JP2022178688A - Mechanical parking lot - Google Patents

Abstract

To provide a relatively inexpensive mechanical parking lot capable of further relatively checking safety in a getting on/off room at the time of entering/exiting.SOLUTION: In a mechanical parking lot, a vehicle that has entered the getting on/off room can be stored in a parking room. The mechanical parking lot is equipped with an imaging device for imaging the inside of the getting-on/off room, a tracking unit 120 for generating tracking information for associating positions of persons in a plurality of images with each other between the plurality of images imaged by the imaging device and a determination unit 122 that determines whether or not to start a closing operation to close an entering/exiting door opened for entrance/exit based on the tracking information.SELECTED DRAWING: Figure 3

Description

The present invention relates to mechanical parking lots.

A mechanical parking lot is known as a parking lot that can efficiently park a large number of vehicles in a small space. A mechanical parking lot generally consists of a loading/unloading compartment for drivers to get into and out of vehicles, a parking compartment provided on the ground or underground, and a lifting device for moving the vehicle between the loading/unloading compartment and the parking compartment. , provided.

JP 2006-077425 A

In the mechanical parking lot, when closing the entrance/exit door of the entrance/exit room opened for entrance/exit, it is confirmed that there is no one in the entrance/exit room.

For example, in order to confirm the safety in the passenger compartment, a method of installing a scanner-type laser sensor in the passenger compartment is conceivable. However, in this method, in order to prevent erroneous detection, walls, pallets, and vehicles, which are moving bodies, must be excluded from the detection range, and a person outside such a detection range cannot be detected.

Further, for example, a method of installing a plurality of wide sensors in the passenger compartment is conceivable. However, with this method, it is necessary to install a wide sensor around the entire circumference of the passenger compartment in order to reduce blind spots, which is not practical in terms of dealing with other equipment.

Further, for example, a method of installing a passive sensor for detecting temperature-moving objects in the passenger compartment is conceivable. However, this method cannot detect a person sitting still in the passenger compartment.

Further, for example, a method of installing LiDAR (Light Detection and Ranging, Laser Imaging Detection and Ranging) in the passenger compartment is conceivable. However, LiDAR is expensive and disadvantageous in terms of cost.

SUMMARY OF THE INVENTION The present invention has been made in view of such circumstances, and one of the exemplary objects of certain embodiments thereof is to provide a relatively inexpensive mechanical parking lot in which the safety of the passenger compartment can be confirmed more reliably when entering and exiting the parking lot. to do.

In order to solve the above problems, a mechanical parking lot according to one aspect of the present invention is a mechanical parking lot that can store a vehicle that has entered the boarding/alighting room in the parking room, comprising: an imaging device that images the inside of the boarding/alighting room; A tracking unit that generates tracking information that associates the positions of people in the plurality of images among multiple images captured by the imaging device, and whether or not to start the closing operation of closing the entrance/exit door opened for entering/exiting. based on the tracking information.

Arbitrary combinations of the above constituent elements, and mutual substitution of the constituent elements and expressions of the present invention in methods, devices, systems, etc. are also effective as aspects of the present invention.

ADVANTAGE OF THE INVENTION According to this invention, the relatively cheap mechanical parking lot which can confirm the safety in the boarding/alighting room|chamber more reliably at the time of entering/exiting can be provided.

It is a sectional view of a mechanical parking lot concerning a 1st embodiment. FIG. 2 is a plan view of the passenger compartment of FIG. 1; 2 is a block diagram showing the functions and configuration of a control device in FIG. 1; FIG. It is a figure explaining a tracking method. FIG. 10 is a diagram illustrating tracking when at least part of an imaging area moves across imaging areas of two imaging devices that are different from each other; It is a figure explaining tracking when there is a blind spot. It is a flowchart which shows an example of operation|movement of the mechanical parking lot at the time of warehousing. It is a flowchart which shows an example of operation|movement of the mechanical parking lot at the time of leaving. It is a block diagram which shows the function and structure of the control apparatus of the mechanical parking lot which concerns on 2nd Embodiment. FIGS. 10(a) to 10(c) are diagrams showing the situation at the time of warehousing in chronological order. FIGS. 11(a) to 11(d) are diagrams showing the situation at the time of warehousing in chronological order. FIG. 11 is a block diagram showing the functions and configuration of a control device according to a third embodiment; FIG. FIG. 12 is a block diagram showing the functions and configuration of a control device according to a fourth embodiment; FIG. FIGS. 12(a) to 12(c) are diagrams for explaining a method for specifying the presence or absence of a difference from a difference image, taking the case of warehousing as an example. FIG. 12 is a block diagram showing the functions and configuration of a control device according to a fifth embodiment; FIG.

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described below based on preferred embodiments with reference to the drawings. The same or equivalent constituent elements, members, and processes shown in each drawing are denoted by the same reference numerals, and duplication of description will be omitted as appropriate. Moreover, the embodiments are illustrative rather than limiting the invention, and not all features and combinations thereof described in the embodiments are necessarily essential to the invention.

(First embodiment)
FIG. 1 is a front view of a mechanical parking lot 10 according to the first embodiment. The mechanical parking lot 10 is an underground mechanical parking lot in which a parking room is provided in the basement of a building 104 provided on the ground. The mechanical parking lot 10 may be a ground-type mechanical parking lot in which a parking room is provided in the building 104 .

The mechanical parking lot 10 includes a B1F parking room 14 that is a parking room on the first basement floor, a B2F parking room 16 that is a parking room on the second basement floor, a B3F parking room 18 that is a parking room on the third basement floor, and a plurality of parking spaces. A pallet 32, a first transfer device 56, a plurality of second transfer devices 58, a lift frame 88, and two or four pillars provided as pillars along a hoistway 96 having a substantially rectangular cross section. A mast 90 , a lifting device 94 , an operation panel 28 provided outside the passenger compartment 12 , and a control device 100 connected to the operation panel 28 .

The mechanical parking lot 10 is a pallet-type parking lot using pallets 32 . In this pallet-type mechanical parking lot 10, a pallet 32 on which vehicles such as automobiles 20, motorcycles 74, and bicycles are mounted can be raised and lowered using a lifting device 94, or a first transfer device 56 and a second transfer device 56 can be used. The vehicle is parked in the parking room by moving it in the plane direction using the transfer device 58 .

A loading/unloading room 12 for putting a vehicle into the mechanical parking lot 10 and taking it out from the mechanical parking lot 10 is provided in the building 104 . The passenger compartment 12 is provided at the upper end of the hoistway 96 . A communication hole 98 is provided in the boarding/alighting compartment 12 to communicate the boarding/alighting compartment 12 and the hoistway 96 .

The hoistway 96 is provided along the vertical direction (z direction). The hoistway 96 communicates with the B1F parking room 14, the B2F parking room 16, and the B3F parking room 18, respectively. A lifting device 94 for lifting the lift frame 88 is installed at the lower end of the hoistway 96 . For example, four masts 90 are provided at the four corners of the hoistway 96 .

The lift frame 88 is supported by four masts 90 so as to be vertically movable, and is vertically moved along a hoistway 96 by a lifting device 94 . A first transfer device 56 is provided above the lift frame 88 . For the lift frame 88, the four masts 90, and the lifting device 94, the lift frame, the four masts, and the lifting device described in Japanese Unexamined Patent Application Publication No. 2009-197417 previously filed by the present applicant are used, respectively. can be applied.

The pallet 32 is a flat member. A vehicle loading surface 32a, which is the upper surface of the pallet 32, is substantially rectangular. Pallet 32 has a uniform size.

B1F parking room 14, B2F parking room 16, and B3F parking room 18 each include a plurality of parking spaces. Each parking space is configured to accommodate a pallet 32 . A parking space is a parking area (unit of parking) within a parking compartment in which a vehicle can be parked with a pallet.

The first transfer device 56 is provided above the lift frame 88 and moves the pallet 32 in the x direction. When there are two masts 90, the first transfer device 56 further moves the pallet 32 in the y direction. The first transfer device 56 moves the pallet between the hoistway 96 and each parking room. The second transfer device 58 is provided in each parking space and moves the pallet 32 in the x direction or the y direction. The second transfer device 58 moves the pallet between parking spaces. For the first transfer device 56 and the second transfer device 58, the transfer device described in Japanese Patent Publication No. 7-29681 previously filed by the present applicant can be applied.

The operation panel 28 functions as a user interface for the users of the mechanical parking lot 10 or the attendants. The operation panel 28 has a touch panel. The operation panel 28 is connected to the control device 100 and transmits control signals to the control device 100 according to operations input to the touch panel.

FIG. 2 is a plan view of the passenger compartment 12 of the mechanical parking lot 10. As shown in FIG. The boarding/alighting compartment 12 is surrounded by walls around the periphery (side surfaces and ceiling surfaces) except for the entrance/exit opening 12a, and an entrance/exit door 26 is provided at the entrance/exit opening 12a. An imaging device 11 is installed in the passenger compartment 12 . The imaging device 11 is a video camera, and images the interior of the passenger compartment 12 at a predetermined cycle to generate a frame image for each cycle. Although only one imaging device 11 is installed in this example, the number of imaging devices 11 is not particularly limited, and a plurality of imaging devices 11 may be installed, for example, to reduce blind spots.

FIG. 3 is a block diagram showing the functions and configuration of the control device 100. As shown in FIG. The control device 100 includes an instruction receiving unit 110, a warehousing processing unit 112, a warehousing processing unit 114, an image acquisition unit 116, a detection unit 118, a tracking unit 120, a determination unit 122, a model storage unit 124, A tracking information storage unit 126 is provided.

The image acquisition unit 116 acquires frame images from the imaging device 11 at a predetermined imaging speed (for example, 60 frames/second).

The detection unit 118 detects (a region of) a person from the frame image acquired by the image acquisition unit 116 . In this embodiment, the detection unit 118 realizes detection of a person using a person detection model M stored in the model storage unit 124 .

The human detection model M uses a large amount of training data including learning images showing people (head, arms, legs, etc., or the whole body) and labels indicating what is contained in the images. A trained model generated by machine learning. Learning of the human detection model M can be performed using various well-known machine learning techniques. Examples of machine learning include deep learning. A specific example of deep learning is error backpropagation using a neural network.

When the detection unit 118 inputs the frame image captured by the imaging device 11 to the human detection model M, the human detection model M outputs the human detection information regarding the frame image. Here, the human detection information includes at least information indicating whether or not a person (including only a part of the head, arms, legs, etc.) is detected, and if a person is detected, it includes the position of the person. good.

Note that the detection unit 118 is not limited to detection using the human detection model M (learned model), and may perform detection using any known algorithm.

The tracking unit 120 tracks the person detected by the detection unit 118 from the frame image. Specifically, the tracking unit 120 identifies the same person among the persons detected in each of the plurality of frame images. The tracking unit 120 associates the position information in the passenger compartment 12 in each frame image of the person identified as the same person in each frame image with each other, generates tracking information, and stores it in the tracking information storage unit 126 . When the detection unit 118 detects multiple people, each of the multiple people is tracked. In this case, multiple pieces of tracking information are stored in the tracking information storage unit 126 .

The detection unit 118 and the tracking unit 120 execute respective processes each time the image acquisition unit 116 acquires a frame image, for example, when warehousing or delivery is started. In other words, the person in the passenger compartment 12 is detected and tracked substantially in real time. The detection unit 118 and the tracking unit 120 may perform respective processes by thinning frame images in order to realize substantially real-time human detection and tracking. In addition, it is not limited to real-time human detection and tracking.

The instruction receiving unit 110 receives instructions (signals) transmitted from the operation panel 28 .

When the vehicle is to be parked, the driver or a staff member stops the vehicle to be parked outside the passenger compartment 12 and exits the vehicle, and inputs an instruction to start parking on the operation panel 28 . Upon receiving the instruction to start warehousing, the operation panel 28 generates a warehousing start signal and transmits it to the control device 100 . When the instruction receiving unit 110 receives the warehousing start signal, the warehousing processing unit 112 controls the lifting device 94, the first transfer device 56, and the second transfer device 58 to move the empty pallet to the loading/unloading room 12. A door opening/closing mechanism (not shown) is controlled to open the loading/unloading door 26 . The warehousing processing unit 112 guides the vehicle to be warehousing to a stop position on the pallet 32 arranged in the boarding/alighting room 12 by means of voice or a vehicle guide light.

After the vehicle enters the boarding/alighting compartment 12 and stops on the pallet, the driver or staff member gets off the vehicle and inputs an instruction to start storing to the operation panel 28. - 特許庁Upon receiving the storage start instruction, the operation panel 28 generates a storage request signal and transmits it to the control device 100 .

When the instruction receiving unit 110 receives the storage request signal, the determining unit 122 determines whether or not to start closing the loading/unloading door 26 based on the tracking information stored in the tracking information storage unit 126 .

If all the tracking information indicates that a person has left the boarding/alighting compartment 12, the determination unit 122 determines that there is no person in the boarding/alighting compartment 12, so that the closing operation of the entrance/exit door 26 can be started. I judge. "Tracking information indicates that a person has left the boarding/alighting room 12" means, for example, that the last positional information in the time-series positional information included in the tracking information is located outside the entrance/exit 12a. or, for example, even if the time-series positional information included in the tracking information ends the positional information indicating the vicinity of the entrance/exit 12a in the passenger compartment 12. good.

If any of the tracking information indicates that a person is in the boarding/alighting compartment 12, the determination unit 122 determines that it is "impossible" to start the closing operation of the loading/unloading door 26 because a person is in the boarding/alighting compartment 12. judge.

When the determination result by the determination unit 122 is "possible", the warehousing processing unit 112 controls the door opening/closing mechanism to close the entrance/exit door 26, and furthermore, the lifting device 94, the first transfer device 56, and the second transfer device A device 58 is controlled to move the pallet on which the vehicle is mounted from the passenger compartment 12 to a predetermined parking space. When the determination result by the determination unit 122 is “impossible”, the warehousing processing unit 112 displays an error on the touch panel of the operation panel 28 without closing the entrance/exit door 26 or moving the pallet 32 .

When leaving the vehicle, the driver or a staff member inputs the vehicle ID identifying the vehicle to be delivered to the control panel 28 . The operation panel 28 generates an exit start signal including the input vehicle ID and transmits the signal to the control device 100 . When the instruction receiving unit 110 receives the delivery start signal, the delivery processing unit 114 controls the lifting device 94, the first transfer device 56, and the second transfer device 58, and specifies the vehicle by the vehicle ID included in the delivery start signal. A pallet loaded with a vehicle to be loaded is moved from a parking space to a boarding/alighting room 12, and an entry/exit door 26 is opened by controlling a door opening/closing mechanism.

A driver or an attendant takes the vehicle out of the passenger compartment 12 and stops the vehicle outside the passenger compartment 12, and inputs an exit instruction to the operation panel 28. - 特許庁When the operation panel 28 receives the instruction to finish leaving the garage, the operation panel 28 generates a closing request signal and transmits it to the control device 100 .

When the instruction receiving unit 110 receives the door closing request signal, the determination unit 122 determines whether or not to start the closing operation of the loading/unloading door 26 based on the tracking information stored in the tracking information storage unit 126 . If all the tracking information indicates that the person has left the boarding/alighting room 12 or the person has gotten into the vehicle to be unloaded, the determination unit 122 determines that the closing operation of the entrance/exit door 26 can be started. I judge. If any of the tracking information indicates that a person is in the boarding/alighting compartment 12, the determination unit 122 determines that the closing operation of the loading/unloading door 26 cannot be started.

When the determination result by the determination unit 122 is “possible”, the retrieval processing unit 114 controls the door opening/closing mechanism to close the entrance/exit door 26 . When the determination result by the determination unit 122 is “impossible”, the retrieval processing unit 114 displays an error on the touch panel of the operation panel 28 without closing the entrance/exit door 26 .

Subsequently, tracking by the tracking unit 120 will be described in more detail. Note that the tracking by the tracking unit 120 is not limited to the method described below, and various known methods can be used.

FIG. 4 is a diagram illustrating an example of a tracking method. Frame images F _t , F _t+1 , and F _t+2 are three frame images temporally continuous in this order. People H _t , H _t+1 , and H _t+2 in frame images F _t , F _t+1 , and F _t+2 are people detected by the detection unit 118 . Note that the persons H _t , H _t+1 and H _t+2 are the same person. The image S is an image in which three frame images F _t , F _t+1 and F _t+2 are superimposed.

In this example, for ease of understanding, the positions of the persons H _t , H _t+1 , and H _t+2 are separated so that they do not overlap, but the time difference between temporally consecutive frame images is, for example, 0.1 seconds or less. Yes, the amount of movement of the person at this time is very small, and in fact the persons _Ht , Ht ₊₁ , _Ht+2 appear to overlap at least partially.

A case will be described below where the person _Ht in the frame image _Ft and the person Ht ₊₁ in the frame image Ft ₊₁ are identified as the same person, and the same person is identified from the frame image Ft ₊₂ . The position information of the person H _t and the person H _t+1 identified as the same person are associated with each other and stored in the tracking information storage unit 126 as tracking information.

The tracking unit 120 identifies the positional information of the person detected in the frame image F _t+2 in the passenger compartment 12 . The position information may be, for example, coordinates in an XY coordinate system with any corner in the passenger compartment 12 as the origin. The tracking unit 120 may specify the position information of the person detected in the frame image F _t+2 from the position of the person's feet in the frame image F _t+2 based on, for example, a known image analysis technique.

Subsequently, the tracking unit 120 identifies the same person between the frame images based on the direction of movement, for example. Specifically, the tracking unit 120 identifies the moving direction D _t+1 from the position of the person H _t to the position of the person H _t+1 for the person H _t and the person H _t+1 who are identified as the same person. Next, the tracking unit 120 identifies the moving direction D _{t+ 2} from the position of the person H _t+1 to the position of the person H _t+2 for the person H t+2 who is the same person candidate. The tracking unit 120 identifies the person H _t+1 and the person H _t+2 as the same person if the acute angle between the moving direction D t+1 and the moving direction D _{t+ 2} is equal to or less than a predetermined threshold angle α.

Also, for example, the tracking unit 120 identifies the same person between frame images based on the amount of movement. Specifically, the tracking unit 120 specifies the distance between the person H _t+1 and the person H _{t+2 who} is the same person candidate as the person H _{t+1, that is, the movement amount L when the person H t+1} and the person H _t+2 are assumed to be the same person. . The tracking unit 120 identifies the person H _t+1 and the person H _t+2 as the same person if the movement amount L is equal to or less than the threshold movement amount. The threshold movement amount may be determined based on the frame rate of the imaging device 11, for example.

Further, for example, the tracking unit 120 may identify the same person between frame images based on both the direction of movement and the amount of movement (that is, vector).

The tracking unit 120 stores the position information of the person H _t+2 in the tracking information storage unit 126 in association with the position information of the person H _t and the person H _t+1 identified as the same person.

FIG. 5 shows tracking when moving from the imaging area of one imaging device 11 toward the imaging area of the other imaging device 11 so that at least a part of the imaging region straddles the imaging regions of the two imaging devices 11 that are different from each other. It is a figure explaining about. Here, it is assumed that the imaging areas of the two imaging devices 11 are all different from each other (that is, do not overlap).

A frame image F1 _i (i=t, t+1, t+2) is a frame image captured by one of the two imaging devices 11, and F2 _i is captured by the other of the two imaging devices 11. It is a frame image captured by the device 11 . The frame images F1 _t , F1 _t+1 , and F1 _t+2 are three frame images temporally continuous in this order. Frame images F2 _t , F2 _t+1 , and F2 _t+2 are three frame images temporally continuous in this order. The frame image F1 _i and the frame image F2 _i are frame images captured substantially at the same time.

The image S1 is an image obtained by overlapping three frame images F1 _t , F1 _t+1 and F1 _t+2 . The image S2 is an image obtained by overlapping the three frame images F2 _t , F2 _t+1 and F2 _t+2 . It is assumed that the person H1 _t based on the frame image F1 _t and the person H1 _t+1 based on the frame image F1 _t+1 are identified as the same person.

Identification of the same person when straddling imaging regions is basically the same as identification of the same person when not straddling imaging regions, and can be identified from at least one of the direction of movement and the amount of movement.

For example, when the same person is identified based on the direction of movement, the tracking unit 120 detects the person H1 t+1 in the frame image F1 _t+1 and the person H2 _{t +2} in the frame image F2, which is the candidate for the same person, from the position of the person H1 _t+1 to the person H1 _t+2. Identify the direction of movement D _t+2 to the position of . The tracking unit 120 identifies the person H _t+1 and the person H _t+2 as the same person if the acute angle formed by the moving direction D t+1 and the moving direction D _{t+ 2} is equal to or less than a predetermined threshold angle.

Further, for example, when identifying the same person based on the amount of movement, the tracking unit 120 determines the distance between the person H1 _t+1 in the frame image F1 _t +1 and the person H2 _t+2 in the frame image F2, which is the candidate for the same person, that is, the distance between the person H1 t+1 and the person H1 _t+1 . Specify the amount of movement when H2 _t+2 is the same person. The tracking unit 120 identifies the person H1 _t+1 and the person H2 _t+2 as the same person if the movement amount is equal to or less than the threshold movement amount.

In this way, the same person can be identified with relatively high accuracy by identifying the same person when moving so as to straddle two imaging areas that do not overlap each other.

FIG. 6 is a diagram for explaining tracking when there is a blind spot. Here, a case where two imaging devices 11_1 and 11_2 are installed in the passenger compartment 12 will be described, but the same explanation applies to a case where one or three or more imaging devices are installed.

6 is a plan view of the passenger compartment 12. FIG. Broken lines in FIG. 6 indicate imaging regions of the imaging devices 11_1 and 11_2. The tracking unit 120 divides the inside of the passenger compartment 12 into a plurality of areas and manages them. In this example, the inside of the passenger compartment 12 is divided into A11 to A55 for management. When the person detected by the frame image of either the imaging device 11_1 or the imaging device 11_2 is lost, that is, the person is removed from the imaging area of either the imaging device 11_1 or the imaging device 11_2, and is no longer detected from any frame image. Consider the case of In this case, the tracking unit 120 estimates the area in which the lost person is located from the position information and moving direction of the person in the frame image before losing sight (for example, immediately before losing sight). In the illustrated example, it is estimated that the user is in area A33 (blind spot) based on the position information and moving direction when the user is in area A24. Then, when a person is detected in an area near (for example, adjacent to) the estimated area, the tracking unit 120 identifies the same person as the lost person.

Next, the operation of the mechanical parking lot 10 configured as described above will be described.

FIG. 7 is a flow chart showing an example of the operation of the mechanical parking lot 10 at the time of parking. A driver or an attendant stops the vehicle to be entered in front of the entry/exit door 26.例文帳に追加A driver or a staff member inputs an instruction to start warehousing to the operation panel 28 . The operation panel 28 accepts a warehousing start instruction (S10). The detection unit 118 and the tracking unit 120 start human detection and tracking based on the frame image (S12). The warehousing processing unit 112 places an empty pallet in the passenger compartment 12 (S14), opens the entry/exit door 26 (S16), and guides the vehicle to the pallet 32 in the passenger compartment 12 (S18).

A driver or attendant parks the vehicle on the pallet 32 . After confirming that there is no one in the passenger compartment 12, the driver or a staff member inputs a retraction start instruction to the operation panel 28. - 特許庁The operation panel 28 accepts a storage start instruction (S20). Based on the tracking information stored in the tracking information storage unit 126, the determination unit 122 determines whether or not the closing operation of the loading/unloading door 26 can be started (S22). If the determination result is "impossible" (N of S22), for example, an error is displayed on the touch panel of the operation panel 28, and the process returns to S20. If the determination result is "possible" (Y of S22), the warehousing processing unit 112 closes the entrance/exit door 26 (S24) and moves the pallet 32 on which the vehicle is mounted to the parking room (S26). The detection unit 118 and the tracking unit 120 end human detection and tracking (S28).

FIG. 8 is a flow chart showing an example of the operation of the mechanical parking lot 10 when leaving the parking lot. A driver or a staff member inputs a delivery start instruction to the operation panel 28 . The operation panel 28 accepts the delivery start instruction (S40). The detection unit 118 and the tracking unit 120 start human detection and tracking based on the frame image (S42). The retrieval processing unit 114 moves the pallet 32 on which the vehicle specified by the management number included in the retrieval start instruction is loaded from the parking room to the boarding/alighting room 12 (S44), opens the entrance/exit door 26 (S46), and moves the vehicle. is guided outside the passenger compartment 12 (S48).

A driver or attendant pulls the vehicle out of the passenger compartment 12 and stops the vehicle outside the passenger compartment 12 . After confirming that no one is in the boarding/alighting room 12, the driver or a staff member inputs a door closing instruction to the operation panel 28. - 特許庁The operation panel 28 receives a door closing instruction (S50). Based on the tracking information stored in the tracking information storage unit 126, the determination unit 122 determines whether or not the closing operation of the loading/unloading door 26 can be started (S52). If the determination result is "impossible" (N of S52), for example, an error is displayed on the touch panel of the operation panel 28, and the process returns to S50. If the determination result is "possible" (Y of S52), the retrieval processing unit 114 closes the entrance/exit door 26 (S54), and the detection unit 118 and the tracking unit 120 terminate human detection and tracking (S56).

7 and 8 are merely examples, and it is possible to add other processes, delete some processes, or change the order of the processes.

Next, the effects of this embodiment will be described. According to this embodiment, the presence or absence of a person in the boarding/alighting compartment 12, and thus whether or not the closing operation of the entrance/exit door 26 opened for entering/exiting the garage can be started, is determined based on the tracking information of the detected person. be done. For example, even if there is a blind spot in the boarding/alighting room 12 and the person who was once detected moves into the blind spot and is no longer detected even though he/she has not left the boarding/alighting room 12, the tracking information indicates that the person is not detected. It can be determined that the closing operation cannot be started because there is a person in the room 12 . That is, according to the present embodiment, it is possible to more reliably determine whether or not there is a person in the passenger compartment 12, and thus whether or not the closing operation should be started. In other words, it is possible to confirm the safety in the passenger compartment 12 at the time of entry and exit. Moreover, since an expensive sensor is not required, it can be realized at a relatively low cost.

Further, according to the present embodiment, a person in the passenger compartment 12 is detected based on the image captured by the imaging device 11 . As a result, detection is possible in a three-dimensional space, not two-dimensionally as in a scanner-type laser sensor or a wide sensor, so that a three-dimensional detection area is formed and blind spots in detection can be minimized. In addition, it is possible to detect a person in a stationary state that cannot be detected by a passive sensor, such as an infant or a person lying down, regardless of age, posture, or movement of the person.

(Second embodiment)
In the second embodiment, by counting the number of people who exited or boarded the vehicle in the passenger compartment and the number of people who entered and exited the passenger compartment, it is determined whether there is a person in the passenger compartment and whether the closing operation can be started. do. The following description will focus on differences from the first embodiment.

FIG. 9 is a block diagram showing the functions and configuration of the control device 200 according to the second embodiment. The control device 200 includes an instruction receiving unit 110, a warehousing processing unit 112, a warehousing processing unit 114, an image acquisition unit 116, a detection unit 118, a determination unit 222, a people counting unit 228, and a model storage unit 124. , and a number storage unit 229 .

The people counting unit 228 includes a getting off people counting unit 230 , a boarding people counting unit 232 , an entering people counting unit 234 , and an exiting people counting unit 236 .

The alighting number counting unit 230 counts the number of people who alighted from the vehicle stopped in the passenger compartment 12 (hereinafter referred to as "alighting number"). More specifically, the exiting number counting unit 230 counts the number of exiting persons when it is specified that the person detected by the detecting unit 118 has exited (came out of) the vehicle based on a known image analysis technique for the frame image. Up. Note that, for example, it is conceivable that a person who got out of the vehicle gets into the vehicle and then gets out of the vehicle again. Considering such a case, the alighting number counting unit 230 counts down the number of alighting people when it is specified that a person has boarded the vehicle.

The number-of-passengers counting unit 232 counts the number of passengers (hereinafter referred to as "passengers") who boarded the vehicle parked in the passenger compartment 12 at the time of leaving the garage. Specifically, the passenger count unit 232 counts the number of passengers when the person detected by the detection unit 118 is identified as having boarded (entered) the vehicle based on a known image analysis technique for the frame image. Up. Note that, for example, it is conceivable that a person who has gotten into the vehicle gets out of the vehicle and gets into the vehicle again. In consideration of such a case, the passenger count unit 232 counts down the number of passengers when it is specified that a person has gotten off the vehicle.

For example, the alighting number counting unit 230 and the boarding number counting unit 232 detect that a person suddenly appears near the vehicle (that is, the detection unit 118 detects a person when the door of the vehicle that is parked in the passenger compartment 12 is open). ), the person may be identified as exiting the vehicle. Further, for example, the alighting number counting unit 230 and the boarding number counting unit 232 detect that a person suddenly disappears in the vicinity of the vehicle while the door of the vehicle that is stopped in the passenger compartment 12 is open (i.e. is no longer detected), the person may be identified as having entered the vehicle.

The entering number counting unit 234 counts the number of persons who entered the boarding/alighting room 12 on foot (hereinafter referred to as "entering number of persons") at the time of entering and exiting. More specifically, the entering number counting unit 234 counts up the number of entering persons when it is specified that the person detected by the detecting unit 118 has walked into the passenger compartment 12 based on a known image analysis technique for frame images. .

The exiting number counting unit 236 counts the number of persons who walked out of the passenger compartment 12 (hereinafter referred to as "exiting number of persons") at the time of entering and exiting the parking lot. Specifically, the exiting person counting unit 236 counts up the number of exiting persons when it is specified that the person detected by the detecting unit 118 walked into the boarding/alighting room 12 based on a known image analysis technique for the frame image. .

If there are multiple entry/exit entrances, such as when there are multiple entry/exit entrances or when there is an emergency exit in addition to the entry/exit exit, the room entry count unit 234 counts the number of people entering the room through all of these exits. The number-of-exiting-persons counting unit 236 may count the number of exits from all of these entrances.

The number of people counted by each of the alighting number counting unit 230 , the boarding number counting unit 232 , the entering number counting unit 234 , and the exiting number counting unit 236 is stored in the number storage unit 237 .

The number of persons counted by the number of persons counting unit 228 is summarized as follows.
(1) When entering the parking lot, the exiting number counting section 230, the entering number counting section 234, and the exiting number counting section 236 respectively count the number of exiting persons, the entering number, and the exiting number.
(2) When exiting the parking lot, the number of passengers counting unit 232, the number of people entering the room counting unit 234, and the number of people leaving the room counting unit 236 count the number of passengers, the number of people entering the room, and the number of people leaving the room, respectively.

The detection unit 118 and the people counting unit 228 execute respective processes each time the image acquisition unit 116 acquires a frame image, for example, when warehousing or delivery is started. In other words, the number of people in the boarding/alighting room 12 is detected and counted substantially in real time. The detection unit 118 and the people counting unit 228 may perform respective processes by thinning frame images in order to realize substantially real-time people detection and people counting. In addition, it is not limited to substantially real-time detection of people and counting of people, and for example, they may be executed collectively when the determination unit 122 makes a determination.

When the following formula (1) is established at the time of entering the vehicle, the determination unit 222 determines that there is no person in the passenger compartment 12 and that the closing operation of the loading/unloading door 26 can be started. If not, it is determined that the closing operation of the loading/unloading door 26 cannot be started because there is a person in the passenger compartment 12 .
(Number of people getting off) + (Number of people entering) = (Number of people leaving) (1)

FIGS. 10(a) to 10(c) are diagrams showing the situation at the time of warehousing in chronological order. In FIG. 10(a), four people who entered the boarding/alighting compartment 12 got off the vehicle, and one person entered the boarding/alighting compartment 12 on foot. In FIG. 10(b), three people have left the boarding/alighting compartment 12, and two people still remain in the boarding/alighting compartment 12. In FIG. In FIG. 10(c), all five people have left the boarding/alighting compartment 12, and no one remains in the boarding/alighting compartment 12. In FIG.

At the timing of FIG. 10(b), the number of people getting off the train: 4, the number of people entering the room: 1, and the number of people leaving the room: 3, so the equation (1) does not hold. Therefore, at the timing of FIG. 10B, the determination unit 222 determines that the closing operation of the loading/unloading door 26 cannot be started. At the timing of FIG. 10(c), the number of people getting off the train: 4, the number of people entering the room: 1, and the number of people leaving the room: 5, and the formula (1) holds. Therefore, at the timing of FIG. 10(c), the determination unit 222 determines that the closing operation of the loading/unloading door 26 can be started.

Returning to FIG. 9, the determining unit 222 determines that the closing operation of the loading/unloading door 26 can be started because there is no person in the passenger compartment 12 when the following formula (2) is established at the time of leaving the garage. If the following equations do not hold, it is determined that the entry/exit door 26 cannot be closed because there is a person in the passenger compartment 12 .
(Number of people entering) = (Number of passengers) + (Number of people leaving) (2)

FIGS. 11(a) to 11(d) are diagrams showing the situation at the time of leaving the garage in chronological order. In FIG. 11(a), five people have entered the passenger compartment 12, and four of them have boarded the vehicle. In FIG. 11(b), a vehicle with four people has left the passenger compartment 12 and one person remains in the passenger compartment 12. In FIG. In FIG. 11(d), the last person has left the boarding/alighting room, and no one remains in the boarding/alighting room 12. In FIG.

At the timing of FIG. 11(c), the number of people entering the room is 5, the number of passengers is 4, and the number of people leaving is 0, so the equation (2) does not hold. Therefore, at the timing of FIG. 11(c), the determination unit 222 determines that the closing operation cannot be started. In the example of FIG. 11(d), the number of people entering the room: 5, the number of passengers: 4, and the number of people leaving the room: 1, and formula (2) is established. Therefore, at the timing of FIG. 11(d), the determination unit 222 determines that the closing operation can be started.

The operation at the time of entering the mechanical parking lot of the present embodiment is basically the same as the operation at the time of entering the parking lot in the first embodiment. should be read as "start of human detection and counting of people", and "end of human detection and tracking" in S28 should be read as "end of human detection and counting of people".

The operation when leaving the mechanical parking lot in this embodiment is basically the same as the operation when leaving the parking lot in the first embodiment. should be read as "start of human detection and counting people", and "end of human detection and tracking" in S56 should be read as "end of human detection and counting people".

Next, the effects of this embodiment will be described. According to the present embodiment, the presence or absence of people in the passenger compartment 12, and furthermore, whether or not the closing operation of the door 26 opened for entering and exiting the car can be started depends on the number of people getting off or on board, the number of people entering and leaving the car. It is determined based on the number of people. For example, even if there is a blind spot in the boarding/alighting room 12 and a person who was once detected moves into the blind spot and is no longer detected even though he/she has not left the boarding/alighting room 12, the result of counting the number of people will be different. Therefore, it can be determined that the closing operation cannot be started because there is a person in the passenger compartment 12 . That is, according to the present embodiment, it is possible to more reliably determine whether or not there is a person in the passenger compartment 12, and thus whether or not the closing operation should be started. In other words, it is possible to confirm the safety in the passenger compartment 12 at the time of entry and exit. Moreover, since an expensive sensor is not required, it can be realized at a relatively low cost.

Here, the number of people in the vehicle is counted from the captured image of the vehicle instead of the number of people getting off and the number of people getting on, and the presence or absence of people in the boarding/alighting room 12 is determined based on the number of people in the vehicle, the number of people entering the room, and the number of people leaving the room. Think about the case. In this case, for example, a person inside the vehicle may enter a blind spot in a part other than the window such as the vehicle housing. In such a case, there is a possibility that the count may be omitted, and as a result, it may not be possible to correctly determine whether or not there is a person in the passenger compartment 12 . On the other hand, according to the present embodiment, the presence or absence of people in the passenger compartment 12 is determined based on the number of people getting off or boarding without any possibility of counting omission. safety can be confirmed.

(Third Embodiment)
In the third embodiment, an ID is assigned to a person detected in the passenger compartment. At the time of entry, it is determined whether there is a person in the passenger compartment and whether the closing operation can be started or not, based on whether or not the person with the ID has left the room. At the time of exiting, it is determined whether or not a person with an ID has left the room or has entered the vehicle, and whether or not there is a person in the passenger compartment, and thus whether or not the closing operation can be started. The following description will focus on differences from the first embodiment.

FIG. 12 is a block diagram showing the functions and configuration of the control device 300 according to the third embodiment. The control device 300 includes an instruction receiving unit 110, a warehousing processing unit 112, a warehousing processing unit 114, an image acquisition unit 116, a detection unit 118, a determination unit 322, a feature extraction unit 338, an ID assignment unit 340, An updating unit 342 and a feature storage unit 344 are provided.

The feature extractor 338 extracts features of the person detected by the detector 118 from the frame image. Features to be extracted are, for example, bone structure, face shape, clothing, and the like. The feature extraction unit 338 may extract a feature that uniquely identifies a person, and may extract one type of feature or multiple types of features. For example, the feature extraction unit 338 may extract only skeleton information, or may extract skeleton information and clothing information.

The ID assigning unit 340 assigns an ID to the person detected by the detecting unit 118 . Specifically, the ID assignment unit 340 stores the features extracted by the feature extraction unit 338 in the feature storage unit 344 . Since the feature information itself is unique data, only the feature information may be stored in the feature storage section assuming that the feature information itself is an ID, or the feature information may be stored in a unique ID (for example, ID001, ID002, . . . ). ) and stored in the feature storage unit 344 . If the feature of the person detected by the detection unit 118 and extracted by the feature extraction unit 338 is the same as the feature stored in the feature storage unit 344, the person has already been given an ID. It will be. In this case, no new ID is given. In other words, the extracted features are not stored in the feature storage unit 344 .

When the updating unit 342 identifies that the person detected by the detecting unit 118 has left the room based on a known image analysis technique for the frame image, the updating unit 342 deletes the feature information of the person from the feature storage unit 344. .

When the updating unit 342 determines that the person detected by the detecting unit 118 has entered or left the vehicle based on a known image analysis technique for the frame image, the updating unit 342 stores the characteristic information of the person. Delete from section 344 .

The detecting unit 118, the feature extracting unit 338, the ID assigning unit 340, and the updating unit 342 execute respective processes each time the image acquiring unit 116 acquires a frame image, for example, when warehousing or warehousing is started. That is, the person in the boarding/alighting room 12 is detected substantially in real time, the feature is extracted, and the ID is assigned or the ID is deleted. The detection unit 118, the feature extraction unit 338, the ID assignment unit 340, and the update unit 342 may thin out the frame images and execute respective processes in order to realize substantially real-time detection and counting of people. Note that the detection unit 118, the feature extraction unit 338, the ID assignment unit 340, and the update unit 342 are not limited to executing these processes substantially in real time, and for example, when the determination unit 122 determines may be executed.

The determination unit 322 determines that no person is in the boarding/alighting room 12 at the time of entry if all the persons to whom IDs have been assigned have left the room, in this example, if no record remains in the storage unit. , it is determined that the closing operation of the loading/unloading door 26 can be started. If there is a person to whom an ID has been assigned who has not left the room, in this example, if the characteristic information of any person remains in the characteristic storage unit 344, the determination unit 322 determines whether there is a person in the passenger compartment 12. Therefore, it is determined that the closing operation of the loading/unloading door 26 cannot be started.

At the time of exiting the vehicle, if all the persons with IDs have entered or left the vehicle, in this example, if no feature information remains in the feature storage section 344, the judgment section 322 Since no one is present, it is determined that the closing operation of the entrance/exit door 26 can be started. If there is a person who has not gotten into the vehicle and has not left the room among the persons to whom the ID has been assigned, the determination unit 322 determines that the characteristic information of any person remains in the characteristic storage unit 344 in this example. If there is, it means that there is a person in the boarding/alighting room 12, so it is determined that the closing operation of the loading/unloading door 26 cannot be started.

The operation at the time of entering the mechanical parking lot of the present embodiment is basically the same as the operation at the time of entering the parking lot in the first embodiment. should be read as "start of human detection, feature extraction, ID assignment and ID deletion", and "end of human detection and tracking" in S28 should be read as "end of human detection, feature extraction, ID assignment and ID deletion". .

The operation when leaving the mechanical parking lot in this embodiment is basically the same as the operation when leaving the parking lot in the first embodiment. should be read as "start of human detection, feature extraction, ID assignment and ID deletion", and "end of human detection and tracking" in S56 should be read as "end of human detection, feature extraction, ID assignment and ID deletion". .

According to this embodiment, the same effects as those of the first embodiment can be obtained.

(Fourth embodiment)
In the fourth embodiment, by comparing a predetermined reference image of the inside of the passenger compartment with an image of the passenger compartment at the desired point of time, it is possible to determine whether or not there is a person in the passenger compartment and thus whether or not the closing operation of the entrance/exit door can be started. judge. The following description will focus on differences from the first embodiment.

FIG. 13 is a block diagram showing the functions and configuration of the control device 400 of the fourth embodiment. The control device 100 includes an instruction receiving unit 110 , a warehousing processing unit 112 , a warehousing processing unit 114 , an image acquisition unit 116 , and a determination unit 422 .

The determination unit 422 compares a reference image (described later) of the interior of the passenger compartment 12 with an image of the interior of the passenger compartment 12 at the point of time to be determined (hereinafter referred to as a “determination image”) to identify the presence or absence of a difference. If there is no difference between the two, the determination unit 422 determines that there is no one in the passenger compartment 12 and that the closing operation of the loading/unloading door 26 can be started. If there is a difference between the two, the determination unit 422 determines that the closing operation of the loading/unloading door 26 cannot be started because there is a possibility that a person is in the passenger compartment 12 .

The reference image at the time of entering is a state in which the vehicle to be entered is parked on the pallet 32 arranged in the passenger compartment 12, and no people are in the passenger compartment 12 except for the people in the vehicle to be entered. 12 is an image captured inside. The reference image at the time of entry is, for example, an image of the interior of the passenger compartment 12 captured when the vehicle is guided into the passenger compartment 12 and stopped at the stop position on the pallet 32 .

The reference image at the time of exit is an image of the interior of the passenger compartment 12 in a state in which an empty pallet 32 is placed in the passenger compartment 12 and no vehicle or person is in the passenger compartment 12 . The reference image at the time of delivery may be prepared in advance, for example.

The determination unit 422 may identify the presence or absence of a difference from a difference image obtained by taking the difference between the determination image and the initial image. FIGS. 14(a) to 14(c) are diagrams for explaining a method of identifying the presence or absence of a difference from a difference image, taking the case of warehousing as an example. FIG. 14(a) shows a reference image. FIG. 14(b) shows a determination image. In the judgment image of FIG. 14(b), there is a person in the passenger compartment 12. FIG. 14(c) shows a difference image obtained by taking the difference between the judgment image and the reference image. If there is a difference between the reference image and the judgment image, the difference image remains in the difference image as shown in FIG. 14(c). The determination unit 422 may specify the presence or absence of a difference based on the presence or absence of this difference portion.

According to this embodiment, the same effects as those of the first embodiment can be obtained.

(Fifth embodiment)
In the fifth embodiment, a learned model that has been learned to output information regarding whether or not to start closing operation of the entrance/exit door when an image of the inside of the passenger compartment is input is used. The following description will focus on differences from the first embodiment.

FIG. 15 is a block diagram showing the functions and configuration of a control device 500 according to the fifth embodiment. The control device 500 includes an instruction receiving unit 110 , a warehousing processing unit 112 , a warehousing processing unit 114 , an image acquisition unit 116 , a determination unit 522 , and a model storage unit 546 .

The model storage unit 546 stores a warehousing model M1 and a warehousing model M2.

The model M1 at the time of entry is a photographed image in the passenger compartment 12 in a state where the vehicle is stopped at a predetermined stop position on the pallet 32 arranged in the passenger compartment 12, and the presence or absence of a person in the photographed image (in other words, the photographed image It is a trained model generated by machine learning using a plurality of teacher data including a label indicating whether or not a person is in the image. A captured image of the interior of the passenger compartment 12 in a state in which the vehicle is stopped at a predetermined stop position is input to the model M1 at the time of entry. The model M1 at the time of warehousing outputs the determination result of the presence/absence of a person in the passenger compartment 12 .

The exit model M2 is generated by machine learning using a plurality of teacher data including captured images of the passenger compartment 12 with empty pallets 32 arranged therein and labels indicating the presence or absence of people in the captured images. It is a trained model that has been trained. A captured image of the interior of the passenger compartment 12 in which an empty pallet 32 is arranged (the vehicle is not stopped) is input to the leaving model M2. The leaving model M2 outputs the determination result of whether or not there is a person in the passenger compartment 12 .

At the time of warehousing, the determination unit 522 uses the warehousing model M1 stored in the model storage unit 546 to determine whether or not the closing operation of the entrance/exit door 26 can be started. Specifically, at the timing when it is desired to determine whether or not to start the closing operation of the entrance/exit door 26 (for example, at the timing when the instruction receiving unit 110 receives the storage request signal), the captured image of the inside of the passenger compartment 12 is captured by the entrance model M1. Enter to The determination unit 522 acquires the determination result output from the model M1 at the time of warehousing. The determination unit 522 determines that the closing operation of the entrance/exit door 26 can be started when the determination result indicates that there is no person in the passenger compartment 12 . If the determination result indicates that there is a person in the passenger compartment 12, the determination unit 522 determines that the closing operation of the loading/unloading door 26 cannot be started.

At the time of leaving, the determination unit 522 determines whether or not to start the closing operation of the entrance/exit door 26 using the model M2 at the time of leaving stored in the model storage unit 546 . Specifically, a captured image of the inside of the passenger compartment 12 at the timing at which it is desired to determine whether or not to start the closing operation of the entrance/exit door 26 (for example, the timing at which the instruction receiving unit 110 receives the door closing request signal) is sent to the exit model M2. input. The determining unit 522 acquires the result of determining whether or not there is a person in the passenger compartment 12 output from the exit model M2. If the determination result of the presence or absence of a person output from the exit model M2 indicates that there is no person in the passenger compartment 12, the determination unit 522 determines that the closing operation of the entrance/exit door 26 can be started, If the determination result indicates that there is a person in the boarding/alighting room 12, it is determined that the closing operation of the loading/unloading door 26 cannot be started.

The model M1 at the time of entering the warehouse and the model M2 at the time of leaving the warehouse may be learned so as to determine whether or not the closing operation of the entering/exiting door 26 can be started. In this case, the determination unit 522 may use the determination result of the model M1 at the time of entering the warehouse and the model M2 at the time of leaving the warehouse as the determination result of whether or not to start the closing operation of the entrance/exit door 26 .

According to the present embodiment, the learning model is used to determine whether or not there is a person in the passenger compartment 12, and thus whether or not to start closing the loading/unloading door 26. Therefore, the determination accuracy is high.

The configuration and operation of the mechanical parking lot according to the embodiment have been described above. It should be understood by those skilled in the art that this embodiment is an example, and that various modifications can be made to the combination of each component, and that such modifications are within the scope of the present invention.

(Modification 1)
You may combine 1st Embodiment and 2nd Embodiment. In this case, the number of people getting off, the number of people getting on, the number of people entering the room, and the number of people leaving the room may be counted based on the tracking information. According to this modified example, these numbers can be counted more accurately.

(Modification 2)
You may combine 1st Embodiment and 3rd Embodiment. In this case, it is possible to more accurately identify whether or not the person detected in each frame image is the same person. For example, even when at least part of the imaging area moves across the imaging areas of two imaging devices 11 different from each other, it is possible to more accurately identify whether the person is the same person. That is, according to this modified example, tracking can be performed more accurately.

(Modification 3)
You may combine 2nd Embodiment and 3rd Embodiment. In this case, since an ID is assigned to each person to be counted, it is possible to prevent the same person from being counted twice. According to this modified example, the number of people can be counted more accurately.

(Modification 4)
Whether or not the closing operation of the entrance/exit door 26 opened for loading/unloading can be started may be determined by any two or more methods of the first to fifth embodiments. In this case, for example, when the determination results of all the methods are "possible", the start of the closing operation of the loading/unloading door 26 may be "possible".

(Modification 5)
When it is determined that the closing operation of the entrance/exit door 26 opened for entering/leaving is "possible", the warehousing processing unit 112 and the warehousing processing unit 114 automatically close without waiting for an instruction from the driver or a staff member. The entrance/exit door 26 may be closed. In this case, the burden on the driver or attendant is reduced. In addition, forgetting to close the entrance/exit door 26 is eliminated.

(Modification 6)
Unlike the embodiment, detection of a person in the passenger compartment 12 may be performed even after the entrance/exit door 26 is closed. For example, a person may be detected all the time. In this case, the control device may notify the management room of the parking lot, the service center of the parking lot, or the like that a person has been detected.

(Modification 7)
In the embodiment and the modified example described above, the case where it is impossible to start the closing operation of the loading/unloading door 26 opened for loading/unloading when there is a person in the passenger compartment 12 has been described. Animals such as animals and external objects may be treated in the same way as people. That is, even when an animal or an external object exists in the boarding/alighting compartment 12, the opening/closing operation of the entrance/exit door 26 may be disabled.

(Modification 8)
In the above-described embodiment and modifications, an example in which the mechanical parking lot is a pallet-type parking lot using pallets has been described, but the present invention is not limited to this. The mechanical parking lot may be, for example, a parking lot that does not use a pallet, such as a slat conveyor type, or a system that transfers vehicles to elevators or storage racks by passing comb teeth. This modified example has the same actions and effects as the above-described embodiment and modified example.

Any combination of the embodiments and modifications described above is also useful as an embodiment of the present invention. A new embodiment resulting from the combination has the effects of the combined embodiment and modifications.

10 mechanical parking lot, 11 imaging device, 12 boarding/alighting room, 26 entrance/exit door, 30 control device, 120 tracking section, 122,222,322,422,522 determination section.

Claims (10)

A mechanical parking lot that can store a vehicle that has entered the boarding and alighting room in the parking room,
an imaging device that captures an image of the inside of the passenger compartment;
A tracking unit that generates tracking information that associates the positions of people in the plurality of images among the plurality of images captured by the imaging device;
a determination unit that determines whether or not to start a closing operation for closing the entrance/exit door opened for entering/exiting based on the tracking information;
A mechanical parking lot with 2. The mechanical parking lot according to claim 1, wherein the determination unit determines that the closing operation cannot be started when the person has not left the boarding/alighting room at the time of entering the parking lot. The determining unit determines that the closing operation cannot be started when the person has not boarded the vehicle to be retrieved from the passenger compartment and has not left the passenger compartment at the time of leaving the garage. 3. The mechanical parking lot according to Item 1 or 2. Another imaging device that captures an image of the inside of the boarding/alighting room and has at least a part of an imaging region different from that of the imaging device,
The tracking unit performs tracking from a person detected from a first image captured by the imaging device and a second image captured by the another imaging device that is temporally later than the first image. 4. The machine according to any one of claims 1 to 3, wherein whether or not the detected person is the same person is specified based on at least one of the movement direction and movement amount of the person detected from the first image. formal parking lot. Another imaging device that captures an image of the inside of the boarding/alighting room and has at least a part of an imaging region different from that of the imaging device,
The tracking unit detects a moving direction of a person detected from a first image captured by the imaging device and a second image captured by the another imaging device, which is temporally later than the first image. If an acute angle formed by the direction of movement of the person detected from the second image is equal to or less than a predetermined threshold angle, the person detected from the first image and the person detected from the second image are the same person. A mechanical parking lot according to any one of claims 1 to 3 to be specified. 6. The determination unit according to any one of claims 1 to 5, wherein when the number of people detected in the passenger compartment does not match the number of people who left the passenger compartment, the determination unit determines that the closing operation cannot be started. The mechanical parking lot described in . When the number of people who entered the boarding/alighting room at the time of exiting the car does not match the number of people who entered the vehicle to be exited from the boarding/alighting room and the number of people who left the boarding/alighting room, 6. The mechanical parking lot according to any one of claims 1 to 5, wherein it is determined that the operation cannot be started. further comprising an ID assigning unit that assigns an ID to the person detected from the image captured by the imaging device;
8. The mechanical type according to any one of claims 1 to 7, wherein the determination unit determines that the closing operation cannot be started when an ID-given person who has not left the room exists at the time of entry. parking. further comprising an ID assigning unit that assigns an ID to the person detected from the image captured by the imaging device;
2. The determining unit determines that the closing operation cannot be started if, at the time of exiting the parking lot, there is a person who has not gotten into the vehicle and has not left the room among the persons to whom the ID has been assigned. 7. The mechanical parking lot according to any one of 7. A learned model generated by machine learning using teacher data including a learning image that is an image of the inside of the passenger compartment, wherein when an image of the passenger compartment captured by the imaging device is input, the closing operation is performed. further comprising a storage unit that stores a trained model that outputs information about whether or not to start,
The mechanical parking lot according to any one of claims 1 to 9, wherein the determination unit determines whether or not to start the closing operation based on an output from the learned model.

Images