JP2013096092A - Vehicle protrusion detection device and method for mechanical parking facility - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect protrusion from a pallet with high precision by accurately extracting a contour of a vehicle from an image picked up in a getting-on/off room of the mechanical parking facilities.SOLUTION: A difference image acquisition part 62 which calculates an inter-frame difference image from two frames picked up at a predetermined interval. An initial contour setting part 68 ORs the inter-frame difference image calculated by the difference image acquisition part 62 from the start of entry of a vehicle to a stop on a pallet, and determines an outer peripheral shape of the ORed image as an initial contour. A contour extraction part 72 extracts a contour of the vehicle in a frame by executing a dynamic contour model for extracting a contour of a body included in the initial contour by converging the initial contour inward according to energy minimization principle. A comparison determination part 74 compares the extracted contour of the vehicle with a reference frame previously set as a frame representing a vehicle stop range on the pallet, and then determines whether the vehicle protrudes depending on the contour of the vehicle being within the reference frame.

Description

  The present invention relates to a technique for detecting the protrusion of a vehicle from a pallet in a mechanical parking lot.

  Generally, a mechanical parking lot is known as a parking lot that can efficiently park a large number of vehicles in a small space. As one of the mechanical parking lots, a pallet type is configured to place a vehicle on a pallet and move it to the vacant pallet space by moving the pallet back and forth or left and right using rails and grooves. There are known mechanical parking lots.

  The area of the pallet is slightly smaller than the opening area of the floor in the passenger compartment. For this reason, if the vehicle is stopped in a state of being deviated from a predetermined position on the pallet, the vehicle may come into contact with a component of the parking lot, for example, the edge of the floor opening of the passenger compartment when the pallet moves. Therefore, it is necessary to detect whether or not the vehicle has stopped at a predetermined position on the pallet, and when it deviates from the predetermined position, it is necessary to notify the driver or staff.

  Patent Document 1 discloses a mechanical parking facility on a transporter having image processing means for performing image processing on an image obtained by capturing a predetermined area including a part of a vehicle entering a passenger compartment. A vehicle loading abnormality detection system is disclosed. The image processing means uses a preset threshold value to determine whether or not the vehicle protrudes from the vehicle entry into the passenger compartment to the vehicle stop position, and the vehicle at the start of the transfer operation from the vehicle stop position to the parking space. The presence or absence of protrusion is detected.

JP 2010-285743 A

  In patent document 1, the outline of a vehicle is recognized in the image imaged in the passenger compartment of a mechanical parking lot, and the presence or absence of the protrusion of the vehicle with respect to a pallet is detected. As a method for extracting a region occupied by a moving object such as a vehicle from a captured image, a background difference method and an inter-frame difference method are mainly known.

  The background difference method is a method of extracting a moving object from a difference between an image when a moving object exists and a background image in a state where no moving object exists. In the background subtraction method, when the vehicle, which is the object to be extracted, and the background are similar in color, a part of the vehicle (for example, a bonnet or the like) may be erroneously recognized as the background, and the vehicle may be detected in a hollow state. is there. In addition, there is a high possibility of being adversely affected by reflected light from the surface of the vehicle body or the shadow of the vehicle. For this reason, there is a high possibility that accurate protrusion detection cannot be performed on the contour of the vehicle extracted by the background subtraction method.

  The inter-frame difference method is a method of obtaining two difference images from continuous images obtained by capturing moving objects at different points in time and extracting a common part of the two difference images as a moving object region. However, since the inter-frame difference method is a method for detecting a moving object in the first place, it is impossible in principle to extract a vehicle stopped on a pallet from a captured image.

  Thus, neither the conventional background subtraction method nor the interframe subtraction method is suitable for extracting from the captured image the area occupied by the vehicle stopped on the pallet.

  The present invention has been made in view of such a situation, and an object of the present invention is to accurately detect an outline of a vehicle from an image captured in a passenger compartment of a mechanical parking lot, thereby detecting a high-precision protrusion to a pallet. It is to provide the technology that makes it possible.

  One embodiment of the present invention is a vehicle protrusion detection device that detects protrusion of a vehicle on a pallet on which the vehicle is placed in a mechanical parking lot. This apparatus is a difference image that calculates an inter-frame difference image from two frames captured at a predetermined interval by an imaging unit that is installed so as to capture a predetermined range including a vehicle and a pallet. The initial part which calculates the logical sum of the inter-frame difference image calculated by the acquisition part and the difference image acquisition part between the start of warehousing of the vehicle and the stop on the pallet, and uses the outer peripheral shape of the logical sum image as the initial contour The contour setting unit and a dynamic contour model that extracts the contour of the object included in the initial contour by converging the initial contour inward according to the energy minimization principle, and executing the contour of the vehicle in the frame The contour extracting unit to be extracted, the contour of the vehicle extracted by the contour extracting unit, and a reference frame preset as representing the vehicle stop range on the pallet are compared, and the vehicle It includes a comparison determination unit for determining the presence or absence of protrusion of the vehicle depending on whether the contour is in the reference frame, a.

  According to this aspect, the contour of the vehicle is extracted from the captured image by applying the dynamic contour model, and the protrusion of the vehicle on the pallet is detected by comparison with a predetermined reference frame. As the initial contour, the outer peripheral shape of the logical sum of the inter-frame difference images of the vehicle moving in the passenger compartment is used. Thereby, even if the vehicle and the background are similar in color, it is possible to set an initial contour that reliably includes the vehicle to be extracted, so that the contour of the vehicle can be accurately extracted. In addition, since the initial contour can be set relatively small, the time required for the convergence process is shortened.

  Note that any combination of the above-described constituent elements, and those obtained by replacing the constituent elements and expressions of the present invention with each other among methods, apparatuses, systems, etc. are also effective as an aspect of the present invention.

  According to the present invention, since the contour of the vehicle can be accurately extracted from the image captured in the passenger compartment of the mechanical parking lot, it is possible to detect the protrusion of the pallet with high accuracy.

It is a figure which shows the boarding / alighting room which is a part of mechanical parking lot with a vehicle. (A), (b) is a figure which respectively shows the example of the picked-up image by the camera when the vehicle is not loaded on the pallet, and when it is loaded. It is a figure which shows the structure of the vehicle protrusion detection apparatus which concerns on one Embodiment of this invention. It is a figure explaining the setting of the initial contour by the initial contour setting part. It is a figure which shows an example of an initial stage outline. It is a figure which shows an example of the vehicle outline extracted by the outline extraction part. It is a flowchart of the vehicle protrusion detection method which concerns on one Embodiment of this invention.

  One embodiment of the present invention is a protrusion detection device that, when a vehicle is parked on a pallet of a mechanical parking lot, extracts a vehicle from an image captured by a camera and detects that the vehicle does not protrude from the pallet. First, a general configuration of the mechanical parking lot will be described, and then the protrusion detection according to the present embodiment will be described.

  FIG. 1 shows an entrance / exit room 100 of an underground type pallet type mechanical parking lot that constructs a parking space downstairs. 1A is a plan view, FIG. 1B is a cross-sectional view of the entrance / exit side from the back of the passenger compartment, and FIG. 1C is a side cross-sectional view.

  The entrance / exit 100 is surrounded by a wall 110 except for the entrance / exit 100a of the vehicle, and an elevating gate 120 is provided at the entrance / exit 100a. A long rectangular opening 200a that allows the long rectangular pallet 10 to be raised and lowered from the floor 200 in the passenger compartment 100 toward the underground parking space is formed. The pallet 10 is made of a metal material such as a steel plate having an area slightly smaller than the area of the opening 200a.

  The pallet-type mechanical parking lot uses a lifting device (not shown) arranged below the pallet 10 to raise or lower the pallet 10 on which the vehicle 20 is loaded, or a transfer device (not shown). ), The pallet is moved in the plane direction in the parking room, and the vehicle 20 is parked in the parking room.

  FIG. 1 shows a state in which a vehicle 20 is loaded on the pallet 10. When the vehicle 20 is loaded and unloaded, the pallet 10 is set at a height position substantially flush with the floor 200 of the passenger compartment 100. In other words, there is a gap of about several centimeters between the edge of the pallet 10 and the edge of the opening 200a, but the vehicle 20 is prevented from entering and leaving.

  A mirror 30 is erected adjacent to the inner wall 110 in the passenger compartment 100. This mirror 30 is for showing the approach state of the own vehicle to the driver of the vehicle that has entered the passenger compartment 100. Also, on the upper part of the mirror 30, there are guide means such as an electric bulletin board that visually guides the driver to enter and stop the vehicle, turning the steering wheel, and the like, and guide means 32 such as a speaker that is audibly guided. Provided.

  When entering the mechanical parking lot, the driver drives the vehicle, enters the entrance / exit 100 from the entrance 100a, and aligns the left and right while looking at the mirror 30 and receiving guidance from the guide means 32. Then, the vehicle is stopped within a vehicle stop position 12 set in advance on the pallet 10. Here, when a sign such as “OK” (or “get off”) comes out by the guide means, the driver turns off the engine and goes out of the vehicle. If the vehicle protrudes from the vehicle stop position 12 on the pallet 10, a part of the vehicle may come into contact with the edge of the opening 200 a or other device in the parking room during subsequent vehicle transportation, thereby damaging the vehicle or device. There is sex.

  Therefore, the boarding / exiting room 100 is provided with a vehicle protrusion detection device 60 for detecting the presence or absence of the protrusion of the vehicle based on the image captured by the camera 34. The configuration of the vehicle protrusion detection device 60 will be described later with reference to FIG.

  The camera 34 is installed on an extension line of one long side of the pallet 10. 2A and 2B show examples of images captured by the camera 34 when the vehicle 20 is not loaded on the pallet 10 and when the vehicle 20 is loaded. As shown in FIG. 2B, when the vehicle 20 is loaded on the pallet 10, the camera 34 captures a predetermined area including at least three sides of the entire vehicle 20 and the pallet 10. The installation position, camera angle, and viewing angle are selected so that they can. The camera 34 is preferably a digital camera that captures moving images at a predetermined frame rate (for example, 60 fps).

  If the camera 34 is in a plane that is parallel to the long side of the pallet and is perpendicular to the floor 200 and can capture a range including at least three sides of the entire vehicle 20 and the pallet 10, It can be installed in any place in the passenger compartment 100.

  FIG. 3 is a block diagram illustrating a configuration of the vehicle protrusion detection device 60 according to the present embodiment. Each block shown here can be realized by hardware such as a computer (CPU) (central processing unit) and other elements and mechanical devices, and software can be realized by a computer program or the like. Here, The functional block realized by those cooperation is drawn. Therefore, it is understood by those skilled in the art who have touched this specification that these functional blocks can be realized in various forms by a combination of hardware and software.

  The vehicle protrusion detection device 60 includes a difference image acquisition unit 62, a warehousing start determination unit 64, a stop determination unit 66, an initial contour setting unit 68, a contour correction unit 70, a contour extraction unit 72, a comparison determination unit 74, and a notification unit 76. .

  The difference image acquisition unit 62 extracts two frames from an image captured by the camera 34 at a predetermined frame rate, and calculates an inter-frame difference image. The two frames to be taken out may be the same as the frame rate of the camera, or may be longer than the frame rate (for example, skipping one frame).

  The warehousing start determination unit 64 determines that the vehicle 20 has started entering the passenger compartment 100 based on the inter-frame difference image calculated by the difference image acquisition unit 62. This difference is detected because the inter-frame difference image calculated at the time of non-warehousing has almost all pixel values zero, whereas the inter-frame difference image calculated at the time of warehousing increases the pixel value. By doing so, it is possible to determine the start of warehousing of the vehicle. Alternatively, the warehousing start determination unit 64 detects the approach of the vehicle 20 by a laser (not shown) installed at the entrance / exit 100a when a parking clerk presses a warehousing process switch (not shown). Or when the weight of the vehicle is detected by a load sensor (not shown) installed on the pallet 10, it may be determined that the warehousing has started.

  The stop determination unit 66 determines that the vehicle 20 has stopped on the pallet 10 based on the inter-frame difference image calculated by the difference image acquisition unit 62. For example, it can be determined that the vehicle is stopped when the difference between the inter-frame difference images substantially disappears. Alternatively, the stop determination unit 66 may be configured such that when a parking clerk operates a warehousing completion switch (not shown) or a laser (not shown) installed in the passenger compartment 100, the rear end of the vehicle 20 is used. It may be determined that the vehicle is stopped when it is detected that is placed on the pallet 10.

  The initial contour setting unit 68 uses the inter-frame difference image calculated by the differential image acquisition unit 62 from the entry of the vehicle into the passenger compartment 100 to the stop at the vehicle stop position, to the contour extraction unit 72. Set the initial contour to be given. The operation of the initial contour setting unit 68 will be described later in detail with reference to FIG.

  The contour correcting unit 70 corrects the contour set by the initial contour setting unit 68 into a shape suitable for processing in Snakes described later. For example, a contraction / expansion process is performed in which each pixel constituting the initial contour is moved inward by one pixel or moved outward by one pixel. By repeating this contraction and expansion a plurality of times, the jagged shape of the initial contour and the isolated portion can be erased to obtain a smooth contour line. Such contraction / expansion processing is also a well-known technique. For example, “10-2-3 contraction / expansion processing”, Digital Image Processing Editorial Committee, Digital Image Processing, Foundation for Promotion of Image Information Education, March 2006 1 day, p. 179-180. After the smooth initial contour is obtained by the contraction / expansion process, the contour correcting unit 70 may execute a thinning process for reducing the number of control points, which will be described later, in order to shorten the calculation time. Note that the above correction processing by the contour correction unit 70 does not necessarily have to be executed.

  The contour extracting unit 72 receives the initial contour from the contour correcting unit 70, applies Snakes known as a dynamic contour model, and extracts the contour of the vehicle 20 from the image captured by the camera 34. Snakes is a particularly excellent method for obtaining a closed curve such as the contour of a vehicle from a captured image. In the following, the case of using Snakes for vehicle contour extraction will be described. However, as long as the calculation method for receiving the initial contour and converging it is adopted, other known contour extraction methods or those developed in the future will be described. A contour extraction method can be applied to this embodiment.

Snakes extracts an edge as a closed curve when a region to be extracted (vehicle in this embodiment) is surrounded by an edge. In Snakes, a contour model is defined by a polygon having _n control points (vertices) CP _{1 to} CP _n . That is, when the position coordinates of each control point CP _i (i = 1 to n) are represented by (X _i , Y _i ), their set S = {(X _i , Y _i ), i = 1, 2,. . . , N} is one contour model. If the intervals between the control points of the contour model are close to a certain degree, a closed curve as the contour of the object can be expressed. A closed curve surrounding the periphery of the vehicle 20 to be subjected to contour extraction is set as an initial contour, and then the position coordinates of each control point CP _i are updated (X _i , Y _i ) so as to minimize a predetermined energy function. By doing so, the closed curve is gradually shortened, and finally the closed curve is pasted along the contour of the object. The contour extraction unit 72 outputs the stuck closed curve as the contour of the vehicle.

  As described above, since the initial contour is gradually contracted inward, there is an advantage that it is difficult to be influenced by the color inside the vehicle as long as an appropriate initial contour can be set outside the vehicle 20.

  Since the contour extraction method by Snakes is well known to those skilled in the art of image processing, further detailed description is omitted in this specification. As for Snakes, for example, Japanese Patent Laid-Open No. 9-6969 and “11-2-3 Region Division Processing Using Edge between Object and Background”, Digital Image Processing Editing Committee, Digital Image Processing, Foundation Image Information Education Promotion Association, March 1, 2006, p. 197-199.

  The comparison determination unit 74 compares the vehicle contour extracted by the contour extraction unit 72 with a predetermined reference frame corresponding to the vehicle stop position 12 on the pallet 10 within the imaging range. Since the camera 34 is always fixed to the pallet 10, a reference frame that forms a quadrilateral within the imaging range of the camera can be determined in advance. The comparison determination unit 74 determines whether all parts of the vehicle outline are inside the reference frame. If it is inside the reference frame, the vehicle 20 is appropriately loaded on the pallet 10.

  The notification unit 76 issues a warning to the driver of the vehicle 20 and / or the parking staff using the guide unit 32 based on the determination result by the comparison determination unit 74. When the comparison determination unit 74 determines that all of the vehicle outline is inside the reference frame, the notification unit 76 has completed parking at the predetermined stop position for the driver and / or the staff. Voice guidance to inform you, or display a message on an electric bulletin board. When the comparison determination unit 74 determines that a part of the vehicle contour protrudes from the reference frame, the notification unit 76 performs the warehousing operation to the appropriate stop position for the driver and / or the staff. Give voice prompts and display messages on electronic bulletin boards.

  Note that the comparison / determination unit 74 may determine which of the right side and the left side of the vehicle protrudes from the reference frame when a part of the vehicle outline protrudes from the reference frame. In this case, the notification unit 76 may give voice guidance or display a message to the vehicle driver informing that the vehicle should be parked on the left side or the right side of the pallet. A notification message may be issued.

  As described above, in order to execute Snakes, it is necessary to provide an appropriate initial contour that includes the vehicle. Since there are various shapes of vehicles that stop on the pallet, the initial contour cannot be set uniformly. On the other hand, if the initial contour is too large, the processing time increases and the practicality is impaired.

  It is also conceivable that a vehicle that is a moving object is detected by a known background difference method or inter-frame difference method, and its outer edge is used as an initial contour. However, as already described, the background difference method and the inter-frame difference method are not suitable for extraction of a vehicle stopped on the pallet in the passenger compartment.

  Therefore, in this embodiment, an initial contour is set using a logical sum of a plurality of inter-frame difference images. This method will be described below.

FIG. 4 is a diagram conceptually illustrating the initial contour setting method by the initial contour setting unit 68. The left column in the figure is a captured image captured at a predetermined frame rate by the camera 34, and images F _{t to} F _{t + 3} captured at times _t to _{t + 3} are shown in the figure, respectively. FIG. 4 shows a state where the vehicle is imaged from the side in order to make the drawing easier to see.

The difference image acquisition unit 62 calculates the inter-frame difference images D _{t to} D _{t + 2} shown in the middle row in the figure by taking the difference between two captured images that are temporally continuous.

In the conventional inter-frame difference method, a moving object is detected from a captured image by obtaining a logical product (AND) of two temporally inter-frame difference images. In contrast, in the present embodiment, a logical sum (OR) of two temporally continuous inter-frame difference images is obtained. In the right column of FIG. 4, OR images OR _t and OR _{t + 1} are shown.

Further, the initial contour setting unit 68 obtains a logical sum of all the logical sum images ORt obtained from images captured from the start of warehousing to stopping on the pallet, and obtains the outer peripheral shape of the obtained image. the initial contour _{C 0.} As a result, the initial contour C ₀ is equivalent to the logical sum of all the inter-frame difference images. It can be said that the initial contour C ₀ is the sum of changes due to minute vehicle movements for each frame.

If the inter-frame difference image acquisition interval is sufficiently short relative to the moving speed of the vehicle in the passenger compartment, as shown in the lower right of FIG. 4, the initial contour C ₀ is the outer periphery of the moving locus of the vehicle in the passenger compartment. Represents the shape.

  When a vehicle that is a moving object is to be detected by taking a logical product (AND) of inter-frame difference images as in the conventional inter-frame difference method, only a vehicle at a specific time point can be detected. Accordingly, when using the initial contour of a vehicle that stops on the pallet, it is necessary to set the acquisition timing of the inter-frame difference image so as to detect the vehicle at a specific time immediately before the stop. On the other hand, as in this embodiment, the movement trajectory of the vehicle in the passenger compartment is obtained using the logical sum (OR) of the inter-frame difference images, and the outer peripheral shape is set as the initial contour, so that Setting of the difference image acquisition timing and the like becomes unnecessary.

  In the example of FIG. 4, the inter-frame difference image is obtained from two captured images that are temporally continuous, but the captured image for obtaining the inter-frame difference image is thinned out, such as skipping one or two images. You may make it do. This thinning-out interval can be arbitrarily set according to the traveling speed of the vehicle moving in the passenger compartment as long as there are overlapping portions in the vehicle area in two temporally continuous inter-frame difference images. .

  In addition, by setting the outer peripheral shape of the movement locus of the vehicle as the initial contour, the initial contour is always a closed curve. Therefore, it is suitable as an initial contour to be given to Snakes.

FIG. 5 shows the vehicle trajectory extracted from the captured image by the initial contour setting unit 68 according to the method described above. Outer peripheral shape of the extraction area in the figure is the initial contour C _0. Preferably, the jaggedness of the initial contour C ₀ is removed by the above-described contraction / expansion process, and then given to the contour extraction unit 72.

  FIG. 6 shows the contour of the vehicle finally extracted by the contour extraction unit 72. The comparison determination unit 74 detects the presence or absence of the vehicle by comparing the outline of the vehicle with a reference frame representing the vehicle stop position on the pallet.

  FIG. 7 is a flowchart of the protrusion detection method according to this embodiment. First, the difference image acquisition unit 62 extracts two frames from an image captured by the camera 34 and calculates an inter-frame difference image (S10). The warehousing start determination unit 64 determines whether or not the vehicle 20 has started entering the passenger compartment 100 based on the inter-frame difference image calculated by the difference image acquisition unit 62 (S12).

When the warehousing start determination unit 64 determines that the vehicle has entered, the initial contour setting unit 68 sets the initial contour C ₀ according to the above-described method based on the logical sum of the inter-frame difference images (S14). The contour correcting unit 70 performs correction for smoothing the initial contour C ₀ as necessary.

The stop determination unit 66 determines whether or not the vehicle 20 has stopped on the pallet 10 based on the inter-frame difference image calculated by the difference image acquisition unit 62 (S18). When it is determined that the vehicle has stopped, the contour extraction unit 72 uses the initial contour C ₀ and the contour extraction unit 72 performs Snakes on the image captured by the camera 34 at a time point after the determination that the vehicle has stopped (S20). Is applied to extract the contour of the vehicle (S22). The comparison determination unit 74 compares the contour extracted by the contour extraction unit 72 with a predetermined reference frame, and determines whether all parts of the vehicle contour are inside the reference frame (S24). .

  When all the outlines of the vehicle are inside the reference frame (Y in S26), the notification unit 76 notifies the driver and / or the staff that parking at the predetermined stop position is completed (S28). When a part of the outline of the vehicle protrudes from the reference frame (N in S26), the notification unit 76 notifies the driver and / or the staff to redo the warehousing operation to the appropriate stop position (S30). In the latter case, when the warehousing operation is performed again in response to this notification, the processing after the vehicle stop determination is repeated. In the case of warehousing after the second time, it is not necessary to set the initial contour again, and the initial contour set at the time of the first warehousing can be reused. This is because the change in the position of the vehicle on the pallet between the first and the second and subsequent times is slight and the change is not so much as to change the initial contour of Snakes, unless the first warehousing is extremely deviated from the pallet. Because it is.

  In the above flow, the step of correcting the initial contour and the step of notifying completion or re-execution of warehousing can be omitted.

  As described above, according to the present embodiment, the contour of the vehicle is extracted from the captured image by applying the active contour model Snakes, and the vehicle protrudes on the pallet by comparison with a predetermined reference frame. To detect. As the initial contour to be given to Snakes, the outer peripheral shape of the logical sum of the inter-frame difference image of the vehicle moving in the passenger compartment is used.

  Since the vehicles parked on the pallet have various shapes, if the initial contour is fixed, the vehicle must be set to be considerably larger than the actual vehicle, and the convergence may be deteriorated or may not be converged. If the initial contour is set using the background subtraction method, an appropriate vehicle region may not be extracted when the background and the vehicle are similar colors. On the other hand, by setting the initial contour using the logical sum of the inter-frame difference images as in this embodiment, the initial stage that reliably includes the vehicle to be extracted even if the vehicle and the background are similar colors Since the contour can be given to Snakes, the contour of the vehicle can be extracted more accurately than the background subtraction method. Further, since the initial contour can be made smaller than the fixed value, the time required for the convergence process is shortened.

  The mechanical parking lot according to the embodiment has been described above. It is to be understood by those skilled in the art that these embodiments are exemplifications, and that various modifications can be made to combinations of the respective components, and such modifications are within the scope of the present invention.

  In the embodiment, it has been described that the contour of the vehicle is extracted when it is determined that the vehicle has stopped on the pallet. Instead, the same contour extraction may be repeatedly executed from the start of warehousing until the vehicle stops on the pallet. In this case, the initial contour setting unit continues to calculate the logical sum of temporally continuous inter-frame difference images. Then, the contour extraction unit performs the above-described contour extraction on the logical sum image calculated at each time point, and the comparison determination unit determines whether or not the extracted vehicle contour is within a predetermined reference frame. judge. Then, when the contour of the vehicle enters the reference frame, it is determined that the vehicle has stopped. That is, the vehicle contour extraction and the reference frame are continuously compared during the vehicle entry operation. In such a modification, it is possible to determine the stop of the vehicle based on the vehicle contour extraction result, so that the stop determination unit is unnecessary.

  In the embodiment, a so-called underground parking lot in which the passenger compartment is located at the upper end of the hoistway has been described. However, the present invention can be equally applied to a three-dimensional parking lot in which the passenger compartment is located at the lower end or middle of the hoistway as long as the vehicle is accommodated by moving the pallet. The same applies to a flat parking lot that does not have a lifting device. The present invention can also be applied to a conveyor-type mechanical parking lot in which vehicles are loaded on a conveyor and conveyed.

  DESCRIPTION OF SYMBOLS 10 Pallet, 12 Vehicle stop position, 20 Vehicle, 34 Camera, 60 Vehicle protrusion detection apparatus, 62 Difference image acquisition part, 64 Storage start determination part, 66 Stop determination part, 68 Initial contour setting part, 70 Contour correction part, 72 Contour Extraction unit, 74 comparison determination unit, 76 notification unit, 100 boarding / exiting room.

Claims (8)

A vehicle protrusion detection device for detecting the protrusion of a vehicle on a pallet on which the vehicle is placed in a mechanical parking lot,
An imaging means installed to image a predetermined range including the vehicle and the pallet;
A difference image acquisition unit for calculating an inter-frame difference image from two frames imaged at a predetermined interval by the imaging means;
Initial contour setting that calculates the logical sum of the inter-frame difference image calculated by the difference image acquisition unit from the start of warehousing to the stop on the pallet, and uses the outer peripheral shape of the logical sum image as the initial contour And
A contour for extracting a contour of a vehicle in a frame by executing a dynamic contour model for extracting a contour of an object included in the initial contour by converging the initial contour in accordance with an energy minimization principle. An extractor;
The vehicle contour extracted by the contour extraction unit is compared with a reference frame preset as representing a vehicle stop range on the pallet, and whether or not the vehicle contour is within the reference frame A comparison / determination unit that determines whether or not the vehicle protrudes;
A vehicle protrusion detection device comprising:   The predetermined interval is set such that there are overlapping portions in the vehicle area in two temporally continuous difference images according to the traveling speed of the vehicle moving in the passenger compartment. The vehicle protrusion detection device according to claim 1.   The vehicle protrusion detection device according to claim 1, wherein the predetermined interval is equal to a frame rate of the imaging unit.   4. The warehousing start determining unit for determining that the warehousing of the vehicle is started when an area change occurs between two temporally consecutive difference images. A vehicle protrusion detection device according to any one of the above.   The vehicle further includes a stop determination unit that determines that the vehicle has stopped on the pallet when no region change occurs between the temporally continuous two-frame difference images after the start of warehousing of the vehicle. The vehicle protrusion detection device according to any one of claims 1 to 4.   When the warehousing of the vehicle is re-executed according to the determination by the comparison and determination unit that the vehicle is protruding, the contour extraction unit uses the initial contour obtained at the previous warehousing as it is. The vehicle protrusion detection device according to any one of claims 1 to 5. The contour extracting unit repeatedly executes the extraction of the contour of the vehicle every time the logical sum of the inter-frame difference image is updated from the start of warehousing to the stop on the pallet,
The stop determination unit determines that the vehicle has stopped on the pallet when no change occurs between two contours continuously extracted by the contour extraction unit. The vehicle protrusion detection apparatus as described in. A vehicle protrusion detection method for detecting protrusion of a vehicle on a pallet on which the vehicle is placed in a mechanical parking lot,
Calculating an inter-frame difference image from two frames imaged at a predetermined interval by an imaging means installed so as to image a predetermined range including a vehicle and a pallet;
Calculate the logical sum of the inter-frame difference image calculated from the start of storage of the vehicle to the stop on the pallet, the outer periphery shape of the logical sum image as the initial contour,
Executing the dynamic contour model that extracts the contour of the object included in the initial contour by converging the initial contour in accordance with the energy minimization principle, and extracting the contour of the vehicle in the frame;
The extracted contour of the vehicle is compared with a reference frame preset as representing the vehicle stop range on the pallet, and whether or not the vehicle protrudes is determined depending on whether or not the contour of the vehicle is within the reference frame. A vehicle protrusion detection method including determining.

Images