions that, when executed by a computer, cause the computer to perform the method described above for training a filled-level classification model.
[0024] According to an eighth aspect, a computer-readable medium is provided, wherein the computer-readable medium includes instructions that, when executed by a computer, cause the computer to perform the method described above for training a filled-level classification model.
[0025] Various exemplary and non-limiting embodiments of the invention (with regard to structure and operation) and their additional objects and advantages will be best understood from the following description of specific exemplary and non-limiting embodiments when read in conjunction with the accompanying drawings.
[0026] The verbs “comprising” and “including” are used herein as open-ended restrictions, neither excluding nor requiring the presence of any unlisted features. Unless otherwise expressly stated, the features recited in the dependent claims may be freely combined with each other. Furthermore, it should be understood that the use of “a” or “an,” i.e., the singular form, throughout the document does not exclude a plurality. Attached Figure Description
[0027] The embodiments of the invention are illustrated in the accompanying drawings by way of example and not limitation.
[0028] Figure 1 An example of an elevator system is illustrated schematically.
[0029] Figure 2 An exemplary implementation of a fill level determination system in an elevator system is illustrated schematically.
[0030] Figure 3 An example of a method for determining the fill level data of an elevator car is illustrated schematically.
[0031] Figure 4 This illustration illustrates an example of the input-output process for determining the fill level data of an elevator car by applying a pre-trained fill level classification model.
[0032] Figure 5 An example of a method for training a filled-level classification model is illustrated schematically.
[0033] Figure 6 An example is illustrated whereby the fill level data of an elevator car is determined from the obtained image data by applying a pre-trained fill level classification model.
[0034] Figure 7An example of a component of the computational unit of a system for determining the fill level is shown schematically.
[0035] Figure 8 An example of a component of a computational entity is shown schematically. Detailed Implementation
[0036] Figure 1 An example of an elevator system 100 is schematically shown. The elevator system 100 includes at least one elevator car 110, which is configured to travel between multiple floors along a respective elevator shaft 120. Figure 1 The example elevator system 100 includes an elevator car 110 traveling along an elevator shaft 120; however, the elevator system 100 may also include an elevator group, i.e., a group of two or more elevator cars 110, each elevator car 110 traveling along a separate elevator shaft 120, the elevator shaft 120 being configured as a unit operation serving the same floor (for clarity, Figure 1 (Multiple floors are not shown). The elevator system 100 also includes an elevator control system, such as an elevator controller 130. The elevator control system 130 is configured to at least partially control the operation of the elevator system 100. The elevator control system 130 may reside, for example, in a machine room (not shown in the diagram for clarity). Figure 1 (As shown in the diagram) or reside in a floor of elevator system 100. Elevator system 100 may also include one or more other known elevator-related entities, such as hoisting systems, user interface devices, safety circuits and devices, elevator door systems, etc., which are not shown in the diagram for clarity. Figure 1 As shown in the diagram. The elevator system 100 also includes a fill level determination system 200 for determining fill level data of the elevator car 110 (for clarity, ...). Figure 1 (The entity of the fill level determination system 200 is not shown in the diagram). The fill level data of the elevator car 110 represents the fill level of the elevator car 110. The fill level of the elevator car 110 defines the degree to which the elevator car 110 is filled.
[0037] Figure 2 An exemplary embodiment of a fill level determination system 200 in an elevator system 100 is schematically illustrated. The fill level determination system 200 includes an imaging device 210 and a computing unit 220. The imaging device 210 is communicatively coupled to the computing unit 220. Communication between the computing unit 220 and the imaging device 210 may be based on one or more known communication technologies, wired or wireless.
[0038] Imaging device 210 is configured to generate image data of the interior of elevator car 110. Imaging device 210 is disposed inside elevator car 110. Imaging device 210 can be arranged (e.g., installed) at different installation locations (i.e., placement) inside elevator car 110. This makes the installation location of imaging device 210 not limited to a specific installation location (e.g., the center of ceiling 202). Imaging device 210 can preferably be placed near the ceiling 202 of elevator car 110, i.e., as high as possible, and so on. Figure 2 As shown in the examples. Some non-limiting examples of mounting placement for the imaging device 210 may include: wall placement, corner placement, and ceiling placement. In a wall placement, the imaging device 210 may be mounted on a wall of the elevator car 110. The wall placement may be, for example, a rear wall placement, wherein the imaging device 210 may be mounted on the rear wall 204 of the elevator car 110, preferably near the ceiling 202. Alternatively, the wall placement may be at any other wall of the elevator car 110, i.e., the imaging device 210 may be mounted on any other wall of the elevator car 110, preferably near the ceiling 202. In a corner placement, the imaging device 210 may be mounted in the upper corner of the elevator car 110. The corner placement may be, for example, at an upper rear corner placement, wherein the imaging device 210 may be mounted at the upper rear corner of the elevator car 110. The upper rear corner of the elevator car 110 may be any of the upper rear corners of the elevator car 110. Alternatively, the corner placement can be a front-upper corner placement, where the imaging device 210 can be mounted at the front-upper corner of the elevator car 110. The front-upper corner of the elevator car 110 can be any one of the front-upper corners of the elevator car 110. In a ceiling placement, the imaging device 210 can be mounted on the ceiling 202 of the elevator car 110, such as... Figure 2 As shown in the example, the imaging device 210 can be mounted such that the image data provided by the imaging device 210 covers the largest possible area of the elevator car 110. Preferably, the imaging device 210 can be mounted such that the image data provided by the imaging device 210 at least completely covers the floor 206 of the elevator car 110. The imaging device 210 can be, for example, an optical imaging device configured to produce optical image data. The imaging device 210 can include, for example, a camera, such as a red-green-blue (RGB) camera or a monochrome camera. The imaging device 210 can provide image data with high resolution and / or wide field of view (FOV) to cover the maximum area of the elevator car 110 with the image data.
[0039] Computing unit 220 can be configured to at least partially control one or more operations of fill level determination system 200. The implementation of computing unit 220 can be as a standalone computing entity or as a distributed computing environment among multiple standalone computing entities (e.g., multiple servers) providing distributed computing resources. Computing unit 220 can be a local computing unit or a remote computing unit. Alternatively, computing unit 220 can be implemented as a combined computing system including both local and remote computing units. Computing unit 220 implemented as a local computing unit can be located in elevator car 110 (e.g., on the roof of elevator car 110 or at any other location within elevator car 110, inside or outside elevator car 110). Alternatively or additionally, computing unit 220 implemented as a local computing unit can be located at any field location within elevator system 100. According to an example, computing unit 220 implemented as a local computing unit can be part of elevator control system 130. Computing unit 220 implemented as a remote computing unit can be located at any off-site location remote from elevator system 100. The computing unit 220, implemented as a remote computing unit, may include, for example, one or more computing entities located remotely from the elevator system 100. The computing unit 220 implemented as a remote computing unit may be, for example, but not limited to, at least one of the following: a cloud-based computing unit (e.g., a cloud server), a service center, a data center, a remote monitoring system, a remote diagnostic system, or any other remote computing unit. The computing unit 220 may be communicatively connected to the elevator control system 130 of the elevator system 100. Communication between the computing unit 220 and the elevator control system 130 may be based on one or more known wired or wireless communication technologies.
[0040] Next, through reference Figure 3 An example of a method for determining the fill level data of elevator car 110 is described. Figure 3 The method is schematically illustrated in a flowchart. The method is executed by the computing unit 220 of the aforementioned fill level determination system 200; that is, the method can be implemented by a computer. It is described using an elevator car 110. Figure 3 The example method is as described above, but if the elevator system 100 includes two or more elevator cars 110, the method can also be applied accordingly to determine the fill level data of each elevator car 110 of the elevator system 100. If the elevator system 100 includes two or more elevator cars 110, and the method is used to determine the fill level data of multiple elevator cars 110 of the elevator system 100, the imaging device 210 can be arranged in each elevator car 110 in which the fill level data is to be determined, and the computing unit 220 can be implemented as a common computing unit and / or a separate computing unit for each elevator car 110 in which the fill level data is to be determined.
[0041] At step 310, the computing unit 220 obtains image data 410 of the interior of the elevator car 110 from the imaging device 210. As described above, the imaging device 210 is configured to generate image data 410 of the interior of the elevator car 110. The image data may be, for example, optical image data. The imaging device 210 is also configured to provide the generated image data 410 to the computing unit 220. Preferably, the imaging device 210 may provide the generated image data 410 to the computing unit 220 substantially immediately after generating the image data 410. The image data 410 generated by the imaging device 210 may, for example, include one or more images (i.e., frames) and / or a video image including multiple consecutive images. According to an example, the imaging device 210 may generate the image data 410 of the interior of the elevator car 110 in response to receiving a control signal from the computing unit 220. The control signal may, for example, include an instruction to trigger the generation of the image data 410 of the interior of the elevator car 110. The generation of control signals from computing unit 220 to imaging device 210 can be triggered, for example, by sensor-based trigger events, user-input-based trigger events, and / or any other trigger events. According to another example, imaging device 210 can continuously generate image data 410 from the interior of elevator car 110 and continuously provide the generated image data 410 to computing unit 220. According to another example, imaging device 210 can continuously generate image data 410 from the interior of elevator car 110 and, in response to receiving a control signal from computing unit 220 including an instruction to provide image data 410 of the interior of elevator car 110, provide the generated image data 410 to computing unit 220.
[0042] In step 320, the computing unit 220 determines the fill level data 430 of the elevator car 110 from the obtained image data 410 by applying a pre-trained fill level classification model 420. In other words, in step 320, the obtained image data 410 is used as the input data of the pre-trained fill level classification model 420, and the input data 410 is processed by the pre-trained fill level classification model 420 to produce the determined fill level data 430 as the output data of the pre-trained fill level classification model 420. Figure 4An example of the input-output process for determining the fill level data 430 of an elevator car 110 from acquired image data 410 by applying a pre-trained fill level classification model 420 is illustrated. The pre-trained fill level classification model 420 can be a neural network-based model, such as a convolutional neural network (CNN)-based model. A non-limiting example of a CNN-based model could be MobileNet V3. However, any other neural network-based model can also be used. The type of neural network can influence the accuracy of the determined fill level data 430 of the elevator car 110. Therefore, the accuracy of the determined fill level data 430 of the elevator car 110 can be optimized (e.g., increased) by selecting a neural network type that leads to high accuracy. The training of the pre-trained fill level classification model 420 will be described in more detail later in this application.
[0043] In step 330, the calculation unit 220 provides the determined elevator car 110 fill level data 430 to the elevator control system 130 for further elevator use. Further elevator use may include, for example, elevator call allocation and / or the presentation of information. For example, in elevator car allocation using the determined elevator car 110 fill level data 430, if the determined fill level data 430 indicates that the elevator car 110 is full, no further floor calls are allocated to the elevator car 110. In other words, the elevator control system 130 determines, based on the determined fill level data 430 received from the calculation unit 220 of the fill level determination system 200, that the elevator car 110, which is determined to be full, bypasses multiple floors between the departure floor and the destination floor, regardless of whether floor calls are generated from the bypassed floors. Using the determined fill level data 430 in elevator car allocation improves the elevator car allocation process. For example, in the presentation of information using the determined fill level data 430, the determined fill level data 430 may be displayed on one or more displays. The one or more displays may, for example, include at least one elevator lobby screen and / or at least one building manager interface. Displaying the determined fill level data 430 on one or more displays allows users of elevator system 100 (e.g., passengers) and / or building managers to receive information about the fill level of one or more elevator cars of elevator system 100. The determined fill level data 430 displayed on one or more displays may, for example, include the determined fill level data 430 in percentage form and / or as a schematic diagram generated based on the determined fill level data 430. Alternatively or additionally, the obtained image data 410 may be displayed on one or more displays. This allows users of elevator system 100 (e.g., passengers) and / or building managers to receive further information about the fill level of one or more elevator cars of elevator system 100.
[0044] As indicated by the name "Pre-trained Fill Level Classification Model," the fill level classification model 420 is trained before it can be used to determine the fill level data 430 of the elevator car 110. Next, by referring to... Figure 5 An example describing the method (i.e., the training method) used to train the filled level classification model 420. Figure 5 The training method is illustrated schematically as a flowchart. The pre-trained fill-level classification model 420 used to determine the fill-level data 430 of the elevator car 110 represents the fill-level classification model 420 trained using the training method. Training of the fill-level classification model 420 may at least include pre-training (i.e., training of the fill-level classification model 420 before using the pre-trained fill-level classification model 420 to determine the fill-level data 430 of the elevator car 110). Training of the fill-level classification model 420 may also include continuous training (i.e., training and / or refining the pre-trained fill-level classification model 420 during the process of using the model 420 to determine the fill-level data 430 of the elevator car 110). This can be achieved by using... Figure 5 The example training method performs both pre-training and continuous training.
[0045] At step 510, computational entity 800 obtains training input data including multiple images of the interior of at least one elevator car. The training input data may, for example, be optical image data including multiple optical images of the interior of at least one elevator car. Computational entity 800 may be computational unit 220 of fill level determination system 200. Alternatively, the computational entity may be an external computational entity outside of fill level determination system 200. Computational unit 220 may be communicatively coupled to computational entity 800. Communication between computational unit 220 and computational entity 800 may be based on one or more known wired or wireless communication technologies. Preferably, the training input data includes a large number of images of the interior of at least one elevator car. The training input data may include multiple images of the interior of the same elevator car 110, the fill level of which will be determined using a pre-trained fill level classification model 420. The training input data including multiple images of the interior of the same elevator car 110 may, for example, be generated by imaging device 210 of fill level determination system 200 arranged within said elevator car 110. Alternatively or additionally, the training input data may include multiple images of the interiors of one or more other substantially similar elevator cars. The training input data including multiple images of the interiors of one or more other substantially similar elevator cars may be generated, for example, by one or more imaging devices arranged within one or more other substantially similar elevator cars. Alternatively or additionally, the training input data may include multiple images of the interiors of one or more elevator cars generated using one or more artificial intelligence (AI) based image generators. The fill level classification model 420 is trained using training input data including multiple images of the interiors of the same elevator car 110, wherein a pre-trained fill level classification model 420 is used to determine the fill level, which improves the accuracy of the determined fill level data 430 of the elevator car 110. For example, the obtained image data 410 used to determine the fill level data 430 by applying the pre-trained fill level classification model 420 may be used alone or in conjunction with the determined fill level data 430 during continuous training of the pre-trained fill level classification model 420. Although training the fill level classification model 420 with training input data including multiple images of the interiors of one or more other substantially similar elevator cars can produce a general-function fill level classification model 420 that can be used to determine the fill level data of multiple substantially similar elevator cars with sufficient accuracy.
[0046] The fill level data 430 of elevator car 110 can be determined by applying a pre-trained fill level classification model 420, considering only human objects (e.g., passengers) within elevator car 110. Alternatively, the fill level data 430 of elevator car 110 can be determined by applying the pre-trained fill level classification model 420, considering both human and non-human objects (e.g., luggage, wheelchairs, and / or cargo) within elevator car 110. If only human objects are considered when determining the fill level data 430 of elevator car 110, the training input data can include multiple images of the interior of at least one elevator car, where only human objects are present. If both human and non-human objects are considered when determining the fill level data 430 of elevator car 110, the training input data can include multiple images of the interior of at least one elevator car, where human objects and / or non-human objects are present. Considering only human objects allows for a simpler pre-trained fill level classification model 420 and / or faster and easier training of the fill level classification model 420. Furthermore, if only human subjects are considered, the size of model 420 can be reduced and the determination speed of model 420 can be increased. Taking into account both human and non-human subjects can improve the accuracy of the determined fill level data 430.
[0047] In step 520, multiple images of at least one elevator car interior contained in the training input data are classified into multiple predetermined fill level levels. Each fill level level represents a different fill level of the elevator car 110. The different fill levels of the elevator car 110 can be expressed as percentages, where a 0% fill level represents an empty elevator car 110, and a 100% fill level represents a full elevator car 110. According to a non-limiting example, the multiple predetermined fill level levels may include 10 fill level levels. These 10 exemplary padding levels may include, for example, the following padding levels: Level 1: 0-9% padding, Level 2: 10-19% padding, Level 3: 20-29% padding, Level 4: 30-39% padding, Level 5: 40-49% padding, Level 6: 50-59% padding, Level 7: 60-69% padding, Level 8: 70-79% padding, Level 9: 80-89% padding, and Level 10: 90-100% padding. This is merely a non-limiting example of multiple predetermined padding levels, and multiple predetermined padding levels may include any other number of padding levels with any other distribution of padding levels. Each image belonging to the multiple images included in the generated training input data is associated with label data indicating the padding level class to which the image is classified. Multiple images associated with the label data are used in step 530 to train a fill level classification model 420 (as described later in this application), such that the pre-trained fill level classification model 420 can be used to determine a fill level corresponding to the fill level of the elevator car 110 in the image data 410 inside the elevator car 110. The label data can also indicate the presence of human and / or non-human objects within the elevator car in the corresponding image, i.e., whether human and / or non-human objects are present in the elevator car in the corresponding image. In addition to classification, one or more known preprocessing operations (e.g., sampling, transformation, denoising, etc.) can be performed on the training input data.
[0048] In step 530, computational entity 800 trains a fill level classification model 420 using training input data classified into multiple predetermined fill level levels to generate a pre-trained fill level classification model 420 for determining the fill level data 430 of the elevator car 110. If computational entity 800 is an external computational entity, it provides the pre-trained fill level classification model 420 to computational unit 220. Training step 530 may include one or more known training phases (e.g., initial training, validation, testing, etc.). After training the fill level classification model 420, computational unit 220 is able to determine the fill level level corresponding to the fill level of the elevator car 110 in the acquired image data 410 by applying the pre-trained fill level classification model 420. In other words, determining the fill level data 430 of the elevator car 110 in step 320 may include determining, by applying a pre-trained fill level classification model 420, the fill level level (belonging to multiple predetermined fill level levels) corresponding to the fill level of the elevator car 110 in the obtained image data 410.
[0049] The determined fill level data 430 of the elevator car 110 may include data indicating the determined fill level class. For example, the determined fill level data 430 of the elevator car 110 may include the determined fill level class and / or the fill level of the elevator car 110 represented by the determined fill level class. Figure 6 A non-limiting example is illustrated, showing how to determine the fill level data 430 of an elevator car 110 from acquired image data 410 by applying a pre-trained fill level classification model 420. Figure 6 In the example, it is assumed that the multiple predetermined fill level levels include the same ten predetermined fill level levels discussed in the non-limiting example above. Image data 410 of the interior of elevator car 110 indicates three person objects (e.g., passengers) 610 inside elevator car 110. Computation unit 220 applies a pre-trained fill level classification pattern 420 to determine the fill level level corresponding to the fill level of elevator car 110 in the acquired image data 410. Figure 6 In a non-limiting example, it is assumed that the fill level corresponding to the fill level of the elevator car 110 in the obtained image data 410 is fill level 3: 20-29% fill level, which indicates that the elevator car 110 is 20-29% full. Therefore, the determined fill level data 430 may include data indicating a fill level of 20-29%. In other words, in this example, the determined fill level data 430 may indicate that the elevator car 110 is 20-29% full.
[0050] The image analysis-based elevator car filling level determination solution based on the method and filling level determination system 200 described above enables the determination of the filling level of the elevator car 110 without calibration, which is a significant advantage compared to conventional image analysis-based elevator car filling level determination solutions that require calibration, as discussed in the background section. Furthermore, the image analysis-based elevator car filling level determination solution based on the method and filling level determination system 200 described above does not impose restrictions on the shape, size, wall material and color, or interior lighting of the elevator car 110. Therefore, the accuracy of the elevator car filling level determination solution is improved. Moreover, the elevator car filling level determination solution can be widely used for various types of elevator cars.
[0051] Figure 7An example of components of the computing unit 220 of the fill level determination system 200 is schematically shown. The computing unit 220 may include: a processing unit 710 including one or more processors, a memory unit 720 including one or more memories, a communication interface unit 730 including one or more communication devices, and possibly a user interface (UI) unit 740. The aforementioned components may be communicatively connected to each other, for example, via an internal bus. The memory unit 720 may store and maintain portions of a computer program (code) 725, a pre-trained fill level classification model 420, acquired image data 410, determined fill level data 430, and any other data. The computer program 725 may include instructions that, when executed by the processing unit 710 of the computing unit 220, cause the processing unit 710 and thus the computing unit 220 to perform desired tasks, such as one or more of the method steps described above and / or the operations of the computing unit 220 described above. Therefore, the processing unit 710 may be arranged to access the memory unit 720 and retrieve any information from and store any information in the memory unit 720. For clarity, the term "processor" herein refers to any unit suitable for processing information and controlling the operation of computing unit 220, as well as other tasks. These operations can also be implemented using a microcontroller solution with embedded software. Similarly, memory unit 720 is not limited to a particular type of memory, but any type of memory suitable for storing the described multiple pieces of information can be applied in the context of this invention. Communication interface unit 730 provides one or more communication interfaces for communicating with any other unit (e.g., imaging device 210, elevator control system 130, computing entity 800, one or more databases) and / or with any other unit. User interface unit 740 may include one or more input / output (I / O) devices for receiving user input and output information, such as buttons, keyboards, touchscreens, microphones, speakers, displays, etc. Computer program 725 may be a computer program product that can be included in a tangible non-volatile (non-transitory) computer-readable medium carrying computer program code 725 embodied therein for use with a computer (i.e., computing unit 220).
[0052] Figure 8An example of components of a computational entity 800 for performing training of a filled-level classification model 420 is schematically shown. The computational entity 800 may include: a processing unit 810 including one or more processors, a memory unit 820 including one or more memories, a communication interface unit 830 including one or more communication devices, and possibly a user interface (UI) unit 840. The aforementioned components may be communicatively connected to each other, for example, via an internal bus. The memory unit 820 may store and maintain portions of a computer program (code) 825, the pre-trained filled-level classification model 420, training input data, and any other data. The computer program 825 may include instructions that, when executed by the processing unit 810 of the computational entity 800, cause the processing unit 810 and thus the computational entity 800 to perform a desired task, such as one or more of the method steps and / or operations of the computational entity 800 described above. Therefore, the processing unit 810 may be arranged to access the memory unit 820 and retrieve any information from and store any information in the memory unit 820. For clarity, the term "processor" herein refers to any unit suitable for processing information and controlling the operation of computing entity 800, as well as other tasks. These operations can also be implemented using a microcontroller solution with embedded software. Similarly, memory unit 820 is not limited to a particular type of memory, but any type of memory suitable for storing the described multiple pieces of information can be applied in the context of this invention. Communication interface unit 830 provides one or more communication interfaces for communicating with any other unit (e.g., communication unit 220, imaging device 210, one or more other imaging devices, elevator control system 130, one or more databases) and / or with any other unit. User interface unit 840 may include one or more input / output (I / O) devices for receiving user input and output information, such as buttons, keyboards, touchscreens, microphones, speakers, displays, etc. Computer program 825 may be a computer program product that can be included in a tangible non-volatile (non-transitory) computer-readable medium carrying the computer program code 825 embodied therein for use with the computer (i.e., computing entity 800).
[0053] The specific examples provided in the description above should not be construed as limiting the applicability and / or interpretation of the appended claims. Unless otherwise expressly stated, the list and groups of examples provided in the description above are not exhaustive.
Claims
1. A method for determining fill level data (430) of an elevator car (110), wherein, The method includes: (310) Obtain image data (410) of the interior of the elevator car (110); By applying a pre-trained filling level classification model (420), the filling level data (430) of the elevator car (110) is determined (320) from the obtained image data (410); and The determined filling level data (430) of the elevator car (110) is provided (330) to the elevator control system (130) for further elevator use.
2. The method according to claim 1, wherein, Determining (320) the fill level data (430) of the elevator car (110) includes determining, from the obtained image data (410) a fill level belonging to a plurality of predetermined fill level levels by applying a pre-trained fill level classification model (420), the fill level levels corresponding to the fill level of the elevator car (110) in the obtained image data (410).
3. The method according to claim 2, wherein, The determined fill level data (430) of the elevator car (110) includes data indicating the determined fill level grade.
4. The method according to any one of the preceding claims, wherein, Further elevator use includes elevator call assignment and / or information presentation.
5. The method according to any one of the preceding claims, wherein, The pre-trained filled level classification model (420) is a model based on a convolutional neural network (CNN).
6. A filling level determination system (200) for determining filling level data (430) of an elevator car (110), wherein, The fill level determination system (200) includes: Imaging device (210), the imaging device being arranged inside the elevator car (110) and configured to generate image data (410) of the interior of the elevator car (110), and The computing unit (220) is configured as follows: Image data (410) of the interior of the elevator car (110) is obtained from the imaging device (210). By applying a pre-trained filling level classification model (420), the filling level data (430) of the elevator car (110) is determined from the obtained image data (410); and The determined filling level data (430) of the elevator car (110) is provided to the elevator control system (130) for further elevator use.
7. The fill level determination system (200) according to claim 6, wherein, Determining the fill level data (430) of the elevator car (110) includes: a computing unit (220) is configured to determine, by applying a pre-trained fill level classification model (420), a fill level belonging to a plurality of predetermined fill level levels from the obtained image data (410), the fill level levels corresponding to the fill level of the elevator car (110) in the obtained image data (410).
8. The fill level determination system (200) according to claim 7, wherein, The determined fill level data (430) of the elevator car (110) includes data indicating the determined fill level level.
9. The fill level determination system (200) according to any one of claims 6 to 8, wherein, Further elevator use includes elevator call assignment and / or information presentation.
10. The fill level determination system (200) according to any one of claims 6 to 9, wherein, The pre-trained filled level classification model (420) is a model based on a convolutional neural network (CNN).
11. An elevator system (100) comprising: At least one elevator car (110) is arranged to travel along a corresponding elevator shaft (120). An elevator control system (130) is configured to control the operation of the elevator system (100); and The fill level determination system (200) according to any one of claims 6 to 10.
12. A computer program (725) comprising instructions which, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 5.
13. A computer-readable medium comprising instructions that, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 5.
14. A method for training a fill-level classification model (420) for determining fill-level data (430) of an elevator car (110), the method comprising: Obtain training input data including multiple images of the interior of at least one elevator car; The plurality of images of the interior of the at least one elevator car are classified into a plurality of predetermined fill level levels; and The filling level classification model (420) for determining the filling level data (430) of the elevator car (110) is trained by using the training input data classified into the plurality of predetermined filling levels.
15. A computer program (825) comprising instructions which, when executed by a computer, cause the computer to perform the method according to claim 14.
16. A computer-readable medium including instructions that, when executed by a computer, cause the computer to perform the method according to claim 14.