JP2000163697A - Picture recognizing device and traffic control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effect picture recognition having high robustness against disturbances. SOLUTION: A picture inputting section IP1 inputs a picture and a change detecting section IP2 detects a changed area in the picture. An area calculating section IP3 calculates the area of the changed area. Since an error caused by a disturbance may be contained in the area, an area correcting section IP4 corrects the error by suppressing the abrupt fluctuation of the area value by detecting a medium value. An area averaging section IP5 successivelyt averages the corrected area values during a fixed period in the past, the errors are scattered and reduced thereby the accuracy of state discrimination by means of a discriminating section. IP6 which discriminates the state by utilizing the area can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic control system for supporting smooth operation of vehicles and the like, and more particularly to an image recognition device for extracting a moving object and a traffic control system using the same.

[0002]

2. Description of the Related Art In an image recognition apparatus for extracting a moving object according to the prior art, a moving object region in an image is extracted by an inter-frame difference process and a background difference process.
In the inter-frame difference processing, first, luminance difference processing between images is performed using images at different shooting times. Next, a threshold value process is performed on the luminance difference value to create a binary image, and an area where the luminance has changed, that is, a moving object area has been extracted. In the background subtraction process, an image in which no object to be extracted is present is set as a background image, and a difference process of the luminance between the background image and the input image is performed to binarize and extract a moving object region.

[0003]

However, in the conventional image recognition apparatus, the image pickup apparatus shakes due to vibrations or strong winds accompanying the vehicle passing, and the so-called disturbance of the passing vehicle or the like or the reflection of headlights. If a luminance change occurs during the process, an erroneous region is also extracted in addition to the moving object region to be originally extracted.Therefore, when the determination process is to be performed using the total area of all the extracted regions, the determination process is performed. The problem was that the accuracy was reduced.

[0004] In the conventional traffic control system, the end of an event or the like at a venue or a stadium is determined based on an increase in the number of passing vehicles on a nearby road or a length of congestion.
There was a delay from the timing when the event or the like was actually ended. Therefore, there has been a demand for a traffic control system that accurately determines that an event or the like has ended at an earlier timing and reflects the result in traffic control.

Further, in the conventional traffic control system, in order to detect the closing state of the railroad crossing, it is necessary to install information transmitting means for actively transmitting that the barrier is closed at the railroad crossing itself.

Therefore, a first object of the present invention is to provide an image recognition apparatus having high robustness against disturbances such as camera shake, light reflection, and shadow.

[0007] A second object of the present invention is to provide a traffic control system in which the end of an event or the like in a venue or a stadium is accurately judged at an earlier timing and reflected in traffic control.

A third object of the present invention is to provide a traffic control system that passively grasps a closed state of a railroad crossing and does not use the information transmitting means at the railroad crossing itself and reflects it in traffic control.

[0009]

In order to achieve the first object, first, a region where a change has occurred in an image is detected, and the area of the region where the change has occurred is calculated. Since the area may include an error due to disturbance, the error is corrected by suppressing a sudden change in the area value by the median value detection processing. Subsequently, by averaging the corrected area values over the past fixed period, errors are further dispersed and reduced, so that the accuracy of the situation determination in the image using the area can be improved.

[0010] In order to achieve the second object, a state in which a pedestrian moves near an exit or a passage of a venue or a stadium is photographed by an imaging device, and a luminance change in an image which changes according to a pedestrian flow rate. By grasping the pedestrian congestion status from the feature amount such as the area of the area and judging the end of the event etc., it is possible to grasp the end of the event etc. at an earlier timing with high accuracy and reflect it in traffic control. it can.

In order to achieve the third object, when a crossing alarm is photographed by an image pickup device and a feature amount in an image generated by the alarm flashing when the crossing interrupter is closed has reached a predetermined value. In addition, by determining that the level crossing barrier is closed, the traffic control system passively activates the level crossing barrier without the need for an information transmission means for transmitting that the level crossing barrier is closed on the level crossing side. The closure can be grasped and reflected in traffic control.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The invention according to a first aspect of the present invention is directed to an image input section for inputting images sequentially output from an image pickup apparatus, and a change detecting section for detecting an area where a change has occurred in the image. An area calculation unit that calculates an area by summing the number of pixels of a region where a change has occurred in the image; an area correction unit that reduces an error that occurs in the area; An area averaging unit that calculates the area, a determination unit that determines the state of the recognition target using the average area, and an output unit that outputs a result of the determination in the determination unit, and an area correction unit. By correcting and reducing errors included in the area of the area where some change has occurred in the image in the area averaging section, the determination section has the effect of improving the accuracy of the determination when determining the recognition target situation.

According to a second aspect of the present invention, a luminance change occurs in the image by an image input unit for inputting images sequentially output from the imaging device and a difference process between images at different imaging times. A change detection unit that detects an area, an area calculation unit that calculates an area by summing the number of pixels of the area where the luminance change has occurred, an area correction unit that reduces errors that occur in the area, An area averaging unit that calculates an average area by averaging periods, a determination unit that determines a state of a recognition target using the average area, and an output unit that outputs a determination result in the determination unit, By correcting and reducing errors included in the area of the area where the luminance change occurs in the image in the area correction unit and the area averaging unit, the determination unit improves the accuracy of the determination when performing the recognition target situation determination. Work Having.

According to a third aspect of the present invention, there is provided an imaging apparatus for capturing an image of a pedestrian moving, and a feature relating to the flow rate of the pedestrian by inputting the image and processing the image. An image recognition device that obtains an amount and determines the congestion state of the pedestrian using the feature amount, a transmission device that transmits a result of the determination by the image recognition device, and a traffic control using the result of the determination. It is equipped with a traffic control device that performs, an image of a pedestrian near the exit or passage of the venue and the stadium is taken by an imaging device, and the image is processed to acquire a feature amount related to the pedestrian, By determining the end of an event by grasping the pedestrian congestion status from the feature amount, it is possible to accurately and quickly determine the end of an event or the like performed at the venue and the stadium at an earlier timing. Having.

According to a fourth aspect of the present invention, there is provided a traffic control system according to the third aspect, wherein the image recognition apparatus is the image recognition apparatus according to the first or second aspect, and By determining the pedestrian congestion state with the area of the brightness change area in the image, which is caused by moving around the exit or passage of the venue and the stadium, as a feature amount, it is possible to accurately determine the situation at an earlier timing. This has the effect that it is possible to determine that an event or the like performed at the venue or the stadium has ended.

According to a fifth aspect of the present invention, there is provided an imaging apparatus for photographing an image of a level crossing alarm, inputting the image, acquiring a characteristic amount associated with blinking of the alarm, and acquiring the characteristic amount from the characteristic amount. An image recognition device that determines the closing state of the railroad crossing, a transmission device that transmits the result of the determination by the image recognition device, and a traffic control device that performs traffic control using the result of the determination, By capturing the image of the alarm and acquiring the characteristic amount associated with the blinking of the alarm when the level crossing barrier is closed, determining the closed state of the level crossing barrier from the characteristic amount, activates the situation at the level crossing itself There is an effect that it is possible to recognize that the railroad crossing is closed without having a means for transmitting information in a localized manner.

According to a sixth aspect of the present invention, there is provided a traffic control system according to the fifth aspect, wherein the image recognizing device is the image recognizing device according to the first or second aspect. At the time of closing, the warning device blinks and the area of the brightness change area generated in the image by the feature amount is determined as the feature value and it is determined that the level crossing is closed, so that the circuit breaker is actively transmitted to the level crossing side. This has the effect that it is possible to grasp that the railroad crossing is closed without having any means for performing the operation.

(Embodiment 1) FIG. 1 shows a first embodiment of the present invention.
3 shows a configuration of an image recognition device according to a first embodiment corresponding to claim 2. In FIG. 1, IP1 denotes an image input unit, which performs an input process of an image of a recognition target captured by an image capturing unit. IP2 is a change detection unit,
An area where some change has occurred in the image is detected. IP3
Is an area calculation unit, which calculates the area by summing all pixels for which a change has been detected by the change detection unit IP2. Area correction unit
In IP4, errors included in the area are corrected. The area averaging unit IP5 averages the corrected area over the past fixed period. The determination unit IP6 determines the state of the recognition target from the averaged area. The output unit IP7 outputs the result of the determination by the determination unit IP6 to the outside. Hereinafter, the change detection unit, the area calculation unit, the area correction unit, the area averaging unit, the determination unit, and the output unit will be described in detail.

(Change Detection Unit IP2) The processing of the change detection unit IP2 will be described with reference to FIG. In the change detection unit IP2,
Detect any changes that occur in the image. A description will be given of a case where a temporal change in luminance of an image obtained by the inter-frame difference processing or the background difference processing is used as a change in an image. First,
By using the image 1IM1 and the image 2IM2 at different shooting times,
The difference processing of the luminance value is performed between the two images. The difference image IM3 is obtained by the difference processing. Since an object that does not exist in the image 1 exists in the image 2, a difference occurs in the difference image. Next, threshold processing is performed on the difference image, and pixels having a threshold value or more are set to a luminance of 1 and other pixels are set to a luminance of 0.
The value is converted to a binary image IM4.

(Area Calculation Unit IP3) In the area calculation unit IP3,
The number of pixels having a luminance value of 1 is calculated using the binary image acquired by the change detection unit IP2. As a method for calculating the area, a method of scanning each pixel sequentially from the upper left of the image and counting the number of pixels having a luminance value of 1, or performing a labeling process on the pixel having a luminance value of 1, A method of summing the area obtained for each label for all labels can be used.

(Area Correction Unit IP4) The area correction unit IP4 corrects an error included in the area. The operation will be described with reference to FIG. Since various disturbances that cause an error have a property that occurs spontaneously without being continuous in time, correction is made using the property. The area values for the past n frames obtained from the area calculation unit IP3 are stored in the memory MEM1,
A median detection process MED for detecting the median of the area values is performed. The area obtained by the median value detection processing is defined as the area after correction. By this processing, it is possible to correct even when an error occurs in which some change occurs in the image due to a disturbance and the area of the change region in the image becomes excessively large or small once.

(Average area IP5) In the area average area IP5,
The errors included in the area value are further reduced. Area average
The processing in IP5 will be described with reference to FIG. The data for m frames, which is the area after the correction in the past fixed period, is held in the memory MEM2, and the averaging process AVG is performed on the data in the memory. The averaging process is performed by reading out all data in the past fixed period from the memory MEM2, adding the data, and dividing by the number m of area data.

(Determination Unit IP6) The determination unit IP6 performs a determination process using the average area obtained by the area average unit IP5. The situation in the image is determined by examining whether the average area indicates a predetermined value.

(Output Unit IP7) The judgment result by the judgment unit IP6 is output from the image recognition device through a predetermined interface such as RS232C, USB, IEEE1394.

As described above, in the image recognition apparatus according to the first embodiment of the present invention, errors included in the area of a region where some change has occurred in an image are corrected by the area correction unit IP4 and the area average unit IP5. By reducing, the judgment unit IP
6 has the effect of improving the accuracy of situation determination.

Although the change detection unit IP2 has described the case of a luminance change as a change in an image, a change in an optical flow, an edge, a texture, or the like may be used as a change in an image. . Also,
The area of a region where a change has occurred in the image can be replaced with an area ratio obtained by dividing the area by the total number of pixels of the image.

(Embodiment 2) Next, a second embodiment corresponding to claims 1 and 2 of the present invention will be described. The second embodiment of the present invention is an image recognition device having an image input unit IP1, a control unit CONT, and an output unit IP9, as shown in FIG. Control unit CONT
Has a change detecting step PR2, an area calculating step PR3, an area correcting step PR4, an area averaging step PR5, and a judging step PR6, and each step is the same as that of the first embodiment.
Change detection section IP2, area calculation section IR3, area correction section IR4
Each processing in the area averaging unit IP5 and the determination unit IP6 is executed by the CPU.
The processing flow is as shown in FIG. Image input unit IP1 and output unit IP
7 is the same as in the first embodiment.

As described above, in the image recognition apparatus according to the second embodiment of the present invention, errors included in the area of a region where a change has occurred in an image are corrected in the area correcting step PR4 and the area averaging step PR5. This has the effect of improving the accuracy of the situation determination performed in the determination step PR6.

(Embodiment 3) Next, claim 3 of the present invention will be described.
A third embodiment corresponding to claim 4 will be described. As shown in FIG. 5, the third embodiment of the present invention is a traffic control system including an imaging device CAM, an image recognition device IP, a transmission device TRN, and a traffic control device CTRL.

The imaging device CAM takes an image of a pedestrian WLK moving through an exit or a passage of a venue or a stadium. The image recognition device IP is, for example, the first embodiment or the second embodiment.
The image recognition device according to the embodiment, wherein the imaging device C
Using the image input from the AM, the congestion state of the pedestrian WLK is grasped by using the area of the brightness change area in the image that changes according to the flow rate of the pedestrian WLK as a feature amount, and the congestion state reaches a predetermined level. In this case, it is determined that the event or the like held at the venue and the stadium has ended. The end of the event or the like is transmitted to the traffic control equipment CTRL by the transmission device TRN, and the traffic is controlled according to the end of the event or the like.

As described above, in the traffic control system according to the third embodiment of the present invention, it is possible to accurately and quickly determine that an event or the like held in a venue or a stadium has ended.

Note that, in the third embodiment, the first
Although the example using the image recognition device described in the embodiment or the second embodiment has been described, aside from using the area of a region where some change occurs in the image as a feature amount for determination, Using an imaging device with an image sensor, multiple images with different viewpoints are taken, and areas having a high correlation are associated with each other, thereby performing three-dimensional sensing in the imaging space by stereo image processing. An image recognition device that determines that an event or the like has ended by grasping the congestion state based on a space occupancy rate or the like, which is a ratio of pedestrians in the space, may be used. In addition, an imaging device is installed at an exit or a passage of a racetrack, a boat racetrack, an amusement park, a theater, a baseball stadium, or the like, and images of pedestrians moving are taken, and it is determined that an event or the like has ended at each location. It is also possible.

(Embodiment 4) Next, claim 5 of the present invention will be described.
A fourth embodiment corresponding to claim 6 will be described. As shown in FIG. 6, the fourth embodiment of the present invention is a traffic control system including an imaging device CAM, an image recognition device IP, a transmission device TRN, and a traffic control device CTRL.

First, the crossing alarm CRS is photographed by the imaging device CAM. The image recognition device IP is, for example, a first
The image recognition device according to the embodiment of the present invention or the second embodiment, wherein an image obtained by photographing the level crossing alarm CRS by the imaging device CAM is input. When the circuit breaker is closed, the crossing alarm is closed when the feature amount becomes a predetermined value with the area or the like of the brightness change area generated in the image accompanying the blinking of the railroad crossing alarm CRS as the feature amount. Is determined.

As described above, in the traffic control system according to the fourth embodiment of the present invention, the image of the level crossing alarm CRS is photographed by the imaging device CAM, and the image recognition device IP
Since the closed state of the circuit breaker is determined by the above, means for transmitting the closed state from the railroad crossing itself to the outside becomes unnecessary.

In the fourth embodiment, the case has been described where the area of the brightness change area generated in the image is used as the feature quantity used by the image recognition apparatus. An image recognition device that uses a feature in which a luminance change area in an image alternates left and right or up and down with blinking, or uses color information as a feature since blinking of an alarm device is red. It may be.

[0037]

As described above, according to the present invention, as a first effect, an error included in the area of a change region in an image due to disturbance is corrected by an area correction unit and an area average unit. Since the reduction is achieved, an effect that a situation determination having high robustness against disturbance can be performed can be obtained.

As a second effect, a pedestrian moving at an exit or a passage in a venue, a stadium, or the like is photographed by an imaging device, and the area of a brightness change area in an image that changes according to the pedestrian flow rate is obtained. It is determined that an event or the like held at a venue or stadium has ended with the feature amount as a feature amount, so that the timing of ending an event or the like can be determined more accurately and earlier than in the past, and the effect can be reflected in traffic control Is obtained.

As a third effect, a level crossing alarm is photographed by an image pickup device, and a level crossing is determined by using the area of the brightness change area in the image caused by the flashing of the alarm when the circuit breaker is closed as a feature amount. Since it is determined that the railroad crossing has been closed, there is no information transmission means for transmitting that the circuit breaker is closed on the railroad crossing side. Can be

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image recognition device according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating an operation of a change detection unit according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram illustrating operations of an area correction unit and an area averaging unit according to the first embodiment of the present invention.

FIG. 4 is a block diagram illustrating a configuration of an image recognition device according to a second embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of a traffic control system according to a third embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a traffic control system according to a fourth embodiment of the present invention.

[Explanation of symbols]

 IP1 Image input unit IP2 Change detection unit IP3 Area calculation unit IP4 Area correction unit IP5 Area averaging unit IP6 Judgment unit IP7 Output unit IP Image recognition device IM1 Image1 IM2 Image2 IM3 Difference image IM4 Binary image MEM1 Memory1 MEM2 Memory2 MED Median value detection process AVG averaging process CONT control unit PR2 Change detection process PR3 Area calculation process PR4 Area correction process PR5 Area averaging process PR6 Judgment process CAM Imaging device WLK Pedestrian TRN transmission device CTRL Traffic control equipment CRS Level crossing alarm

Claims (6)

[Claims] An image input unit for inputting an image sequentially output from an imaging device, a change detection unit for detecting a region where a change has occurred in the image, and a pixel number of the region where the change has occurred are summed up. An area calculating unit that calculates an area by calculating an area, an area correcting unit that reduces an error occurring in the area, an area averaging unit that averages the area for a predetermined period in the past to calculate an average area, and a recognition target that uses the average area. An image recognition device, comprising: a determination unit that determines the state of (i), and an output unit that outputs a result of the determination by the determination unit. 2. An image input unit for inputting images sequentially output from an imaging device, a change detection unit for detecting an area where a luminance change has occurred in the image due to a difference process between images at different imaging times, An area calculation unit that calculates the area by summing the number of pixels of the area in which the luminance change has occurred, an area correction unit that reduces an error that occurs in the area, and an area that calculates the average area by averaging the area over a fixed period of time An image recognition apparatus comprising: an averaging unit; a determining unit that determines a state of a recognition target using the average area; and an output unit that outputs a result of the determination performed by the determining unit. 3. An imaging device for capturing an image of a pedestrian moving, inputting the image, processing the image to obtain a characteristic amount related to the flow rate of the pedestrian, and using the characteristic amount. A traffic control system including an image recognition device that determines the congestion state of the pedestrian, a transmission device that transmits a result of the determination by the image recognition device, and a traffic control device that performs traffic control using the result of the determination. 4. The image recognition apparatus according to claim 1, wherein
The traffic control system according to claim 3, which is the image recognition device according to (1). 5. An image pickup device for capturing an image of a level crossing alarm, image recognition for inputting the image, acquiring a characteristic amount associated with blinking of the alarm, and judging a closing state of the level crossing from the characteristic amount. A traffic control system including a device, a transmission device that transmits a result of the determination by the image recognition device, and a traffic control device that performs traffic control using the result of the determination. 6. The image recognition apparatus according to claim 1, wherein
The traffic control system according to claim 5, which is the image recognition device according to (5).