JP3806508B2 - Parking lot image monitoring device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for monitoring a vacant state of a parking space (vehicle compartment) in a parking lot, and more particularly to a parking lot image monitoring device that measures a vehicle occupancy state and monitors a running path using a television camera.
[0002]
[Prior art]
In the management of the parking lot, the occupancy management of the passenger compartment is performed and if there is an empty space, the parking lot user (user) is informed. As described in this publication, sensors such as ultrasonic sensors are installed in each vehicle compartment to determine whether a vehicle is present in each vehicle compartment. In addition, the parking lot manager monitors the video of the TV camera installed in the aisle to monitor the state of the car running on the road, and always displays a monitor screen for any abnormalities such as accidents on the road. Monitor by watching. On the other hand, in JP-A-4-290199 as a parking lot management by image processing, the entire parking lot is photographed by a camera, a vehicle is extracted by inter-frame difference processing of this camera image, and the coordinate data of the extracted image and the parking lot in advance are extracted. There is one that manages a parking position and an empty vehicle position in a parking lot by associating with a map creating data and obtaining a movement locus of the car. Further, according to the invention described in Japanese Patent Laid-Open No. 9-16897, a parking lot is photographed from directly above with a camera, and a vehicle is extracted by differential processing between this image and a pre-captured parking lot background image. The vehicle presence management is performed by managing the empty vehicle position.
[0003]
[Problems to be solved by the invention]
Sensors such as ultrasonic sensors themselves are inexpensive, but there is a problem that including the construction cost for installing a large number of sensors is problematic, and people stand above or below the sensors. When the vehicle is present, there is a problem that the result of the vehicle presence determination is output even when the vehicle is not present. Moreover, since the manager of a parking lot needs to always monitor the monitor screen and check whether there is an abnormality, it becomes a very difficult task. On the other hand, the problem of the prior art by image processing is that the image processing according to Japanese Patent Laid-Open No. 4-290199 does not take into account that the shooting angle of view of the camera becomes invisible due to overlapping of vehicles, so that accurate processing cannot be performed. In addition, there is a problem that all the vehicles cannot be detected only by the inter-frame difference process (stopped vehicles cannot be detected). Japanese Patent Laid-Open No. 9-16897 discloses an effective method in a place where the camera can take a picture of the parking space from directly above, but such an image is used when the ceiling is low such as an underground parking lot. There is a problem that it cannot be set to the corner.
[0004]
[Means for Solving the Problems]
In the present invention, in order to solve the above-described problem, a parking lot image monitoring apparatus according to the present invention captures a plurality of passenger compartments that are parking spaces in a parking lot and a traveling road at the same time, and images of the television camera. A parking lot image monitoring apparatus comprising an image processing apparatus for inputting and processing an image, comprising: an image input means for inputting video from a camera; an image memory for storing an image input by the image input means; and an image input means Area setting means for setting the processing area of the road area and the area of the passenger compartment with respect to the input camera image, and monitoring for monitoring the running of the vehicle or the presence of the vehicle in each area set by the area setting means Means. In this way, a plurality of passenger compartments in the parking lot are photographed from the front rear by a single TV camera, and this image is automatically determined for the presence of vehicles in each passenger compartment by image processing technology. is there. As camera installation conditions at this time, it is assumed that one of the vehicle compartments side by side facing the vehicle is photographed in a bird's-eye view from the front rear side with a passage in between. With such an angle of view, it is possible to determine whether or not the vehicle compartment is hidden by the presence or absence of a vehicle on the road, so that the presence determination of the vehicle compartment can be accurately performed. In addition, since the traveling road is photographed at the same time, it is possible to monitor abnormality of the traveling road. This abnormality monitoring of the travel path is a function that outputs an alarm when the vehicle has been in the travel path area for a long time. In other words, when processing a single image, it is possible to solve the problem that accurate vehicle compartment measurement was difficult with the conventional method by separating the route monitoring and the vehicle compartment monitoring. It is also possible to monitor the road abnormality.
[0005]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[0006]
FIG. 1 shows a shooting state of the television camera of the present invention.
[0007]
The TV camera was taken from this place, with a plurality of passenger compartments lined up facing each other installed in the rear of the front across the passage. Cabs 2-1 to 2-6 are shown in the video 1 of the TV camera, and the road 3 is shown in front of it. A parked vehicle 4 enters the passenger compartment, and a passing vehicle 5 is shown on the road. By setting such a field of view, it is possible to measure on-board management of a plurality of passenger compartments with a single camera and to simultaneously monitor an abnormality in the traveling path.
[0008]
FIG. 2 shows the system configuration of the present invention.
[0009]
Image input means 8 for inputting the video of the camera, image storage means 9 for storing the input image, area setting means 10 for setting a processing area on the image memory, and travel path monitoring means 11 for monitoring the travel path The vehicle compartment monitoring means 12 for executing the vehicle occupancy determination and the communication means 13 for transmitting the measurement result.
[0010]
First, the vehicle compartment monitoring means 12 will be described. When monitoring each compartment area, if a vehicle is present on the road, the vehicle hides the compartment, which may cause an error in the vehicle presence determination in the compartment. Therefore, before the vehicle presence determination in the passenger compartment, a function for determining whether a vehicle exists on the road is provided, and the vehicle presence determination in the passenger compartment is executed when there is no vehicle on the road. ing. That is, as shown in FIG. 1, a processing area 6 for monitoring the road is defined in advance, and a measurement area 7 is defined in a predetermined area of each passenger compartment. As shown in FIG. 3, the process inputs an image 14, sets a processing area in the travel road area of the image 15, determines whether a vehicle is present on the travel road 16, and if it does not exist, The vehicle compartment monitoring means 12 for measuring the state of each vehicle compartment is executed. Thus, by dividing the monitoring into the traveling road and the monitoring of the passenger compartment, the vehicle presence determination in the passenger compartment is not erroneously performed depending on the vehicle on the traveling road, so that it is possible to accurately determine the presence of the passenger compartment. Determination of whether a vehicle is present on the travel path will be described later. The vehicle compartment monitoring means 12 in FIG. 3 has a flow as shown in FIG. First, the cabin number loop process 18 is performed, a processing area is set 19 in a predetermined area of each compartment, and the area is differentiated 20. Since the vehicle has many horizontal stripes, the pattern appears when it is differentiated. If there is no vehicle, nothing will appear. The concept of this differentiation process is shown in FIG. The differential image is further binarized with a predetermined threshold value 21 and the area of the binary value is calculated 22. When the binary area is larger than the set value, it is determined that the vehicle is in an on-vehicle state 24; 25. With such processing, it is possible to determine the presence of each vehicle compartment. Here, the presence of the vehicle is determined using differential processing, but there is also a shading pattern matching method.
[0011]
The principle is shown in FIGS.
[0012]
As shown in FIG. 6, areas to be pattern-matched to predetermined areas in the passenger compartment 2 are set as 26-1 and 26-2, and an image when no vehicle exists in advance is stored for those areas. This is called a template image. When pattern matching is performed at the same place in the template image and the input image, the similarity is high because the image is almost the same as the stored image in the case of an empty car, and the similarity is significantly lowered when a car enters as shown in FIG. Due to such a difference, it is possible to determine whether the vehicle is in the passenger compartment. In this embodiment, two pattern matching areas are set. However, the same effect can be obtained by setting one large area and performing pattern matching processing on the area. In the two-location method, when the degrees of similarity are different in the two areas, it can be grasped that the car is moving, and therefore, the determination condition is that the degrees of similarity are lowered at the two locations. Thus, since the pattern matching method does not use binarization processing, there is an advantage that determination accuracy is high. However, since it is necessary to store the state of the empty vehicle in advance, there is a drawback that an erroneous determination is made when the vehicle happens to be present when the system is started up. Therefore, as shown in FIG. 8, a process combining two methods of a differentiation method and a pattern matching method can be considered. Since the vehicle compartment monitoring means 17 determines whether or not there has been a vacant state in each vehicle compartment in the past, and if there is no vacant state in the state of the relevant vehicle compartment in the past, a background image (template image) has not been created. Car presence determination is performed by a differentiation method 34 to 37, and when it is determined that the vehicle is empty, an empty state flag is turned on 41 and a pattern matching background image (template image) 42 is created. Also, if there is an empty state in the past, a background image (template image) has already been created, so pattern matching 29 is performed between this background image (template image) and the input image to determine whether the similarity is low. It is determined 30. If it is low, it is determined 31 that the vehicle is present, and if it is high, it is determined 32 that the vehicle is empty. If it is determined that the vehicle is in an empty state, the stored background image is updated 33. In this update process, the brightness of the image in the empty state changes due to the influence of illumination, dirt on the camera lens, and so on. For example, averaging with the input image is executed to update the stored background image. Here, it is optimal to use normalized correlation processing, which is a basic function of image processing, for the pattern matching processing. By the way, the vehicle presence determination can be performed by the pattern matching processing and the differentiation processing, but there is also a method of determining by executing both functions simultaneously. FIG. 9 shows the flow. The difference from FIG. 8 is that pattern matching is executed when there is an empty state in the past, and differentiation processing is always executed regardless of whether or not there is an empty state. As a result, two determination results can be obtained. If both types of determination results are the same, the result is output as it is, and if it is different, the pattern matching process is given priority (when the pattern matching process is executed) and output. It is a thing. By adopting such processing, it is possible to improve the accuracy of determination. FIG. 10 shows the hardware configuration of the present invention. In the image monitoring device 47, the video signal of the television camera 48 is taken into the image memory 50 via the A / D converter 49. An image processor 51 that reads the contents of the image memory 50 and executes various image processes, a D / A converter 52 for displaying the contents of the image memory 50 on a monitor, a CPU 53 that executes various controls, and a program storage memory A RAM / ROM 54, an image processing parameter file 55 for storing various processing areas, and a recognition result communication unit 56 are provided. Such a configuration is not particularly different from general image processing hardware. Next, the travel route monitoring means will be described. The travel path monitoring means 11 has two functions. That is, a function for determining whether or not a vehicle is present on the travel path for determining whether or not to operate the passenger compartment monitoring means 12 and an abnormality monitoring function for the travel path. First, the traveling road vehicle presence determination function will be described. The determination as to whether or not a vehicle is present on the travel path is performed by differentiating the processing region 6 in FIG. 1 in the same manner as in FIG. 5 and binarizing the differential image with a predetermined threshold value. Can be determined that the vehicle is on the road. In addition, a moving object can be detected by performing inter-frame difference processing (images are captured at predetermined time intervals and subjected to differential processing), so determination may be made by this method. However, since the vehicle cannot be detected when the vehicle is stopped, the vehicle compartment monitoring means may be erroneously activated. In addition, there is a method in which an image in which no vehicle is present on the road is input in advance, and the vehicle is extracted by a difference process between the background image and the input image. Normally, there is no vehicle traveling in the aisle at the start of business (in the case of a passenger compartment, there may be a vehicle that has cleared the night), so a background image can be stored before the parking lot starts business. As described above, when the background image can be stored, the pattern matching process is executed between the input image and the background image, similarly to the passenger compartment monitoring means, and when the similarity is low, it is determined that the vehicle exists. You can also By such processing, vehicle presence confirmation on the traveling road can be executed. On the other hand, the abnormality monitoring function of the traveling road monitoring means monitors a state where the vehicle has stopped for a long time on the traveling road, and is realized by the processing shown in FIG. Since it is a determination whether the vehicle is stopped, two images are input at predetermined time intervals, and the difference processing of the images is executed. Since the part where the difference in brightness is small is a non-moving area, this image is created, and the vehicle is extracted by the difference process between this image and the background image stored in advance and the input image. If an AND process is performed on an area with a small difference between the frames, an image of the stopped vehicle can be obtained. The center position coordinate (xc, yc) of this image is calculated, and when this coordinate has not moved from within the predetermined range, the time is measured, and when the predetermined time has elapsed, an alarm signal is notified. it can. Alternatively, the same output can be obtained by using only the background difference image and managing the center coordinates of the extracted image. By such processing, a travel road abnormality monitoring function can be realized. Moreover, it is also possible to grasp the state of the traveling path of the entire parking lot by extracting the vehicle on the traveling path by background difference processing and sending the central coordinates to the host computer. The conceptual diagram is shown in FIG. The host computer 57 receives the coordinates of the vehicle on the road from a plurality of image processing devices (parking lot image monitoring devices) 47, and this data is converted into a map of the entire parking lot in advance, and the correspondence between this data and the received data. If the data is plotted on the map 58 of the entire parking lot, the state of the vehicle on the entire travel path can be grasped. The image monitoring apparatus and the host computer are connected by a network 59. Further, as shown in FIG. 13, each image processing apparatus is provided with an image compression / transmission function, and in response to a request from the host computer, the video of the television camera being shot is compressed and transmitted. The host computer decompresses the received data, It can be displayed as a designated image on the map 58 of the entire parking lot. The compression method is selected according to the transmission time and required image quality. As a result, when an abnormality such as a traveling road occurs, the administrator can view the video instantly, so that the management operation becomes very smooth. The traveling road monitoring means and the passenger compartment monitoring means have been mainly described above, but the vehicle presence determination of the passenger compartment can be accurately executed by performing image processing separately for the traveling road monitoring means and the passenger compartment monitoring means.
[0013]
【The invention's effect】
By occupying a plurality of passenger compartments and traveling roads with a single TV camera and providing traveling road monitoring means and passenger compartment monitoring means, it is possible to realize occupancy management of a parking lot at a low cost. It is possible to detect an abnormality in the travel path. In addition, by sending these data to the host computer, it is possible to collect the vehicle management on the vehicle and the vehicle status on the road. Furthermore, by providing an image compression / transmission function in the image processing device, an image decompression function, and a display function in the host computer, if an abnormality occurs on the road, the administrator can immediately check the video in the vicinity. There is an effect of smooth operation.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating an example of a shooting angle of view according to the present invention and names of respective units.
FIG. 2 is a diagram showing a system configuration of an image monitoring apparatus of the present invention.
FIG. 3 is a diagram showing a vehicle presence determination flow according to the present invention.
FIG. 4 is a diagram showing a flow of an embodiment of the vehicle compartment monitoring means.
FIG. 5 is a diagram showing a concept of differential processing for vehicle detection.
FIG. 6 is a diagram illustrating an example of region setting when a vehicle is detected by a pattern matching method.
FIG. 7 is a diagram illustrating an example when a vehicle is present in a pattern matching area.
FIG. 8 is a diagram showing a flow of one embodiment of the vehicle compartment monitoring means when the differentiation method and the pattern matching method are combined.
FIG. 9 is a diagram showing a flow of one embodiment of the vehicle compartment monitoring means when the differentiation method and the pattern matching method are executed simultaneously.
FIG. 10 is a diagram illustrating a hardware configuration of an image monitoring apparatus according to the present invention.
FIG. 11 is a conceptual diagram of processing of an abnormality monitoring function of a traveling path monitoring unit.
FIG. 12 is an example of a system configuration diagram in a case where a host computer and an image monitoring apparatus are connected to grasp the entire parking lot.
FIG. 13 is an example of a system configuration diagram in a case where a host computer and an image monitoring apparatus are connected to grasp an entire parking lot to which an image transmission function is added.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 8 ... Image input means, 11 ... Traveling route monitoring means, 12 ... Vehicle compartment monitoring means, 13 ... Communication means, 47 ... Image monitoring apparatus.

Claims (4)

A television camera for photographing a passenger compartment in which a vehicle in a parking lot is parked, an image input means for inputting a video from the television camera, and an image memory for storing an image input by the image input means. A parking lot image monitoring device that monitors the vehicle by image processing,
The TV camera, for a plurality of passenger compartments arranged across the traveling road facing the running road of the parking lot, simultaneously facing the plurality of passenger compartments across the traveling road from the front rear of the passenger compartment, It is installed at a position to take a bird's-eye view so that the traveling path is in front of the facing vehicle compartments,
Area setting means for setting each processing area of the area of the plurality of facing rooms and the area of the traveling path for the image input from the image input means,
For the area of the road set by the area setting means, a road monitoring means for detecting a moving body by inter-image calculation with a background image stored in the image memory;
The vehicle compartment state set by the region setting means is based on the similarity of the vehicle compartment image stored in advance in the image memory and the similarity of the input image from the television camera by pattern matching processing. Vehicle compartment monitoring means for monitoring
A parking lot image monitoring device, wherein, when the travel route monitoring means determines that no vehicle is present on the travel route, the vehicle compartment monitoring means monitors the presence of a vehicle in each compartment. The parking lot image monitoring device according to claim 1,
The travel path monitoring means, characterized in that a function determining the travel or presence of a vehicle by determining whether the vehicle traveling road running differential processing is present for the area of the traveling path which has been set Parking image monitoring apparatus according to. The parking lot image monitoring device according to claim 1,
The casing monitoring means has a function of determining the set the running differential processing for an area of a plurality casing facing standing car, the cabin and input images of empty stored in advance for the area A parking lot image monitoring apparatus, wherein the presence of a vehicle in each vehicle compartment is determined by a function of performing a pattern matching process between the vehicle and determining the presence of the vehicle . In the parking lot image monitoring device according to claim 2 ,
The travel path monitoring means set the the region of the running path extracting vehicle, extracted Parking image monitoring apparatus and outputs the coordinates of the vehicle.

Images