JP3767571B2 - Method and apparatus for controlling exposure of camera-type vehicle detector - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a camera-type vehicle detector that can detect a vehicle on a road by photographing a road with a camera and processing the image, and particularly relates to an exposure control method and apparatus for the camera.
[0002]
[Prior art]
As a vehicle detector for detecting a vehicle on a road, an ultrasonic type is well known. This ultrasonic type measures the vehicle by emitting ultrasonic waves from the top to the bottom of the road and detecting the arrival time of the reflected wave, but detects the characteristics of each vehicle. I can't do that.
Therefore, a camera-type vehicle detector is known that detects the characteristics of a vehicle and identifies a vehicle license plate by photographing a road with a camera and processing the image.
[0003]
[Patent Document 1]
Japanese Patent Laid-Open No. 2001-021958
[Problems to be solved by the invention]
The exposure control performed by this camera-type vehicle sensor changes the control contents at night, daytime, and morning and evening. At night, the exposure time of the camera is set to be short for the purpose of detecting the headlight, and for the purpose of photographing the vehicle body during the day, the brightness of the photographed screen is controlled to be constant. And in the morning time zone, when the daytime shifts from daytime to daytime, transient control is performed to gradually open the exposure.
[0005]
However, if the dawn time is mistakenly detected and the control switching timing is shifted, the screen will be dark even though it is in the morning and the body cannot be detected, or the head will be overexposed even at night. The light cannot be detected correctly.
SUMMARY OF THE INVENTION An object of the present invention is to provide an exposure control method and apparatus for a camera-type vehicle sensor that can accurately determine dawn.
[0006]
[Means for Solving the Problems]
In the exposure control method for a camera-type vehicle sensor according to the present invention, a screen shot by the camera-type vehicle sensor is input every predetermined time, and brightness data of the input screen or a plurality of screens input over the past time are input. Based on the luminance data, calculate the number of pixels whose luminance is less than or equal to the low luminance threshold (referred to as low luminance pixels), and set the maximum value from the number of low luminance pixels acquired from a certain past time to the present And determining whether or not the calculated number of low luminance pixels is lower than a predetermined ratio from the maximum value of the number of low luminance pixels, and when the lower than the predetermined ratio has continued for a certain time or longer. The exposure control is switched from night control to morning time zone control (claim 1).
[0007]
According to the above method, the maximum value of the number of low-luminance pixels is obtained, and the number of low-luminance pixels decreases as the night goes on, so the number of low-luminance pixels decreases from the maximum value below a predetermined ratio. Therefore, it is determined that the day has come to a dawn when it continues for a certain time or more, and the exposure control is switched from night control to control in the morning time zone.
The maximum value is obtained from the number of low luminance pixels acquired from a certain past time to the present time. Therefore, when the number of low luminance pixels increases, the maximum value is updated to a new value. become. Since the maximum value can be updated without being fixed in this way, the dawn can be accurately determined even when the brightness of the screen at night changes slightly.
[0008]
The number of low luminance pixels can be calculated by creating a histogram of the number of pixels versus luminance based on the input screen.
The low-brightness threshold can be obtained by extracting pixels from the lower luminance in the screen at night and setting the luminance of pixels exceeding a predetermined number as the low-brightness threshold. Further, at night, pixels can be extracted from the lower luminance in the screen, and the luminance that is one step lower than the luminance of the pixels exceeding the predetermined number can be set as the low luminance threshold. In the latter case, when the number of pixels changes greatly before and after the predetermined number, if the luminance of the pixels exceeding the predetermined number is set to the low luminance threshold value, the number of low luminance pixels is abnormally large, and this phenomenon is prevented. In order to achieve this, the luminance that is one step lower than the luminance of the pixels exceeding the predetermined number is set as the low luminance threshold.
[0009]
Even if the low brightness threshold is determined at night, the low brightness threshold may fluctuate if the weather changes at night. In this case, it is preferable to employ a low brightness threshold value in a time zone immediately before the dawn determination. Therefore, it is preferable to update the low luminance threshold value to a new value at night.
An exposure control apparatus according to the present invention is an apparatus according to the same invention as the exposure control method according to claim 1 (claim 6).
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
FIG. 1 is a schematic block diagram of a camera type vehicle sensing system to which an exposure control apparatus of the present invention is applied.
A camera-type vehicle sensor 1 is attached to an upper portion of a pole 2 installed in the vicinity of the road. A communication line 5 is connected from the camera-type vehicle sensor 1 to a personal computer 4 of a traffic information center 3 that collects traffic conditions on the road. The communication line 5 may be either wired / wireless or a combination thereof, and may be a dedicated line or a public communication line.
[0011]
FIG. 2 is a block diagram showing the internal configuration of the camera-type vehicle detector 1. The camera-type vehicle sensor 1 includes a camera 11, a camera DSP (Digital Signal Processor) that controls the operation of the camera 11, an image processing device 13, and a communication interface 14.
The camera 11 includes a lens system 15 and an image sensor 16. The aperture stop amount in the lens system 15 is adjusted based on the stop control signal from the camera DSP 12.
[0012]
The field of view of the camera 11 is directed to a road (not shown), and the camera 11 captures a road state and forms a captured image on the image sensor 16. Data of the image formed on the image sensor 16 is output to the camera DSP 12. The exposure time (corresponding to the shutter speed) of the image sensor 16 is adjusted based on the exposure time control signal from the camera DSP 12.
The image processing device 13 performs vehicle detection, license plate reading, vehicle speed detection, traffic jam length detection, and the like based on the captured image data of the camera 11 to obtain various information for grasping the traffic flow. These pieces of information are transmitted to the traffic information center 3 through the communication interface 14 and the communication line 5.
[0013]
Moreover, the image processing device 13 generates image data of a road surface area where no vehicle exists by deleting the vehicle from the captured image data of the camera 11. An image of a road surface area where no vehicle exists is referred to as a “reference image”. Then, the image processing device 13 executes exposure control of the camera 11 based on the reference image data. The result of exposure control is used for exposure time control of the image sensor 16 and aperture control of the lens system 15 through the camera DSP 12. The interval for creating and taking in the reference image data is hereinafter referred to as k (seconds).
[0014]
FIG. 3 is a graph for explaining an outline of exposure control performed by the image processing apparatus 13. The horizontal axis represents time, and the vertical axis represents luminance and exposure time. The exposure control is executed for each night, morning, daytime, and evening. At night, the purpose is to shoot a headlight, and as shown in FIG. 3, the exposure time is fixedly controlled. In the daytime, the purpose is to photograph the vehicle body and the license plate, and control is performed to keep the brightness of the reference image constant. This constant luminance is called “target luminance”.
[0015]
In the morning, control is performed to increase the exposure time in order to connect from night control to daytime control. The purpose of increasing the exposure time is to shoot a headlight at night, so the exposure time may be relatively short. However, in the morning, it is necessary to increase the exposure amount of the image sensor 16 to shoot a vehicle body or the like. This is because it is necessary to gradually increase the exposure time for this purpose. In the evening, control is performed to shorten the exposure time in order to connect from daytime control to nighttime control.
[0016]
Hereinafter, the procedure of the present invention for determining the timing for shifting from night control to morning control will be described in detail. This procedure is divided into a low-luminance threshold determination procedure performed during nighttime control, and a daylight determination procedure for determining the timing for shifting from nighttime control to morning control using the determined low-luminance threshold.
FIG. 4 is a flowchart for explaining a low-luminance threshold value determination procedure performed during nighttime control. First, the “reference image distribution number” is determined (step S1). This is the number of pixels serving as a reference for how many percent or less of the pixels of the entire screen shot at night counted from the lowest luminance are regarded as low luminance pixels.
[0017]
The procedure for determining the reference image distribution number will be described in detail with reference to FIGS.
FIG. 5A is a histogram created based on the luminance of the pixels on the reference screen. The horizontal axis represents luminance, and the vertical axis represents the number of pixels. This luminance is expressed by the number of gradations such as 256, for example. FIG. 5B is a graph showing a value obtained by integrating the number of pixels in the histogram with luminance (referred to as pixel distribution number).
FIG. 6 is a flowchart showing a procedure for determining the reference image distribution number. In step T1 in FIG. 6, the number of pixel distributions is added in order from the luminance 0 of the histogram. This is the same as tracing the curve of FIG. A constant A (for example, 20%) is set, and the pixel distribution number when the pixel distribution number exceeds the constant A is set to y (step T2; see FIG. 5B). Then, the number of pixel distributions with luminance one lower than y is set to x (step T3). y is compared with 2A (step T4). If y <2A, y is set as the reference image distribution number (step T5).
[0018]
By the way, since the shooting is performed at night, the luminance of the pixels on the reference screen often has the same value. Therefore, the pixel distribution number may rise steeply near the constant A. At this time, y may be 2A or more. In this case, if y is the reference image distribution number, y is too large, and the previous x is set as the reference pixel distribution number (step T6).
Next, referring back to FIG. 4, the upper limit value of the luminance of the pixel corresponding to the reference image distribution number is set as a low luminance threshold candidate (step S2). Then, the low-luminance threshold candidates determined as described above are compared with each other with respect to the latest N (for example, N = 10) reference images, and the majority of the low-luminance threshold candidates is determined. The candidate value is set as the “low luminance threshold” (step S3).
[0019]
Further, it is checked whether or not the low luminance threshold set in this way is equal to the previously set low luminance threshold (step S4). If they are equal, it can be determined that the night darkness in the field of view of the camera 11 is stable, so the low luminance threshold is stored in the memory of the image processing device 13 (step S5). If there is a low brightness threshold previously stored in memory, it is overwritten. In this way, the latest and stable nighttime low brightness threshold is preserved.
[0020]
In step S4, if the previously set low luminance threshold is not equal to the low luminance threshold set this time, it is considered that the environment has changed at night. For example, it has started to rain or snow, the moon has gone from cloudy to sunny. In this case, since it cannot be said that it is a stable night environment, the low luminance threshold is reset, and the low luminance threshold is created again by returning to the start.
Next, it is determined whether or not the time immediately before shifting to the dawn determination is reached (step S6). If the camera type vehicle sensing system has a clock, this determination may be made based on whether it is time for dawn. However, if the timepiece does not have a clock, the reference image data creation and capture time interval k (seconds) is known in advance, so the number of times the reference image data has been captured is determined, for example, by counting from the time of night. be able to.
[0021]
When the time immediately before shifting to the dawn judgment is reached, the latest low-luminance threshold value stored in the memory is fixed. As a result, the process shifts to dawn determination using the low brightness threshold value in the night environment immediately before dawn.
FIG. 7 is a flowchart for explaining the control procedure of the dawn determination. Briefly explaining the control procedure of this daylight determination, the number of pixels with lower luminance than the low luminance threshold is counted, and the number of pixels decreases as it becomes brighter. Is determined.
[0022]
First, the variable “maximum value” is set to 0 (step U1). In the pixels constituting the reference screen, the number of pixels having a luminance lower than the low luminance threshold is calculated, and this is calculated as the latest N images (N is an integer of 1 or more; N = 1 or 2 or more) ) For each (step U2). The number of pixels whose luminance is lower than the low luminance threshold is called “low luminance pixel number”. It is determined whether or not the number of low luminance pixels is greater than the maximum value (step U3). Since the maximum value is initially set to 0, the number of low luminance pixels is larger than the maximum value. This maximum value is overwritten in the memory and updated (step U4).
[0023]
Next, it is determined whether or not the number of low luminance pixels has decreased from a maximum value below a certain ratio (step U5). The constant ratio is set to 40%, for example. It is determined whether or not the state in which the ratio is lower than the predetermined ratio has continued for a predetermined time (step U6). For example, if the fixed time is 5 minutes, this corresponds to the above-described decrease determination being continued 30 times when the reference image capturing period k is 1 second and N = 10. Further, assuming that the predetermined time is 30 minutes, the reference image capturing period k is 1 second, and N = 10, this corresponds to the reduction determination being continued 180 times.
[0024]
If it continues for a certain period of time, it is determined to be dawn, and exposure control of the camera 11 is switched to morning control (see FIG. 3) (step U7). If not continued, the process returns to step U2 to continue the update of the maximum value and the determination of a constant ratio decrease from the maximum value.
FIG. 8 is a graph in which the number of low luminance pixels is plotted against time. The horizontal axis represents time t, and the vertical axis represents the number of low luminance pixels. The maximum value of the number of low-luminance pixels that is continuously updated is represented as “maximum value” on the vertical axis. It is determined that the dawn has ended at a time point t0 when a state where the ratio falls below 40% from the maximum value continues for a certain time T.
[0025]
By the above determination procedure, the number of low-luminance pixels decreases as the night dawns. Therefore, when the number of low-luminance pixels has decreased from the maximum value to a predetermined percentage, the night has come By determining, the exposure control can be switched from night control to control in the morning time zone.
Although the embodiments of the present invention have been described above, the embodiments of the present invention are not limited to the above-described embodiments. For example, in the camera type vehicle sensing system, the image processing device 13 is provided in the camera type vehicle sensor 1, but the present invention is not limited to this, and is connected to the camera type vehicle sensor 1 through the communication line 5. It may be provided in the personal computer 4. In addition, various modifications can be made within the scope of the present invention.
[0026]
【The invention's effect】
As described above, according to the present invention, since accurate dawn determination can be performed, the exposure control logic can be smoothly switched from night control to dawn control.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram of a camera type vehicle sensing system to which an exposure control apparatus of the present invention is applied.
FIG. 2 is a block diagram showing an internal configuration of the camera-type vehicle detector 1;
FIG. 3 is a graph for explaining an outline of exposure control;
FIG. 4 is a flowchart for explaining a low-luminance threshold value determination procedure performed during nighttime control.
5A is a histogram created based on the luminance of pixels on a reference screen, and FIG. 5B is a graph showing the number of pixel distributions obtained by integrating the number of pixels in the histogram with luminance.
FIG. 6 is a flowchart illustrating a procedure for determining a reference image distribution number.
FIG. 7 is a flowchart for explaining a control procedure for dawn determination;
FIG. 8 is a graph in which the number of low luminance pixels is plotted against time.
[Explanation of symbols]
1 camera type vehicle detector 2 pole 3 traffic information center 4 personal computer 5 communication line 11 camera 12 camera DSP
13 Image Processing Device 14 Communication Interface 15 Lens System 16 Image Sensor

Claims (6)

Enter the screen shot with the camera-type vehicle detector every predetermined time,
Based on the luminance data of the input screen or the luminance data of a plurality of screens input over the past time, the number of pixels whose luminance is equal to or lower than the low luminance threshold (referred to as the low luminance pixel number) is calculated.
Find the maximum value from the number of low luminance pixels acquired from a certain past time to the present,
Determining whether the calculated number of low luminance pixels is lower than a predetermined ratio from the maximum value of the number of low luminance pixels;
An exposure control method for a camera-type vehicle detector, wherein the exposure control is switched from night control to morning time zone control when the lowering of the predetermined ratio continues for a predetermined time or more. 2. The exposure control method for a camera-type vehicle sensor according to claim 1, wherein the number of low luminance pixels is calculated by creating a histogram of the number of pixels versus luminance based on an input screen. 3. The low luminance threshold is characterized in that, at night, pixels are extracted from the lower luminance in the screen, and the luminance of pixels exceeding a predetermined number is set as the low luminance threshold. Exposure control method for a camera-type vehicle sensor. The low-brightness threshold is characterized in that, at night, pixels are extracted from the lower brightness in the screen, and the brightness that is one step lower than the brightness of pixels exceeding a predetermined number is set as the low-brightness threshold. An exposure control method for a camera-type vehicle detector according to claim 2. 2. The camera-type vehicle sensor exposure control method according to claim 1, wherein when the low-brightness threshold fluctuates at night, the low-brightness threshold is updated to a new value. Screen luminance calculating means for calculating the luminance of the screen based on the data of the image taken by the camera of the camera-type vehicle detector;
Based on the screen brightness data calculated by the screen brightness calculation means or the brightness data of a plurality of screens input over the past time, the number of pixels whose brightness is lower than the low brightness threshold (referred to as the number of low brightness pixels) Low luminance pixel number calculating means for calculating
The maximum value is obtained from the number of low luminance pixels acquired from a certain past time to the present time, and the calculated number of low luminance pixels is lower than a predetermined ratio from the maximum value of the low luminance pixels And a control switching means for switching the exposure control from night control to morning time zone control when it has been determined that it has fallen below the predetermined ratio for a certain period of time or longer. An exposure control device for a vehicle sensor.

Images