JP2023179121A - elevator system - Google Patents

Abstract

To prevent error detection by a mirror, by setting a mirror area on a photographed image, without needing information concerning the mirror.SOLUTION: An elevator system according to one embodiment, in which a mirror is installed in an elevator car, is provided with photographing means, feature amount extracting means, mirror image block detecting means and mirror area generating means. The photographing means photographs a range including a landing near an entrance from the inside of the elevator car. The feature amount extracting means extracts feature amounts of an image photographed by the photographing means, per block unit. The mirror image block detecting means determines a pair of blocks that perform symmetrical movements in a pair of block images that performs symmetrical movements in the image, on the basis of a change in feature amounts of the image extracted by the feature amount extracting means and detects either of the pair of blocks as a mirror image block. The mirror area generating means generates a mirror area from an aggregate of the mirror image blocks detected by the mirror image block detecting means.SELECTED DRAWING: Figure 1

Description

Embodiments of the present invention relate to an elevator system that detects users using a camera installed in a car.

Conventionally, a camera is installed inside the elevator car, and by processing the images taken by this camera, it is possible to detect the number of passengers in the car, etc., and the detection results are sent to the elevator. A system that reflects this on driving control is known.

This type of system requires accurate detection of users through image processing. However, if a mirror is installed inside the car, the mirror may falsely detect a user, resulting in double counting of the number of passengers.

JP2020-145511A Patent No. 5230793

As a method for preventing the above-mentioned false detection caused by the mirror, a common method is to mask an area (mirror area) corresponding to the installation location of the mirror on the photographed image. However, the installation location of the mirror varies depending on the specifications of the car. Therefore, when using the above method, it is necessary to check the specifications of the car for each property, provide information regarding the mirror (installation location, etc.) to the system, and set the mirror area.

The problem to be solved by the present invention is to provide an elevator system that can set a mirror area on a photographed image without requiring information about mirrors and prevent false detection caused by mirrors.

An elevator system according to one embodiment is an elevator system in which a mirror is installed in a car, and includes an imaging means, a feature extraction means, a mirror image block detection means, and a mirror area generation means. The imaging means photographs a range from inside the car to a landing area near the entrance. The feature extracting means extracts the feature of the image obtained by the imaging means in units of blocks. The mirror image block detection means determines a pair of blocks having symmetrical motion in the image based on a change in the feature amount of the image extracted by the feature amount extraction means, and One is detected as a mirror image block. The mirror area generating means generates a mirror area from the aggregate of the mirror image blocks detected by the mirror image block detecting means.

FIG. 1 is a diagram showing the configuration of an elevator system according to an embodiment. FIG. 2 is a diagram showing the configuration of the area around the entrance and exit in the car in the same embodiment. FIG. 3 is a diagram showing an example of an image taken by the camera in the same embodiment. FIG. 4 is a diagram showing an example of an image of frame N in the same embodiment. FIG. 5 is a diagram showing an example of an image of frame N+1 in the same embodiment. FIG. 6 is a diagram showing an example of a change in brightness gradient between images in the same embodiment. FIG. 7 is a diagram showing examples of changes in brightness gradients, brightness values, and histograms between images in the same embodiment. FIG. 8 is a diagram showing an example of marking of a mirror image block in the same embodiment. FIG. 9 is a diagram for explaining a method of generating a mirror area in the same embodiment. FIG. 10 is a diagram for explaining another method of generating a mirror area in the same embodiment. FIG. 11 is a flowchart for explaining the operation of the elevator system in the same embodiment. FIG. 12 is a flowchart showing details of the mirror area generation process executed in step S16 of FIG. 11 above.

Hereinafter, embodiments will be described with reference to the drawings.
Note that the disclosure is merely an example, and the invention is not limited to the content described in the embodiments below. Modifications that can be easily conceived by those skilled in the art are naturally included within the scope of the disclosure. In order to make the explanation more clear, in the drawings, the size, shape, etc. of each part may be changed from the actual embodiment and shown schematically. In some drawings, corresponding elements are designated by the same reference numerals, and detailed description thereof may be omitted.

FIG. 1 is a diagram showing the configuration of an elevator system according to an embodiment. Note that although one car will be described as an example here, the configuration is the same for a plurality of cars.

A camera 12 used as an imaging means is installed above the entrance of the car 11. Specifically, the camera 12 is installed in a curtain plate 11a that covers the upper part of the entrance/exit of the car 11 with its lens portion facing directly below. The camera 12 has an ultra-wide-angle lens such as a fisheye lens, and photographs a wide range of objects including the inside of the car 11 with a viewing angle of 180 degrees or more. The camera 12 is capable of continuously capturing images at several frames per second (for example, 30 frames/second).

Note that the camera 12 does not need to be installed above the entrance of the car 11 as long as it is near the car door 13. For example, any location may be used as long as it is possible to photograph the entire interior of the car, including the entire floor surface of the car 11, such as the ceiling near the entrance of the car 11, and the landing area 15 near the entrance when the door is open.

At the landing 15 on each floor, a landing door 14 is installed at the arrival gate of the car 11 so as to be openable and closable. The landing door 14 engages with the car door 13 to open and close when the car 11 arrives. Note that the power source (door motor) is located on the side of the car 11, and the hall door 14 only opens and closes following the car door 13. In the following description, it is assumed that when the car door 13 is open, the landing door 14 is also open, and when the car door 13 is closed, the landing door 14 is also closed.

Each image (video) continuously captured by the camera 12 is analyzed in real time by the image processing device 20. In FIG. 1, the image processing device 20 is shown removed from the car 11 for convenience, but in reality, the image processing device 20 is housed together with the camera 12 in the curtain plate 11a.

The image processing device 20 includes a storage section 21 and a detection section 22. The storage unit 21 sequentially stores images taken by the camera 12 and has a buffer area for temporarily storing data necessary for processing by the detection unit 22. Note that the storage unit 21 may store images that have been subjected to processing such as distortion correction, enlargement/reduction, and partial cropping as preprocessing for the photographed images.

The detection unit 22 is composed of, for example, a microprocessor, and detects a user inside the car 11 or at the landing 15 using an image taken by the camera 12. The detection section 22 is functionally divided into a feature extraction section 22a, a mirror image block detection section 22b, a mirror area generation section 22c, and a detection processing section 22d. Note that these may be realized by software, hardware such as an IC (Integrated Circuit), or a combination of software and hardware. Alternatively, the elevator control device 30 may have some or all of the functions of the image processing device 20.

The feature amount extraction unit 22a reads out each image stored in the storage unit 21 in chronological order, and extracts feature amounts from these images in units of blocks. A "block unit" is a unit in which an image is divided into blocks of a predetermined size in a matrix. “Feature amount” refers to a numerical value that quantitatively represents the feature/characteristic of an image, and includes, for example, a brightness gradient, brightness value, histogram, and the like. The mirror image block detecting section 22b determines a pair of blocks having symmetrical motion in the image based on the change in the feature amount of the image extracted by the feature amount extracting section 22a. The mirror image block detection unit 22b detects one of the pair of blocks as a mirror image block based on the position of the car sill 47 (see FIG. 3) of the car 11 on the photographed image. The "mirror image block" refers to a block recognized as an image in the mirror 50 (see FIG. 3) installed inside the car 11. The mirror area generation unit 22c generates a mirror area from a collection of mirror image blocks detected by the mirror image block detection unit 22b.

The detection processing unit 22d detects the presence or absence of a user in the car 11 or in the landing 15 based on changes in the brightness of images obtained from the camera 12 in chronological order. Note that in addition to the user, objects such as wheelchairs may also be included in the detection target. At this time, the detection processing section 22d performs processing to prevent false detection of the user reflected in the mirror 50, such as masking the mirror area recognized by the mirror area generation section 22c.

The elevator control device 30 is composed of a computer including a CPU, ROM, RAM, and the like. The elevator control device 30 includes an operation control section 31, a door opening/closing control section 32, and a notification section 33. The operation control unit 31 controls the operation of the car 11. The door opening/closing control unit 32 controls opening/closing of the car door 13 when the car 11 arrives at the landing 15. Specifically, the door opening/closing control section 32 opens the car door 13 when the car 11 arrives at the landing 15, and closes the door after a predetermined period of time has elapsed.

Here, for example, when the detection processing unit 22d detects a user near the car door 13 during the door opening operation of the car door 13, the door opening/closing control unit 32 detects a door accident (drawing into the door pocket). Door opening/closing control is performed to avoid accidents. Specifically, the door opening/closing control unit 32 temporarily stops the door opening operation of the car door 13, moves it in the opposite direction (door closing direction), or slows down the door opening speed of the car door 13. The notification unit 33 alerts the users in the car 11 based on the detection result of the detection processing unit 22d.

FIG. 2 is a diagram showing the structure of the area around the entrance and exit in the car 11. As shown in FIG.
A car door 13 is provided at the entrance of the car 11 so as to be openable and closable. In the example of FIG. 2, a two-door, double-swing type car door 13 is shown, and two door panels 13a and 13b forming the car door 13 open and close in opposite directions along the frontage direction (horizontal direction). . Note that the "frontage" is the same as the entrance and exit of the car 11.

Front pillars 41a and 41b are provided on both sides of the entrance and exit of the car 11, and surround the entrance and exit of the car 11 together with the curtain plate 11a. The "front pillar" is also called an entrance pillar or an entrance frame, and a door pocket for storing the car door 13 is generally provided on the back side. In the example of FIG. 2, when the car door 13 is opened, one door panel 13a is stored in a door pocket 42a provided on the back side of the front pillar 41a, and the other door panel 13b is placed on the back side of the front pillar 41b. It is stored in the door pocket 42b.

A display 43, an operation panel 45 with a destination floor button 44, and a speaker 46 are installed on one or both of the front pillars 41a, 41b. In the example of FIG. 2, a speaker 46 is installed on the front pillar 41a, and a display 43 and an operation panel 45 are installed on the front pillar 41b.

Further, as shown in FIG. 3, a rectangular mirror 50 is installed on the back surface 49 of the car 11 at a location facing the entrance/exit. This mirror 50 is used, for example, as a rearview mirror when a wheelchair user exits the car 11. Note that the mirror 50 includes not only a "glass type" but also a "stainless steel mirror type."

A camera 12 having an ultra-wide-angle lens such as a fisheye lens is installed in the center of the curtain plate 11a above the entrance of the car 11. The camera 12 photographs the inside of the car 11 and the landing area 15 near the entrance at a predetermined frame rate (for example, 30 frames/second). The image taken by this camera 12 is given to the image processing device 20 shown in FIG. 1, and used for detection processing to detect a user or an object.

FIG. 3 is a diagram showing an example of an image taken by the camera 12. With the car door 13 (door panels 13a, 13b) and the landing door 14 (door panels 14a, 14b) fully open, photograph the entire car interior and the landing 15 near the entrance at a viewing angle of 180 degrees or more from above the entrance of the car 11. This shows the case where The upper side is the hall 15, and the lower side is inside the car 11.

At the landing 15, three-sided frames 17a and 17b are provided on both sides of the arrival gate for the car 11, and a strip-shaped landing sill 18 having a predetermined width is installed on the floor 16 between the three-sided frames 17a and 17b. 14 along the opening/closing direction. Further, a band-shaped car sill 47 having a predetermined width is disposed on the entrance/exit side of the floor surface 19 of the car 11 along the opening/closing direction of the car door 13.

Here, if the mirror 50 is installed inside the car 11, a problem arises in that the user reflected in the mirror 50 on the photographed image is erroneously detected, resulting in double counting of the number of passengers. In this case, false detection by the mirror 50 can be prevented by masking an area (mirror area) corresponding to the installation location of the mirror 50 on the photographed image. However, in order to set the mirror area, it is necessary to provide the system with information regarding the mirror 50 (installation location, etc.) for each property in advance, which is very time-consuming.

Below, a method for correctly setting a mirror area in a portion corresponding to the installation location of the mirror 50 on a captured image without requiring information regarding the mirror 50 will be described.
(a) Detecting symmetrical motion FIG. 4 is a diagram showing an example of an image of frame N, and FIG. 5 is a diagram showing an example of an image of frame N+1 (N is an arbitrary integer). P1 and P2 in the figure indicate users. The user P1 is riding in the car 11 at a position that is not reflected in the mirror 50 (in this example, near the side surface 48a). The user P2 is near the entrance of the car 11 and is about to get into the car 11. A mirror 50 is installed at a position facing the entrance/exit on the back side 49 of the car 11, and the user P2 is reflected in this mirror 50.

On the photographed image, the real image of the user P2 and the mirror image of the user P2 reflected in the mirror 50 move symmetrically. Therefore, in order to detect the position of the mirror 50, it is sufficient to detect a symmetrically moving portion on the photographed image. Specifically, a photographed image is divided into predetermined blocks, and a feature amount (for example, a brightness gradient) is extracted for each block. Then, the amount of change is obtained when the feature amounts of the image of frame N and the image of frame N+1 are compared for each block, and a pair of blocks having symmetrical motion is determined from the amount of change.

FIG. 6 is a diagram showing an example of a change in brightness gradient between images. Part A is the upper side of the image (the landing side), and part B is the lower side of the image (the back side of the car) (see FIGS. 4 and 5). The arrows in the figure indicate the direction of the brightness gradient.

Assume that in the image of frame N, the luminance gradient of block b (Xb, Yb) in portion A is 270°, and the luminance gradient of block k (Xk, Yk) in portion B is 90°. Assume that in the image of frame N+1, the luminance gradient of block b (Xb, Yb) in portion A is 225°, and the luminance gradient of block k (Xk, Yk) in portion B is 135°.

The amount of change in the brightness gradient of block b (Xb, Yb) is −45° (225°−270°). The amount of change in the brightness gradient of block k (Xk, Yk) is +45° (135°-90°). The amount of change in brightness gradient represents movement. Therefore, it can be seen that symmetrical movement occurs between block b and block k in the image. In this case, the difference when comparing the amount of change in the brightness gradient of block b and the amount of change in the brightness gradient of block k is approximately zero.

In the example shown in FIG. 6, the description focuses on block b in part A and block k in part B, but in reality, the brightness gradient is extracted from the entire image in units of blocks, and the captured image is determined from the difference in the amount of change. A pair of blocks having symmetrical motion is determined above, and one of the blocks in the pair is detected as a mirror image block. As will be described later, since the mirror 50 is often installed on the back surface 49 of the car 11, a block near the bottom of the photographed image is detected as a mirror image block with the position of the sill 40 as a reference.

As shown in FIG. 7, in addition to the brightness gradient, a brightness value or a histogram may be extracted as the feature amount, and a mirror image block may be detected based on the amount of change thereof. In the example of FIG. 7, "luminance gradient" is extracted as the feature amount 1, "luminance value" is extracted as the feature amount 2, and "histogram" is extracted as the feature amount 3. In this case, a value obtained by adding the amount of change in the brightness gradient, the amount of change in the brightness value, and the amount of change in the histogram according to a preset weighting value is used to detect the mirror image block.

Assuming that the difference in the amount of change in the brightness gradient extracted as feature 1 is "-θ _b ° - θ _k °" and the weighting value is "G _θ ", the score S _θ of feature 1 is calculated as follows. It will be done.
S _θ = (-θ _b ° - θ _k °) × G _θ
Assuming that the difference in the amount of change in the brightness value extracted as the feature amount 2 is "β _b - β _k " and the weighting value is "G _β ", the score S _β of the feature amount 2 is obtained as follows.
S _β = (β _b − β _k )×G _β
Assuming that the difference in the amount of change in the histogram extracted as the feature quantity 3 is "Hist _b -Hist _k " and the weighting value is "G _hist ", the score S _hist of the feature quantity 3 is obtained as follows.
S _hist = (Hist _b −Hist _k )×G _β
Here, G _θ >G _β >G _hist . The reason why the weighting value G _θ of the luminance gradient is set higher than the luminance value and the histogram is that the luminance gradient has directional properties and is most reliable when determining symmetrical movement. .

If the sum of the score S _θ of feature amount 1, the score S _β of feature amount 2, and the score S _hist of feature amount 2 is within a preset threshold TH1, block b and block k move symmetrically. It is judged as a pair of blocks with . The threshold value TH1 is set in consideration of the lighting environment of the car 11, and has an ideal value of 0±error. In other words, the change in the feature amount of block b and the change in the feature amount of block k are symmetrical, and the closer to zero the difference is when comparing the changes in the feature amount between the two, the more symmetrical the movement is judged to be. Ru.

In this way, by using the brightness value or histogram in addition to the brightness gradient, it is possible to more accurately determine a pair of blocks that have symmetrical movement in the photographed image. Alternatively, the determination may be made by adding feature amounts other than the brightness value and histogram.

(b) Generation of mirror area FIG. 8 is a diagram showing an example of marking of a mirror image block. FIG. 9 is a diagram showing an example of generation of a mirror area. "M" is a mark indicating that the block has been detected as a mirror image block. "Mb" is a mark indicating that the mirror image block has been determined as a part of the mirror.

When the landing 15 is placed on the upper side of the photographed image, the block existing below the car sill 47 among the pair of blocks having symmetrical movements detected in (a) above is detected as a mirror image block. This is because the mirror 50 is generally installed on the back surface 49 of the car 11 in many cases. If both of the pair of blocks are located below the car sill 47, the one closer to the bottom of the photographed image is detected as the mirror image block.

As shown in FIG. 8, marks M are attached to blocks detected as mirror image blocks. A mirror image block is detected for each image obtained in chronological order as a photographed image, and when the same block is detected as a mirror image block a certain number of times C1 or more (at least 2 times or more), the mirror image block is treated as a part of the mirror 50. It is confirmed and used to generate the mirror area ME. At this time, mark M is replaced with mark Mb. The reason for setting the number of times C1 or more is to exclude blocks that are irregularly erroneously detected as mirror image blocks due to the influence of noise or the like.

Here, as shown in FIG. 9, when the collection of mirror image blocks of the mark Mb occupies 75% or more and forms a rectangular area, this rectangular area is generated as the mirror area ME. The reason why the area is rectangular is that mirrors are generally rectangular in shape.

Another method is to expand the rectangular area to be searched on the condition that it includes mirror image blocks of mark Mb at a rate of 75% or more, and use the finally obtained rectangular area as the mirror area ME. You can also generate it.

A specific example is shown in FIG. Assume now that mirror image blocks indicated by “Mb1” to “Mb10” have been determined as part of the mirror 50.
First, a first rectangular area including a mirror image block of "Mb1" is searched. Next, a second rectangular area including mirror image blocks of "Mb2" and "Mb3" adjacent to the first rectangular area is searched. Similarly, a third rectangular area containing mirror image blocks of "Mb4", "Mb5" and "Mb6" adjacent to the second rectangular area → mirror image blocks of "Mb7" and "Mb8" adjacent to the third rectangular area. Expand the rectangular area to be searched under the above conditions, such as the fourth rectangular area containing → the fifth rectangular area containing the mirror image blocks of "Mb9" and "Mb10" adjacent to the fourth rectangular area... The fifth rectangular area finally obtained is defined as the mirror area ME.

If such a method is used, it is possible to detect if a block that is not recognized as a mirror image block is partially included, for example, if the image contains noise or if a part of the mirror 50 is dirty. It is also possible to correctly generate the mirror area ME.

Next, the operation of the elevator system in this embodiment will be explained.
FIG. 11 is a flowchart for explaining the operation of the elevator system. The processing shown in this flowchart is mainly executed by the image processing device 20.

Now, assume that the car 11 is on an arbitrary floor with its door open. A camera 12 installed inside the car 11 photographs the inside of the car 11 and the landing area 15 near the entrance at a predetermined frame rate. Images captured by the camera 12 are stored in the storage unit 21 of the image processing device 20 in chronological order.

The detection unit 22 of the image processing device 20 reads out each image stored in the storage unit 21 in chronological order (step S11). The detection unit 22 divides these images into predetermined block units, and extracts, for example, a brightness gradient as a feature quantity in each block unit (step S12). For each image, the detection unit 22 compares the amount of change in the feature amount in the photographed image on a block-by-block basis, and calculates the difference in the change in the feature amount (step S13).

Here, if a pair of blocks in which the difference in feature amount change is within the preset threshold TH1 is detected, the detection unit 22 determines the blocks as having symmetrical motion (step S14). The detection unit 22 detects, as a mirror image block, the block located at the lower side of the photographed image among the pair of blocks with reference to the position of the car sill 47 (step S15).

Specifically, as explained in FIG. 6, when comparing the amount of change in the brightness gradient of block b and the amount of change in the brightness gradient of block k between the image of frame N and the image of frame N+1, If the difference in the amount of change is within the threshold TH1, block b and block k are determined to be a pair of blocks having symmetrical movements. In this case, since block k exists below the photographed image, it will be detected as a mirror image block.

In this way, when some mirror image blocks recognized as part of the mirror 50 are detected from the photographed image, the detection unit 22 generates a mirror area ME on the photographed image from the aggregation of the mirror image blocks ( Step S16). FIG. 12 shows details of the mirror area generation process executed in step S16.

The detection unit 22 marks a block detected as a mirror image block on the photographed image (step S21). At this time, the detection unit 22 counts the number of times each block is detected as a mirror image block (step S22). When the same block is detected as a mirror image block a certain number of times C1 or more (Yes in step S23), the detection unit 22 determines the mirror image block as a part of the mirror 50 (step S24), and detects the determined mirror image. A mirror area ME is generated on the photographed image from the collection of blocks (step S25).

In detail, as explained in FIG. 9, when a rectangular area is formed by a set of mirror image blocks (Mb) determined as a part of the mirror 50 occupying a certain proportion or more, the edge of the rectangular area is A mirror area ME is generated based on this. Alternatively, as explained in FIG. 10, the rectangular area to be searched is expanded on the condition that mirror image blocks (Mb) are included in a certain proportion or more, and the search is performed based on the edges of the finally obtained rectangular area. Generate it as a mirror area ME.

Once the mirror area ME is generated on the photographed image, the detection unit 22 masks the mirror area ME and performs user detection processing (step S17). Thereby, without erroneously detecting the users reflected in the mirror 50, it is possible to perform detection processing only on existing users and accurately count, for example, the number of users on board. In the example of FIG. 5, when user P2 gets into the car 11 when the door is opened, the mirror image of user P2 reflected in the mirror 50 is ignored, and the current It is possible to correctly count the number of passengers as two people.

As described above, according to the present embodiment, the mirror area ME is set in the part where the mirror 50 is installed, focusing on the part having symmetrical movement on the photographed image. Therefore, without checking the installation location of the mirror 50 for each property, it is possible to correctly set the mirror area ME on the photographed image and prevent false detection by the mirror 50.

In the above embodiment, the case where the door of the car 11 is open has been described as an example, but the same applies even when the door of the car 11 is closed. That is, by detecting one of a pair of blocks having symmetrical movement as a mirror image block from an image taken by the camera 12 while the door is closed, it is possible to generate a mirror area ME from a collection of the mirror image blocks. .

Further, even when the mirror 50 is installed on the side surface 48a or the side surface 48b, the mirror area ME can be generated using the same method as described above. In this case, when a pair of blocks having symmetrical movement is determined from the photographed image, a block located below the car sill 47 and close to the left or right end of the photographed image is detected as a mirror image block, The mirror area ME may be generated from the collection of mirror image blocks.

Furthermore, in order to increase the likelihood of the detection target, the amount of speed change (strength) of the feature amount may be added. The “amount of speed change in feature amount” refers to the speed at which the feature amount changes between frame images. In other words, if there is a pair of blocks whose feature values change at the same speed, one of the blocks is detected as a mirror image block.

According to at least one embodiment described above, it is possible to provide an elevator system that can set a mirror area on a captured image and prevent false detection caused by mirrors without requiring information regarding mirrors.

Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention, as well as within the scope of the invention described in the claims and its equivalents.

11... Car, 11a... Curtain board, 12... Camera, 13... Car door, 14... Landing door, 15... Landing area, 16... Landing floor, 17a, 17b... Three side frame, 18... Landing sill, 19... Riding area Floor surface of the car, 20... Image processing device, 21... Storage unit, 22... Detection unit, 22a... Feature extraction unit, 22b... Mirror image block detection unit, 22c... Mirror area generation unit, 22d... Detection processing unit, 30... Elevator control device, 31... Operation control unit, 32... Door opening/closing control unit, 33... Notification unit, 41a, 41b... Front pillar, 41a-1, 41b-1... Inner side surface of front pillar, 42a, 42b... Door pocket, 43 ...Display, 44...Destination floor button, 45...Operation panel, 46...Speaker, 47...Car sill, 48a, 48b...Side, 49...Back, 50...Mirror, ME...Mirror area.

Claims (10)

In an elevator system where a mirror is installed inside the car,
an imaging means for photographing a range from inside the car to a landing area near the entrance;
Feature extracting means for extracting feature quantities of the image photographed by the imaging means in block units;
A pair of blocks having symmetrical motion is determined in the image based on a change in the feature amount of the image extracted by the feature amount extraction means, and one of the pair of blocks is detected as a mirror image block. Mirror image block detection means;
An elevator system comprising: mirror area generating means for generating a mirror area from an aggregate of the mirror image blocks detected by the mirror image block detecting means. The mirror image block detection means is
2. The elevator system according to claim 1, wherein one of the pair of blocks is detected as the mirror image block based on the position of the sill of the car on the photographed image. The mirror image block detection means is
2. The elevator system according to claim 1, wherein a pair of blocks in which a difference is within a preset threshold value when a change in feature amount of the image is compared on a block-by-block basis is detected. The mirror area generation means described above is
2. The elevator system according to claim 1, wherein when the same block is detected as the mirror image block a certain number of times or more in the image, the block is used to generate the mirror area. The mirror area generation means described above is
2. The elevator system according to claim 1, wherein when the mirror image blocks occupy a certain proportion or more to form a rectangular area, the rectangular area is generated as the mirror area. The mirror area generation means described above is
2. The elevator according to claim 1, wherein the rectangular area to be searched is expanded on the condition that the mirror image block is included in a proportion equal to or higher than a certain level, and a finally obtained rectangular area is generated as the mirror area. system. The elevator system according to claim 1, wherein the feature amount includes a brightness gradient. 2. The elevator system according to claim 1, wherein the feature amount includes a brightness gradient, a brightness value, and a histogram. 9. The elevator system according to claim 8, wherein weighting values are set for each of the brightness gradient, the brightness value, and the histogram. 10. The elevator system according to claim 9, wherein a weighting value for the brightness gradient is set higher than the brightness value and the histogram.

Images