NL2014154B1

NL2014154B1 - System and method for detecting the occupancy of a spatial volume.

Info

Publication number: NL2014154B1
Application number: NL2014154A
Authority: NL
Inventors: Van Dijk Roel
Original assignee: Lumi Guide Fietsdetectie Holding B V
Priority date: 2015-01-19
Filing date: 2015-01-19
Publication date: 2017-01-05
Also published as: NL2014154A

Abstract

There is provided a system for detecting the occupancy of a spatial volume, for instance a parking position for a vehicle. The system comprises at least two spaced apart optical sensors that are positioned alongside each other and both target the spatial volume; and computing means configured to accept synchronous images originating from the optical sensors and, from the images received, generate occupancy data for the spatial volume. A method for detecting the occupancy of a spatial volume is also disclosed.

Description

SYSTEM AND METHOD FOR DETECTING THE OCCUPANCY OF A SPATIAL

VOLUME

TECHNICAL FIELD

The invention relates to a system for detecting the occupancy of a spatial volume. The invention in particular relates to a system for detecting the occupancy of a parking position for a vehicle, such as a car or bicycle. The invention also relates to a method for detecting the occupancy of a spatial volume, making use of the invented system.

BACKGROUND ART

Systems for detecting the occupancy of a spatial volume are presently used for assessing the availability of parking positions in parking lots or garages for cars or bicycles for instance. Obtaining such information about availability may be useful to users, but also to managers of parking lots, irrespective of whether they are privately owned or public. Privately owned parking lots for bicycles may comprise parking lots of railroad companies, hospitals, and commercial companies, whereas public parking lots may comprise municipal parking’s for instance. Systems that detect occupancy may be used for free and for paying parking lots.

The known systems may be able to answer questions like the total amount of available parking positions at some instant, or the actual location of the available parking positions in the parking lot. Other questions that may also need to be answered however include the total amount of available parking positions in several parking lots in a municipality or town, how to direct a user to such an available parking position, the location of a user specific parking position, and the duration of occupancy of a particular parking position. A known system for detecting availability comprises counting the number of vehicles that enter a parking lot and leave the same. Such systems use induction for instance to detect passage of a vehicle. A counting system is sensitive to interference and is not based on a direct count of available parking positions. Frequent correction is necessary to increase confidence levels, for instance by a manual count. Vehicles that have been left behind (‘orphan vehicles’) are not detected by such a system either.

Other systems use ultrasonic or infrared sensors for each parking position. To obtain a desired confidence level, virtually any parking position needs to be equipped with such sensors. A vehicle moreover should not be placed too far from the sensor.

SUMMARY OF THE INVENTION

It is an aim of the present invention to provide a system for detecting the occupancy of a spatial volume, the system having a higher confidence level than achieved with the state of the art systems.

This and other aims are provided by a system comprising at least two spaced apart optical sensors that are positioned alongside each other and both target the spatial volume; and computing means configured to accept synchronous images originating from the optical sensors and, from the images received, generate occupancy data for the spatial volume. By using at least two spaced apart optical sensors that are positioned alongside each other and both target the spatial volume it becomes possible to perceive depth of an object within the spatial volume. This allows using a limited number of sensors, and yet obtaining a high level of confidence in determining occupancy. The at least two optical sensors create a stereo image of a spatial volume of interest.

The optical sensors may be any sensor known to one skilled in the art, such as 360° optical cameras, pan-tilt-zoom cameras or the more common cameras, and/or combinations thereof. The computing means are configured to generate occupancy data for the spatial volume by analysing synchronous images received from the optical sensors. The computing means may comprise personal computers, or, in another embodiment, may comprise computerboards especially adapted for image processing. Analysing is typically performed by image processing software. The software carried out by the computing means in an embodiment also controls the optical sensors.

In an embodiment of the invention, a system is provided wherein the computing means are configured to calculate distance data of an object within the spatial volume to the at least two optical sensors, and, from the calculated distance data, generate the occupancy data for the spatial volume. A further embodiment provides a system wherein the computing means are configured to store distance data corresponding to an unoccupied state of a spatial volume, and to generate the occupancy data of the spatial volume by comparing the calculated distance data of the object with the unoccupied distance data of the spatial volume.

Image processing software configured to calculate distance data from two images is known per se. Particularly suitable software is based on solving a stereo correspondence problem. In the context of the present application, the correspondence problem refers to the problem of ascertaining which parts of the image generated by a first optical sensor correspond to which parts of the image generated by a second optical sensor, and whether differences are due to the lapse of a time period and/or movement of an object within a target area of the images. In an embodiment of the invention, the computing means are configured to solve the correspondence problem by matching two images simultaneously received from two spaced-apart cameras. In another embodiment, the computing means are configured to solve the correspondence problem by matching N images simultaneously received from N spaced-apart cameras, where N>2. It is also possible to obtain occupancy information by receiving images from one or more cameras that are moving relative to the targeted spatial volume.

In one embodiment of the invention, the correspondence problem is solved by checking if one location in a first image looks similar or the same as another location in a second image. In another embodiment, a subset of features is located in a first image and it is checked whether the layout of the subset of features is similar in the first and a second image. A small window is passed over a number of positions in the first image, and each position is analysed with respect to how well it compares with the same location in a second image. When a fit is lacking or not good enough, the feature may not be present in both images.

Another embodiment of the invention provides a system wherein the computing means are configured to store a spatial relationship between the at least two optical sensors and any spatial volume, intended to be occupied by an object, and store distance data corresponding to an unoccupied state of each spatial volume.

Yet another embodiment relates to a system wherein the occupancy data comprise identification data related to the object. This embodiment allows a user to be informed about the exact location of an object that belongs to him, for instance his vehicle or bicycle, in an environment such as a parking lot.

Another embodiment relates to a system that further comprises a central server coupled to the computing means and configured to accept the occupancy data originating from the computing means. The coupling between the central server and the computing means may be wired or wireless. This embodiment may for instance provide a user of the system with information about the total amount of available parking positions in several parking lots in a municipality or town, and/or may direct a user to such an available parking position.

Another embodiment of the system according to the invention further comprises monitoring means coupled to the computing means and/or the central server, and configured to make the occupancy data accepted by the central server visually available. Preferably, such monitoring means comprise electronic billboards, smart phones, and the like.

The invention also relates to a method for detecting the occupancy of a spatial volume, for instance a parking position for a vehicle. The method in accordance with the invention makes use of the invented system and comprises positioning at least two optical sensors alongside each other and mutually spaced apart, such that they both target the spatial volume; provide a computing means, and send synchronous images originating from the optical sensors to the computing means, and, from the images received, generate occupancy data for the spatial volume.

Useful embodiments of the invention comprise methods wherein the computing means calculate distance data of an object within the spatial volume to the at least two optical sensors, and, from the calculated distance data, generate the occupancy data for the spatial volume; wherein the computing means store distance data corresponding to an unoccupied state of a spatial volume, and generate the occupancy data of the spatial volume by comparing the calculated distance data of the object with the unoccupied distance data of the spatial volume; wherein the computing means store a spatial relationship between the at least two optical sensors and any spatial volume, intended to be occupied by an object, and store distance data corresponding to an unoccupied state of each spatial volume; and/or wherein the occupancy data comprise identification data related to the object.

Other useful embodiments relate to methods that further provide a central server that is coupled to the computing means and that send the occupancy data originating from the computing means to the central server; that further provide monitoring means that are coupled to the computing means and/or the central server, and make the occupancy data accepted by the central server visually available; and methods wherein the monitoring means comprise electronic billboards, smart phones, and the like.

The system and method in accordance with the invention may also be used for other purposes, including but not limited to providing security to a parking lot for instance, or to occupancy of mooring jetties in harbors, and the like.

Embodiments of the system and method of the invention are particularly useful in bicycle sheds or parking’s, since bicycles are not easily located, for instance because they are frequently placed against each other and outside reserved parking positions. They may also be parked for longer periods of time, which may be weeks or even months. A bicycle user preferably needs to know beforehand if a parking position is available in a parking lot, or if parking positions are available in which parking lots, if several of such lots are optional. Larger cities tend to provide larger and larger bicycle parking’s, for instance close to railroad stations. In such cases, the system and method of the invention not only provide a bicycle user with instant information about the exact location of an available parking position, but in some embodiments may also provide him with directions.

Embodiments of the system and method of the invention are also particularly useful to a manager of a parking lot. The invention for instance may provide a manager with information about how long a bicycle (or other vehicle) occupies a particular parking position. It then becomes possible to identify bicycles that have been left behind, for instance by residents that use a free parking space.

It is expressly stated that the embodiments of the invention disclosed in the present application may be combined in every possible combination of these embodiments, and that each embodiment separately may be the subject of a divisional application.

DETAILED DESCRIPTION OF THE INVENTION

The present invention will now be described in more detail with reference to the figures without however being limited to the illustrated embodiments. In the figures:

Figure 1 schematically illustrates a system for detecting vacant parking positions in a bicycle parking lot in accordance with an embodiment of the invention; and

Figure 2 schematically illustrates a flow chart of the steps taken by the computing means in analysing images received from two optical sensors in accordance with an embodiment of the invention.

With reference to figure 1 a parking rack 1 having two floors (la, lb) is shown. The parking rack 1 is positioned within a larger parking lot. Parking rack 1 can accommodate 10 bicycles 3 on a ground floor lb and another 10 bicycles 3 on a first floor la at a distance of about 120 cm above the ground floor lb. Next to the parking rack 1, a number of ground compartments or sections (2a, 2b) are provided, for instance by painting demarcation lines on the floor of the parking lot. These sections may be used for parking off-size bicycles, such as those provided with children seats or transport crates. Attached to a ceiling 5 at a height of about 250 cm above ground floor lb, two Axis 360° cameras (model M3007PV) (4a, 4b) are provided. The cameras (4a, 4b) are spaced apart over a distance 6 of about 15 cm. The cameras (4a, 4b) are positioned such that they provide a field of vision of about 360° around a central axis (7a, 7b) of each camera (4a, 4b), whereby the central axes (7a, 7b) extend downwards in a vertical direction. The cameras (4a, 4b) are connected to a computer network that comprises a computing means 8 by means twisted pair cables (9a, 9b). Image processing software installed on the computing means 8 collects simultaneous images that originate from both cameras (4a, 4b), and that are transmitted via an application programming interface (API) of both cameras (4a, 4b). The images may be transmitted within a range of frequencies, and in a typical example are transmitted at a frequency of 10 Hz.

In an embodiment of the invention, the software installed on the computing means is configured, i.e. each parking position in a parking lot is defined in the images produced by the cameras (4a, 4b). Such configuration is typically carried out when the software is installed for the first time, or when the spatial arrangement of the parking positions, and/or the position of the cameras relative to the targeted parking positions has been changed.

After having configured the software, a calibration step is typically carried out for each parking position. Calibration involves determining the distance from an unoccupied parking position to the camera’s position. Since the cameras are spaced apart over some (small) distance (in the order of 10 to about 100 cm), the distance of the parking position to the cameras corresponds to the average distance to the cameras. The computing means 8 are configured to store the distance data corresponding to an unoccupied state of the spatial volumes, i.e. the parking positions. The software installed on the computing means 8 determines the distances by solving a stereo correspondence problem. Indeed, for each parking position, the software shifts the targeted part of the images of both cameras until these image parts overlap. The distance to the cameras is determined from the amount of shift, needed to bring corresponding points of the images into an overlapping arrangement.

After configuration and calibration of the software installed on the computing means 8, the computing means of the system according to the invention have stored a spatial relationship between the cameras (4a, 4b) and any parking position, intended to be occupied by a bicycle 3, and have stored distance data corresponding to an unoccupied state of each parking position. The system is now ready for determining the availability of parking positions within a parking lot. The occupancy data of any parking position are generated by comparing the calculated distance data of an object present in the parking position with the unoccupied distance data related to said parking position.

In an embodiment, the configured software detects any change that occurs in the produced images. If such change occurs, the software installed on the computing means 8 typically calculates whether the change represents a new ‘static’ situation, or whether the change is ‘dynamic’ in the sense that the change is temporary only, such as when a person is walking past the imaged parking positions and therefore crosses the produced images. A change will be considered ‘static’ when it lasts for a predetermined amount of time, for instance in the range of (tenths of) seconds.

The steps carried out by the software are illustrated by the flow chart of Figure 2. In step 10 of the method, a left and a right camera (4a, 4b) collect images in a predefined imaging area. A next step 11 determines whether any change in the images produced is ‘static’ or ‘dynamic’. If it is determined at step 11 that the change in the images is ‘dynamic’, i.e. a motion is detected, the produced images are neglected and the images last generated are withheld. If it is determined at step 11 that the change in the images of the imaging area is ‘static’, i.e. no motion is detected, the images obtained by the camera 7b on the right are transformed in a next step 12. This transformation involves matching the image of the right camera 7b with the local vanishing point of the image produced by the camera 7a on the left. A further step 13 involves looking for matching points (pixels) between the left and right images. This is for instance carried out by searching the left image for interesting points, defined as parts of the spatial volume (or parking position) that may also be found with a high level of confidence on other images of the same spatial volume. Corresponding points describing the same part of the spatial volume are then located in the right image.

Corresponding points in the left and right images form a pair. A next step 14 of the method involves filtering these pairs of points. Filtering may conveniently be carried out by defining an axis and filter the pairs with respect to the length of a shift of the pairs and the angle of said shift relative to the axis. Other ways of filtering may also be used however.

After having calculated such a pixel shift in step 15, a decision is taken by the software about the availability of a targeted parking position. Such a decision making step 16 may be performed by setting the limit for the amount of pairs of points having a correct length and angle after the filtering step. When a sufficiently high amount of pairs of points having the correct length and angle remain after the filtering step (i.e. the limit value is exceeded), a parking position is defined as ‘occupied’, a state denoted as 17 in the flowchart of Figure 2. When the amount of pairs of points having the correct length and angle after the filtering step does not exceed the set limit value, a parking position is defined as ‘free’, a state denoted as 18 in the flowchart of Figure 2.

After having analysed substantially all image areas (20a, 20b, 20c) according to the procedure described above, the software installed in the computing means 8 requests new images from the cameras.

The system and method of the invention allow detecting the occupancy of spatial volumes within a larger spatial volume with a high accuracy. The optical sensors used in the invented system and method generate images that are subsequently analysed to determine a distance of the cameras to a targeted spatial volume. The targeted spatial volume is provided with enough light to ensure visibility by the cameras of the spatial volume. This may be achieved by lightening in the visible frequency range, but also by infrared lightening for instance. The system is able to detect a change in distance of 1 cm and less at a camera distance of 7 m when an image size of 1600 x 1200 pixels and a lens aperture of 80° are used for instance.

The system and method of the invention provide additional advantages and possibilities over the prior art systems and methods. Objects that have been left behind (‘orphan objects’) for instance are easily and accurately detected. This is conveniently carried out by a method in which the software installed on the computing means 8 keeps the last image taken of a parking position in an occupied state, and compares this image with the first image of the same parking position after a change has occurred in the image area. Keeping track of the times at which both images were taken allows to calculate the lapsed time frame. If this time frame exceeds a set time limit, it can be decided that the occupied parking position is occupied by an orphan object. A manager of the parking lot can then take the appropriate action such as removing the object.

The system and method may also be used to direct a user to his own object (vehicle, bicycle). This embodiment requires that the object carries some identification means that is visible to the optical sensors or cameras. Suitable ID means comprise a license plate or a QR code for instance. The user in this embodiment provides the system with the correct ID of his object, for instance through a suitable interface, such as his mobile phone.

Since the optical sensors of the system of the invention provide visual information about parking positions and about an imaging area that preferably covers more volume than such parking positions, the system may simultaneously be used as a security means. The system may also be used to collect statistical data, such as those related to the use of the parking lot or garage, favourite walking routes, unused or barely used parking positions, and other useful information.

Claims

1. System for detecting the occupancy of a spatial volume, for instance a parking position for a vehicle, the system comprising at least two spaced apart optical sensors that are positioned alongside each other and both target the spatial volume; and computing means configured to accept synchronous images originating from the optical sensors and, from the images received, generate occupancy data for the spatial volume.

2. System according to claim 1, wherein the computing means are configured to calculate distance data of an object within the spatial volume to the at least two optical sensors, and, from the calculated distance data, generate the occupancy data for the spatial volume.

3. System according to claim 2, wherein the computing means are configured to store distance data corresponding to an unoccupied state of a spatial volume, and to generate the occupancy data of the spatial volume by comparing the calculated distance data of the object with the unoccupied distance data of the spatial volume.

4. System according to any one of the preceding claims, wherein the computing means are configured to store a spatial relationship between the at least two optical sensors and any spatial volume, intended to be occupied by an object, and store distance data corresponding to an unoccupied state of each spatial volume.

5. System according to any one of the preceding claims, wherein the occupancy data comprise identification data related to the object.

6. System according to any one of the preceding claims, further comprising a central server coupled to the computing means and configured to accept the occupancy data originating from the computing means.

7. System according to claim 6, further comprising monitoring means coupled to the computing means and/or the central server, and configured to make the occupancy data accepted by the central server visually available.

8. System according to claim 7, wherein the monitoring means comprise electronic billboards, smart phones, and the like.

9. Method for detecting the occupancy of a spatial volume, for instance a parking position for a vehicle, the method comprising positioning at least two optical sensors alongside each other and mutually spaced apart, such that they both target the spatial volume; provide a computing means, and send synchronous images originating from the optical sensors to the computing means, and, from the images received, generate occupancy data for the spatial volume.

10. Method according to claim 9, wherein the computing means calculate distance data of an object within the spatial volume to the at least two optical sensors, and, from the calculated distance data, generate the occupancy data for the spatial volume.

11. Method according to claim 10, wherein the computing means store distance data corresponding to an unoccupied state of a spatial volume, and generate the occupancy data of the spatial volume by comparing the calculated distance data of the object with the unoccupied distance data of the spatial volume.

12. Method according to any one of claims 9-11, wherein the computing means store a spatial relationship between the at least two optical sensors and any spatial volume, intended to be occupied by an object, and store distance data corresponding to an unoccupied state of each spatial volume.

13. Method according to any one of claims 9-12, wherein the occupancy data comprise identification data related to the object.

14. Method according to any one of claims 9-13, further providing a central server that is coupled to the computing means and sending the occupancy data originating from the computing means to the central server.

15. Method according to claim 14, further providing monitoring means that are coupled to the computing means and/or the central server, and making the occupancy data accepted by the central server visually available.

16. Method according to claim 15, wherein the monitoring means comprise electronic billboards, smart phones, and the like.