JPH03201098A - Mobile body detecting method - Google Patents

Abstract

PURPOSE:To obtain a method scarcely affected by the change of outside luminance by supplying time difference to the shutter speed of a first and second video signals, and finding information with respect to a mobile body based on a differential value obtained by processing the difference of density data. CONSTITUTION:Incident light is distributed in two directions with a half mirror 1, and an electric charge accumulated in a first CCD 2A is read out in an ordinary mode as the first video signal, and the electric charge accumulated in a second CCD 2B is read out in a shutter mode as the second video signal. Then, after noise components are eliminated from the signals at first and second LPFs 4A, 4B, they are converted to first and second digital density data. A table conversion part 6 performs the multiplication by constant times of the second digital density data so as to go to the same level as that of the first digital density data, and an MPU 10 finds the information with respect to the mobile body based on the differential value obtained by processing the difference of the digital density data. In such a way, it is possible to easily and stably find the information with respect to the mobile body being affected scarcely by the change of the outside luminance even outside a room.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a moving object detection method capable of processing image data to obtain information regarding a moving object such as a vehicle.

[Conventional technology]

Conventionally, image data such as roads taken with a television camera is subjected to binarization processing, and difference processing is applied to the binarized image data after a predetermined period of time has elapsed from this binarized image data. There is a method that calculates information regarding a moving object (vehicle), such as vehicle detection, amount of change, and moving speed, based on a change in brightness at a specific position of the difference value obtained.

As such a moving object detection device, for example, Japanese Patent Laid-Open No. 60
There is one shown in Japanese Patent No.-27997.

[Problem to be solved by the invention]

However, in the case of the above-mentioned conventional moving object detection device, since the difference processing is performed after the binarization processing, if the density difference between the vehicle and the background is small, information regarding the vehicle cannot be stably measured.

Furthermore, outdoors, the brightness of the background changes significantly depending on the weather or time of day, making it difficult to set the darkness value to an appropriate value for binarization processing, making it difficult to obtain accurate image data. There was an inconvenience that it would no longer be possible.

This invention was made to solve the above-mentioned problems, and is a moving object detection method that is hardly affected by changes in external brightness even outdoors and can easily and stably obtain information about moving objects. The present invention provides a method.

[Means to solve the problem]

The moving object detection method according to the present invention distributes incident light in two directions and supplies it to the first and second photoelectric conversion elements, and the first incident light is detected at the first incident light sensing time when the moving object is blurred. A first video signal is read out from the photoelectric conversion element of The first digital density data and the second digital density data obtained by reading out and processing the first and second video signals respectively
Information regarding the moving object is obtained based on the difference value obtained by the difference processing of the digital density data.

[For production]

The moving object detection method in this invention includes first and second
The difference value obtained by the difference processing of the first digital density data and the second digital density data obtained by processing the video signals of , that is, the blur component of the moving object, and the first and second incident light sensing Information about the moving object can be determined based on the time difference.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

FIG. 1 is a block diagram showing the configuration of a moving object detection device to which an embodiment of the present invention is applied. Reference numeral 1 indicates a half mirror as an incident light distribution means. This is to distribute the incident light photographed in two directions at the same time.

2A and 2B are first and second solid-state image sensors (CODs) to which the respective incident lights distributed by the half mirror 1 are supplied.
), and basically has a structure in which a large number of electrodes are arranged on a semiconductor substrate via an insulating film, and as shown in the explanatory diagram of FIG. 2, the photosensitive portion 21A.

21B, . . . and a transfer section 22A that transfers an electric signal (hereinafter referred to as charge) corresponding to the amount of incident light incident on the photosensitive sections 21A, 21B, . . . during the horizontal blanking period and the vertical blanking period. 22B, . . . and a transfer section 23.

Regarding charge transfer, please refer to the literature such as "rCCD Camera Technology (Hiroo Takemura, Radio Gijutsusha)".

To explain the shutter function using a CCD, in the case of the first incident light sensing time (ill normal mode), for example, the readout pulse (vertical blanking The photosensitive sections 21A, 21B, .
, . . . and transferred via the transfer unit 23.

In addition, in the case of the second incident light sensing time (shutter mode), as shown in FIGS. 4(a) to 4(C), the photosensitive portion 21A is activated in synchronization with, for example, the horizontal blanking pulse.

By resetting the charges accumulated by 21B, . . . , the photosensitive portion 21A is
, 21B, ... shorten the accumulation time for accumulating charges,
The charge accumulated by the read pulse is transferred to the transfer unit 22A,
22B, . . . and transferred via the transfer section 23, high-speed shutter operation is realized.

Therefore, the horizontal blanking pulse causes the photosensitive area 2 to
The shutter speed can be controlled by controlling the number of times the charges accumulated in 1A, 21B, . . . are reset.

For details regarding this shutter speed control, please refer to "High Resolution CCD" published by the Tokai Branch Federation of Electrical Related Societies, October 1989.

In the normal mode, the moving object is blurred, for example, 1/60th of a second, and in the shutter mode, the moving object is in a stationary state, for example, 1/1000th of a second.

Furthermore, in FIG. 1, 3A and 3B are the first and second
The first CCD control section 3A performs control to read out the charges accumulated in the first CCD 2A as the first video signal in the normal mode described above, and the second The CCD control unit 3B of
The charge accumulated in 2B is transferred to the second
This controls the reading of the video signal as a video signal.

4A and 4B are first and second low pass filters (
L P F), the first ■, PF4A is the first CC
This is to remove the noise component of the first video signal outputted by the D control section 3A, and to prevent aliasing noise in the first analog-to-digital converter' (AD converter) 5A, which will be described later. 2 LPF4B is the second C
This is to remove the noise component of the second video signal output by the CD control section 3B, and to prevent aliasing noise in the second AD converter 5B, which will be described later.

5A and 5B are first and second AD converters that convert the first and second video signals output by the first and second LPFs 4A and 4B into digital density data, respectively; 6 is a second AD converter A table conversion unit 7A, 7 that multiplies the second digital density data outputted by 5B by a constant, for example.
B indicates first and second area adders, and the first area adder 7A converts the first digital density data output from the first AD converter 5A in a direction perpendicular to the moving direction of the moving object;
For example, in the width direction of the vehicle (see FIG. 5(a), which will be described later).

(C) and (in the horizontal direction in 81), the second area addition section 7B outputs the digital density data output from the table conversion section 6 in the horizontal direction. This outputs a one-dimensional second added density value added to .

Reference numeral 8 denotes a subtracter for subtracting the second addition density value output from the second area addition unit 7B from the first addition density value output from the first area addition unit 7A; 9 represents a position comparison unit; The position comparator 9 uses a comparator or the like to calculate the rising points of the first and second added density values obtained by the first and second area adders 7A and 7B using a comparator or the like. In other words, the tip portion (position) of the vehicle is determined.

Reference numeral 10 denotes a microprocessor (MPU), which calculates information such as vehicle detection, amount of change, and moving speed based on the outputs of the subtracter 8 and the position comparator 9.

FIGS. 5(a) to (f) are diagrams showing digital image diagrams of each part and added density values obtained by adding digital density data in the horizontal direction, and FIGS. 5(a), (C), and (e)
The lower side corresponds to the traveling direction of the moving object (vehicle M).

FIG. 6 is a diagram showing an example of a conversion data table of the table conversion section 6.

Next, the operation will be explained.

First, the incident light whose focus and iris are adjusted by a TV camera lens etc. is distributed in two directions by the half mirror 1, and the incident light distributed to each direction is supplied to the first and second CCDs 2A and 2B. .

At this time, as described above, the charges accumulated in the first CCD 2A are converted into the first video signal (see FIG. 5(a)) in the normal mode.
), and the charge accumulated in the second CCD, 2B, is read out as a second video signal (Fig. 5 (C)) in shutter mode.
Read as .

In this way, the first and second video signals read out from the first and second CCDs 2A and 2B have their noise components removed by the first and second LPFs 4A and 4B, respectively. The data is converted into first and second digital density data by second AD converters 5A and 5B.

However, since there is a difference in shutter speed between the first and second digital density data, if you take the difference between them, a background component that does not move will remain in the difference image, so the time difference in shutter speed between normal mode and shutter mode is Since it is constant, the table converter 6 converts the second digital density data so that it has the same level as the level of the first digital density data (the level of the background component) output by the first AD converter 5A. Multiply by a constant.

The first and second parts adjusted to the same level in this way
Since the digital density data of
The area adding units 7A and 7B move within the detection area indicated by the dashed line in a direction perpendicular to the moving direction of the vehicle M, as shown in FIG. 5(a).
Since the projection results are added in the horizontal direction (in (C) and (e)), the first and second added density values are as shown in FIG. 5(b).
, (as shown in dl, it becomes a one-dimensional added density value.

At this time, some blurring due to the normal mode occurs in FIG. 5(b).

Then, the subtracter 8 outputs an added density value obtained by subtracting the first added density value from the second added density value, that is, an added density value with only the blur portion shown in FIG. 5(f). The image as a whole becomes a digital image shown in FIG. 5(Q).

Further, the position comparison unit 9 compares the first and second added density values with the rising point of the added density value (point yF in FIG. 5(b)).
b, point yz in FIG. 5(C)) is determined and output.

Therefore, MPU10 is based on the output of the subtracter 8,
Width y in which the added density value rises and returns as shown in FIG. 5(f)
. (amount of change), and calculate the width y0 by subtracting the time T2 shown in Fig. 4 (J) from the preset shutter speed time difference (time T shown in Fig. 3 (C)).
By dividing, the moving speed of the vehicle M can be obtained, and based on the output of the position comparison unit 9, it is compared whether the positions of the point yFb and the point y8 are the same, and it is determined whether both points yF and y□ are the same. If it is the location, it is determined that vehicle M exists.

Thus, according to this embodiment, the first and second C
By giving a time difference to the shutter speeds of the first and second video signals read from the CDs 2A and 2B, that is, rather than focusing on temporal changes in brightness, the vehicle M is extracted from multiple images at the same location at the same time. As a result, information such as the moving speed, position, and presence of the vehicle M can be easily and stably obtained outdoors without being affected by changes in external brightness due to the movement of the sun, clouds, etc.

Also, since the data is converted into a one-dimensional structure, the vehicle M
Since no time is required for data processing even if the speed is high, tracking of the vehicle M becomes easy.

FIG. 7 is a block diagram showing the configuration of a moving object detection device to which another embodiment of the present invention is applied, and the same parts as in FIG. 1 are given the same reference numerals.

In FIG. 7, reference numeral 11 indicates that the first digital density data obtained by adding the first digital density data output from the first AD converter 5A for a predetermined frame is added to the first area adding unit 7A.
12 is a delay unit that delays the second digital density data output from the second AD converter 5B by a predetermined frame amount and a predetermined frame value added by the adder unit 11, and outputs the delayed data to the table conversion unit 6. shows.

In this moving object detection device, when there is almost no difference between the first video signal and the second video signal even if the shutter speeds are different, that is, when the vehicle M is traveling at a low speed due to traffic jams, the adding unit 11 , the first digital density data is added for a predetermined frame, for example, several frames, to increase the apparent shutter speed, and the delay section 12 adds the first digital density data for a predetermined frame, for example, several frames.
By delaying the digital density data by several frames to make the leading position of the vehicle M the same, processing similar to that of the moving object detection device shown in FIG. 1 can be performed. A similar effect can be obtained.

In addition, in the above-mentioned moving object detection device, the first incident light sensing time is 1/60th of a second, and the second incident light sensing time is 100 seconds.
Although explained as 1/0 second, it is not limited to this time, and the first incident light sensing time may be any time when the moving object is blurred, and the second incident light sensing time is the time when the moving object is blurred. Any time period during which the object is in a stationary state is sufficient.

Although the example in which the half mirror 1 is used as the incident light distribution means has been described, a prism or the like that functions in the same manner may also be used.

Further, although an example has been described in which the first and second C0D2A and 2B are used as the first and second photoelectric conversion elements, other photoelectric conversion elements that function similarly may be used.

Further, the table conversion unit 6 may refer to the first digital density data and multiply the second digital density data by a constant so that the high density levels become the same level, or the table converter 6 may multiply the second digital density data by a constant so that the high density level becomes the same level, or the table converter 6 may refer to the first digital density data and multiply the second digital density data by a constant, Based on
The table converter 6 may be omitted by correcting the reference level to be determined (FIG. 5 (portion L of fl)).

Further, although the presence of the vehicle M was determined based on the output of the position comparison unit 9, the presence of the vehicle M can also be determined based on the output of the subtractor 8 based on the amount of blur.

Furthermore, although the digital density data has been described with a one-dimensional structure, it may be a two-dimensional structured digital density data.

〔Effect of the invention〕

As described above, according to the present invention, incident light is distributed in two directions and supplied to the first and second photoelectric conversion elements, and the first incident light is detected at the first incident light sensing time when blurring of a moving object occurs. A first video signal is read from one photoelectric conversion element, and a second video signal is read from the second photoelectric conversion element at the same time as the first video signal at a second incident light sensing time when the moving object is in a stationary state. Information regarding the moving object is obtained based on the difference value obtained by differential processing between the first digital density data and the second digital density data obtained by reading and processing the first and second video signals, respectively. Even outdoors, information about moving objects can be obtained easily and stably, almost unaffected by changes in external brightness.

Furthermore, even if the moving object is moving at high speed, no time is required for data processing, so there is an effect that tracking of the moving object becomes easier.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a moving object detection device to which an embodiment of the present invention is applied. FIG. 2 is an explanatory diagram showing the configuration of a solid-state image sensor (CCD). ) is an explanatory diagram of the first incident light sensing time, FIGS. 4(a) to (C) are explanatory diagrams of the second incident light sensing time, and FIGS. 5(a) to (f') are additive density values. FIG. 6 is a diagram showing an example of a conversion data table of the table conversion section, and FIG. 7 is a block diagram showing the configuration of a moving object detection device to which another embodiment of the present invention is applied. DESCRIPTION OF SYMBOLS 1... Half mirror 12A... 1st solid-state image sensor, 2B... 2nd solid-state image sensor, 3A... 1st solid-state image sensor control part, 3B... 2nd solid-state image sensor Element control section, 4A...first low pass filter, 4B...
・Second low bass filter, 5A...First analog-to-digital converter, 5B...Second analog
Digital converter 1.6...Table conversion section, 7A.
...first area addition section, 7B...second area addition section,
8... Subtractor, 9... Position comparison section, 10... Microprocessor unit, 11... Addition section, 12
...Delay part. 1A 2A 1B 2B Figure 2 4 Figure (a) (C) (e) Added density value (b) Added density value (d) Added density value (f)

Claims (1)

[Claims] The captured incident light is divided into two directions and the first and second
reading a first video signal from the first photoelectric conversion element at a first incident light sensing time when the moving object is blurred; and at a second time when the moving object is in a stationary state. At the incident light sensing time, a second video signal is read out from the second photoelectric conversion element at the same time as the first video signal, and the first and second video signals are processed, respectively. A moving object detection method, comprising: obtaining digital density data of the first digital density data and second digital density data, and obtaining information regarding the moving object based on the difference value obtained by differential processing between the first digital density data and the second digital density data.