JP3268096B2 - Image-based stop vehicle detection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image-based vehicle sensing method for processing a video signal obtained by capturing an image of a road and detecting a vehicle on the road by changing a luminance level corresponding to the same position on the road. In particular, the present invention relates to a method for appropriately detecting a vehicle that has been stopped for a long time.

[0002]

2. Description of the Related Art A method and apparatus have been put to practical use in which a road is imaged by an image pickup means in order to grasp a traffic situation, and an obtained video signal is processed to detect a vehicle passing through the road. Such an apparatus is disclosed, for example, in Japanese Patent Application Laid-Open
No. 188005 (Japanese Patent Application No. 2-31892: image-type vehicle detector). In addition, the present applicant has disclosed a method of appropriately extracting a luminance level, that is, reference road surface level data in the case of only a road surface without a passing vehicle, which is included in an image processing process for vehicle detection, by Japanese Patent Application Laid-Open No. H05-205,197.
No. 40818 (Japanese Patent Application No. 3-43011: image-type vehicle sensing method) has already been proposed. In these proposals, the vehicle is detected by comparing the pattern correlation between the road luminance pattern and the vehicle luminance pattern of the measurement line.

In the image-type vehicle sensing, for example, an image-type vehicle sensing device shown in the block diagram of FIG. 15 is used. The illustrated image-type vehicle sensing device has substantially the same configuration as that of the above-described proposed device, and generally includes an imaging unit 10, a luminance data conversion unit 20, an arithmetic processing unit 30, and a control unit 40. FIG.
Is a diagram showing an example of an actual arrangement.
0 (CCD camera) is installed, and a video signal of an image of the road is input to the image processing units (20, 30, 40). 70
Is a near-infrared illumination device for night operation.

The image pickup means 10 is, for example, a CCD video camera. The image pickup means 10 picks up an image of a road detection area as shown in FIG. 3A from above the road and converts the corresponding NTSC video signal into an image processing section. To the luminance data conversion unit 20.

[0005] The image processing unit uses this video signal to move to a plurality of measurement point positions set on a straight line (measurement line, L) in the image shown in FIG. Based on the difference between the luminance data groups obtained at predetermined time intervals, the presence of the vehicle on the measurement line (L) at a certain time is determined based on a difference between the luminance data groups obtained as a set of luminance levels in the luminance signal corresponding to time. To detect.

That is, the reference road surface level data (Dr) determined from the luminance data group (FIG. 3B: road luminance pattern) when no vehicle exists on the measurement line, and the vehicle exists on the detection line. The luminance data group when the luminance level changes greatly along the detection line (FIG. 3C:
The vehicle is detected by detecting a difference from the vehicle brightness pattern. In addition, set multiple detection lines,
The speed of the vehicle can also be obtained based on the vehicle arrival time at both sides.

The luminance data converter 20 is a synchronizing signal extracting circuit 21 for obtaining a horizontal synchronizing signal and a vertical synchronizing signal from a video signal. Based on these synchronizing signals, the present scanning position on a scanning line is digitized and corresponding horizontal coordinates are obtained. A horizontal address counter 22 and a vertical address counter 23 for obtaining vertical coordinates; a digital comparator 24 for comparing these outputs with the coordinates sequentially designated by the microprocessor 41 and generating a capture signal at the moment of coincidence; It comprises a video A / D conversion circuit 25 for digitizing the luminance at the corresponding instant of the video signal according to the signal and outputting the luminance data. The input unit is provided with a video amplifier 27 for amplifying the transmitted video signal to an appropriate level, and a clamp level fixing circuit 28 for reproducing the DC level of the video signal prior to A / D conversion. I have.

The arithmetic processing unit 30 is a circuit for receiving luminance data from the video A / D conversion circuit 25 and averaging a plurality of luminance data and calculating a correlation. Therefore, a signal processor 31 is used, and a storage unit necessary for storing data is also provided.

The control unit 40 controls the above-mentioned units, receives the data from the arithmetic processing unit 30 and performs further arithmetic processing on the data to detect the presence of a vehicle, calculate the speed of the vehicle, and perform the following processing. A microprocessor 41, a ROM storing a control program, which is output to the device;
42, a RAM 43 for data storage, and an I / O circuit 44 for output. Reference numeral 45 denotes a CPU bus through which the microprocessor 41 exchanges data and addresses, control commands and response signals with other parts.

[0010] This device also includes a DC power supply 5.
1. Clock circuit 52 supplied to each unit, imaging unit 10
D / A that sends out a signal for increasing the sensitivity of the imaging means 10 when the level of the video signal from
A conversion circuit 53 and the like are provided.

In order to detect a running vehicle by processing a video signal using such an apparatus, an image-based vehicle sensing method generally includes the following steps. That is, 1) a video signal corresponding to an image of a region including a road to be sensed is obtained by imaging means such as a CCD camera. 2) In the video signal, a plurality of positions (temporal positions in the video signal) corresponding to measurement points located on a straight line (measurement line) in the cross direction of the road are set in advance, and luminance data groups corresponding to these positions are set. Are sequentially obtained. 3) A luminance data group corresponding to only a road surface on which no vehicle exists on the measurement line (detection line) is extracted at a predetermined time and held as reference road surface level data. The reference road surface level data is appropriately corrected with the passage of time according to the conditions such as the illuminance of the road. 4) The luminance data group along the measurement line is compared with the reference road surface level, and if a sufficient difference is recognized (both high and low cases), it is assumed that a vehicle is present and a sensing signal is transmitted.

As described above, in the conventional image-based vehicle sensing method, the reference road surface level extraction (learning of the reference road luminance pattern) is performed at a constant frequency in order to increase the detection accuracy. For example, Japanese Unexamined Patent Application Publication No.
In the proposal of No. 818, when the same luminance pattern of the measurement line is intermittently obtained as shown in FIG. 6, it is adopted (updated) as a reference road luminance pattern and held. As a result, in a traffic situation where only a road image can be obtained between vehicle images on the screen portion corresponding to the measurement line during a traffic jam, etc., and a sufficient level of road luminance level cannot be obtained. However, it accurately learns the reference road luminance pattern.

However, according to the conventional method as well as the above method, when the vehicle stops on the measurement line for a long time,
This also includes the reference road surface level, and the learning of the true reference road luminance level cannot be performed. As a result, if the vehicle has been stopped (parked) on the measurement line for a long time, or traffic has become congested, and the vehicle has been on the measurement line for a longer time, the state of the vehicle, that is, the vehicle The luminance level in a state including the luminance level to be extracted is extracted as a reference road surface level, and as a result of comparing the reference road surface level group and the luminance level group obtained thereafter, there is no difference, and the actual The inconvenience that sensing cannot be performed (only a semi-presence operation can be expected) occurs.

When the vehicle exits from a long stop, the luminance data of the reference road is determined taking into account the existence of the vehicle, and is not yet luminance data corresponding to a true road. In this state, the sensing output is kept output even if there is no vehicle. For this reason, in the conventional method of this type, accurate vehicle detection can be performed only when the stop time of the vehicle is about 20 seconds or less. This is because the detection level (several%) drifts during the road luminance pattern of 20 seconds.

[0015]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described situation, and can reliably continue a sensing state even when a vehicle is present on a measurement line for a long time, that is, a presence operation is possible. An object of the present invention is to propose a method for detecting a stopped image vehicle. According to the present invention, even if the vehicle keeps stopping on the measurement line for a long time such as 30 minutes or several hours, it is possible to continuously detect the stopped vehicle.

[0016]

In order to solve the above-mentioned problems, in the image-based stopped vehicle sensing method according to the present invention, a video signal corresponding to an area including a road to be sensed is obtained from an image pickup means, and the image signal is displayed on a screen. (Dt) is sequentially obtained as a group of luminance data at positions in the video signal corresponding to a plurality of measurement points temporally corresponding to a plurality of measurement points located on a detection line assumed in the cross direction of the road, and a vehicle exists on the detection line. If not, extract a luminance data group corresponding to the reference road surface level data (Dr) at a predetermined frequency.
Is performed, and a luminance data group (D
An image type vehicle that sends out a sensing signal indicating that a vehicle is present on the measurement line when a difference of not less than a predetermined value is found by comparing the luminance data group (Dv) based on t) with the reference road surface level data (Dr). In the sensing method, when the state of sensing the vehicle continues for a predetermined time (T1) or more, the vehicle shifts to a vehicle stop mode in which only the update processing of the reference road surface level data (Dr) is stopped. When the difference of the brightness data group (Dv) based on the obtained brightness data group (Dt) with the reference road surface level data (Dr) no longer exceeds a predetermined value, the transmission of the sensing signal is stopped and the brightness data The group is promptly updated with the reference road surface level data (Dr) to release the vehicle stop mode.

[0017]

In the above-described method, first, the vehicle is detected in the same vehicle detection process as in the prior art, and the state in which the vehicle is detected is maintained for a predetermined time (T
1) Detect the case where the above is continued, and shift to the vehicle stop mode. That is, only the update processing of the reference road surface level data (Dr) is stopped, and the determined vehicle stop determination operation is repeated until the vehicle exits using the same reference road surface level data (Dr). As a result, it is possible to continue to reliably sense a long-time stop state of the vehicle including parking and the like. Of course, when the vehicle exits (when the vehicle is no longer detected), the reference road surface level data (Dr) does not match the actual situation. (Cancel the vehicle stop mode). If the vehicle is not in the stop mode,
It is possible to perform vehicle detection, speed measurement of a traveling vehicle, and the like, which are equivalent to those in the related art.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the accompanying drawings. The apparatus suitable for carrying out the method of the present invention is substantially the same as the conventional apparatus, and for example, the apparatus shown in FIG. 15 described above can be used. The outline of each part configuration is
As already explained. Of course, the control content for vehicle detection, that is, the vehicle detection method is different.

In operation of the method of the present invention, a measurement line is set so as to be orthogonal to the direction of the road to be imaged, as schematically shown in FIG. The figure shows a case where the cross-section traffic flow measurement up to four lanes is performed. As for the measurement line, a start measurement line A, a stop measurement line B, and a plurality of measurement lines are set between the measurement lines, and are sequentially scanned and processed.
This is for detecting a stopped vehicle in a wide range, but is an additional technical problem, and will be described later. First, the detection of the stopped vehicle is performed on a single arbitrary measurement line (for example, the measurement line S). I will mainly explain what you do.

FIGS. 1 and 2 are flow charts showing an embodiment of an image-based stopped vehicle sensing method according to the present invention. In the method according to the embodiment, first, a video signal corresponding to an image of an area including a road to be sensed is obtained by an imaging unit such as a CCD camera as in the related art.

Next, in the video signal, positions (temporal positions in the video signal) corresponding to a plurality of preset measurement points located on a predetermined scanning line (almost a straight line in a direction transverse to the road: measurement line). Are sequentially obtained (S
1, S2). In the embodiment, a target measurement line is sequentially scanned in a frame unit toward the lower part of the screen at a predetermined cycle, and image A / D data of a different measurement line is taken in for each frame. (Details will be described later. Not shown in the flowchart.)

Next, the above-mentioned luminance data group along the measurement line (scanning measurement line) is compared with the stored reference road surface level data (Dr: how to determine and update timing will be described in detail later). If a difference is found (both high and low), the vehicle is assumed to be present and a sensing signal is sent (S4, S5). In addition,
If it is already in the sending state, this means continuing this. If there is not much difference between the reference road surface level data and the luminance data group, no sensing signal is transmitted (S6). If the sensing signal is being sent, the sending is stopped (released).

Since the absolute level of the luminance of the measurement line changes according to the illuminance of the outside world, a correct reference road luminance pattern cannot be learned. However, the illuminance of the external world may be considered to be uniform (increase / decrease uniformly) on the measurement line. Therefore, even if the absolute level of the road luminance pattern changes, the overall distribution (that is, the form of the luminance level: pattern) will be the external world. It does not change depending on the illumination. In the example,
Taking this idea into account, both the luminance data group and the reference road surface data are normalized (patterned) and then compared (pattern correlation comparison).

That is, as shown in FIG. 4, an average value (a) of the luminance level group of the measurement line is obtained, and a value (pattern) obtained by normalizing with this average value is used as a luminance change rate pattern.
However, the normalized luminance data is obtained from Rn (i) = R (i) / a (1). Here, here, R (i): measured luminance data Rn (i): normalized luminance data On the other hand, the reference road surface data used for comparison is normalized in exactly the same manner to determine a change rate pattern, and then the two patterns are compared to detect the presence of a vehicle. Thereby, even if the illuminance changes, there is no malfunction and the reliability is increased.

The process following the above process (mainly the process of updating or keeping the reference road surface level data (Dr) as it is) greatly differs depending on whether a vehicle is detected in the above process. Above all, a process corresponding to a case where a vehicle is detected or a case where the vehicle is departed after stopping for a long time, which is described in detail below, is a characteristic part of the method of the present invention. Note that each process corresponding to the case where there is no vehicle detection is substantially the same as the conventional process.

First, a subsequent process (equivalent to the conventional process) when no vehicle is detected will be described. If there is no vehicle detection (S4, S6), the extraction (update) of the reference road surface level data is performed in the following process similar to the process disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 5-40818.

That is, the luminance data group (Dt) obtained corresponding to a predetermined number (N) of frames at predetermined frame (M) intervals is averaged to obtain primary reference road surface level data (Dr).
1), and the primary reference road surface level data (Dr1)
And the primary reference road surface level data (Dr1) obtained in the same manner as the previous time are determined to be identical. When it is detected that the number of determinations has been continuously obtained a predetermined number of times (Q times) or more, the primary reference road surface level data (Dr1) is formed as the secondary reference road surface level data (Dr2), and the same 2 When the next reference road surface level data (Dr2) is continuously obtained a predetermined number of times (R times), the secondary reference road surface level data (Dr2) is extracted as the reference road surface level data (Dr). It is updated and stored for use in the process for sensing (process: S10).

Thus, when the vehicle stops on the measurement line due to traffic congestion or other reasons, or when another vehicle frequently enters the screen, for example, the luminance based on the road surface intermittently appears on the measurement line. Even in a bad environment that can only appear, it is possible to skillfully select and capture the detection line of only the road surface that appears intermittently, accurately extract and update the reference road surface level data, and correctly distinguish between roads and vehicles Can be performed.

The updating process described above is not a requirement of the present invention, and the updating itself may be performed by another method. In short, it is only necessary to extract a luminance data group corresponding to a case where no vehicle is present on the detection line and update the reference road surface level data (Dr) at a predetermined frequency with data reflecting the extracted luminance data group. As a result, the reference road surface level data is suitably corrected with the lapse of time according to the situation such as the illuminance of the road. However, any such update process,
If the operation is performed uniformly regardless of the vehicle detection result, only the semi-presence operation can be expected as pointed out above.

Next, a description will be given of a related process and a process relating to handling of reference road surface level data when a vehicle is detected, which is a feature of the present invention. When a vehicle is detected by comparing the two pattern change rates, a sensing output is transmitted. Then, the vehicle sensing state is equal to or longer than a predetermined period of time (for example, 5 seconds. If the measurement line is being scanned, it is also sensed in the next scan, that is, it is considered that the measurement line has been continued during the scan cycle). If continued, the mode is shifted to the vehicle stop mode (steps S7 and S8). Actually, the result may be stored in a memory or the like so that it can be referred to.

When the vehicle shifts to the vehicle stop mode, the above-mentioned reference road surface level is not updated. In the example,
It is determined whether or not the vehicle is in the vehicle stop mode (S9). If the vehicle is in the vehicle stop mode, it is further determined whether or not the sensing signal is transmitted (S1).
1) If no sensing signal is sent (during sensing: sending a sensing signal), the reference road surface level data is not updated,
Data acquisition (S1) of the next new measurement line is performed. If the measurement line is not scanned, the processing is repeated from data acquisition (S1).

In this manner, the reference road surface level data (Dr) before vehicle detection is referred to until the vehicle leaves.
As a result, it is possible to continuously transmit the sensing signal in response to the case where there is a stopped vehicle.

If the stopped vehicle leaves after shifting to the vehicle stop mode, this fact is detected in step (S4), and the sensing output is canceled correspondingly (step: S).
6). Thereafter, through the steps (S9) and (S11), the vehicle stop mode is released (S12), and only in this case, the reference road surface level data corresponding to the present condition is immediately prepared in preparation for the next vehicle detection. (S13). For this purpose, the updating process of the reference road surface level is described as Q =
0, R = 1. That is, the same luminance pattern is P
This pattern is immediately adopted and held as a reference road luminance pattern when it is obtained consecutively (for P frames), so as to correspond to the subsequent vehicle sensing operation. This is shown in FIG.

Alternatively, the reference road surface level data may be updated with the latest single brightness level group. The point is that when there are no vehicles stopped for a long time, the transmission of the sensing signal is stopped and the vehicle stop mode is released, and by using a different algorithm or by using a parameter different from the normal from the road surface luminance data, An update process to the reference road surface level data (Dr) according to the current situation may be performed as soon as possible.

The method of detecting a stopped vehicle has been described above. However, at night, the brightness level of the road surface is extremely low, and the above-described detection method cannot be effective in a natural state. Therefore, consideration must be given to illuminating the monitoring target area with an appropriate lighting device so that necessary luminance can be obtained. Many proposals have already been made for nighttime illumination technology and the like.

As described above, according to the image-based vehicle sensing method of the present invention, the luminance change rate pattern on the measurement line is obtained and compared with the previously learned road luminance change rate pattern (pattern correlation comparison). If the road and the vehicle are distinguished from each other and the sensing continues for a certain period of time (T1) or more, the vehicle shifts to the stop mode to perform the vehicle sensing, and the presence of the stopped vehicle is continuously determined. Of course, this determination (sensing) is continuously performed until the vehicle leaves, and a so-called presence detection operation is achieved.

The method of the present invention has been described with reference to the embodiments. As can be seen from the above description, according to the present invention, a stopped vehicle on an arbitrary measurement line can be detected on the measurement line. However, the measurement line to be processed is not fixed, and a plurality of measurement lines are provided.
It is also possible to scan in order to detect a stopped vehicle on all measurement lines, and to detect a wide range.

In the method of the present invention, in order to detect the stop of the vehicle or the parking, the respective measurement lines must
Since it is sufficient to determine the presence or absence of a vehicle by acquiring data at a cycle of 3 to 4 seconds, the measurement lines are scanned in order to collect image A / D data of each measurement line, and to detect vehicle stoppage and parking. Necessary image data can be obtained.
In this way, it is possible to determine whether the vehicle has stopped in a wide area between the start measurement line and the stop measurement line. This can be used effectively to detect vehicles parked illegally on the roadside zone. Such processing is also possible with the illustrated device. Hereinafter, this application to which the present invention is applied will be described.

Although not shown in the flow charts of FIGS. 1 and 2, the exemplary apparatus moves the measurement line between the start measurement line and the stop measurement line periodically (for example, 3 to 4 seconds) downward in a frame unit. By sequentially scanning, image A / D data of a measurement line different for each frame is acquired. FIG.
Shows an example of the arrangement of measurement lines corresponding to the case where cross-sectional traffic flow measurement up to four lanes is performed. For the start measurement line, any one of A, C, and D is changed to an appropriate one according to the traffic situation. Use switching. For example, measurement (vehicle detection) is performed from a start line A to a stop line B with a small traffic volume.

For this reason, in the example device, as shown in FIG. 10, at the end of the even field, the image A /
A measurement address (designating a measurement line) at which the D data is to be loaded is loaded into the measurement address memory 26 via the host CPU. Note that the measurement address memory 26
The measurement address data is stored in O (fart-in first-out). In the next frame, the image A / D data of the measurement line corresponding to the measurement address is fetched. In this way, in the exemplary apparatus, the measurement line is periodically scanned (for example, 3 to 4 seconds) downward in the frame between the start measurement line A and the stop measurement line B sequentially in frame units, and different measurement is performed for each frame. The image A / D data (luminance data group) of the line is fetched.

As described above, between the start line and the stop line, the determination of the stopped vehicle including the above-described vehicle stop mode is repeatedly performed for each scanning line for each target scanning line in the cycle (T2). For example, as shown in FIG. 11, vehicle stop data indicating the presence of a stopped vehicle in each lane is obtained between the start measurement line and the stop measurement line. Since the data as shown in FIG. 11 is obtained for each cycle (T2), if there is vehicle stop data at the same measurement position for more than the vehicle stop determination time (T1: second), the data between the start line and the stop line may be obtained. It is determined that there is a stopped vehicle.

As shown in FIG. 11, when there is a stopped vehicle on the roadside zone, it is possible to identify the stopped vehicle. If the vehicle is stopped for more than this fixed time, this is regarded as illegal parking. It is possible to make a determination and take an appropriate response process (for example, a warning to stop illegal activities, or a clerk going to a parking place and performing a necessary process). This leads to elimination of traffic congestion near intersections and prevention of danger.

It should be noted that the above-described method of the present invention functions effectively when sufficient illuminance is obtained in the daytime, and when a lighting device or an infrared camera is used even at night. Was. However, even when a general visible light TV camera is used and no illumination is used, it is possible to detect a stopped vehicle to some extent with the same device as described above. As a technique related to the present invention, an image-based vehicle sensing method that can be performed by the apparatus alone for use at night will be briefly described below.

In the night vehicle detection using the apparatus alone, the headlight pair is detected and the stopped vehicle can be determined by utilizing the fact that the brightness value of the headlight is much higher than the surroundings. This pattern is shown in FIG. start line~
Between the stop lines, image A / D data is obtained by scanning at a cycle (T2) as in the case described above (for daytime). As a result, the data shown in FIG. 13 is obtained in substantially the same manner as in FIG. 3 described above. Therefore, when the data corresponding to the headlight data that is lit at the same measurement position continuously for a fixed time (T1) or more is obtained. Is a vehicle stop. In this method, since the headlight is detected, if there is a lighting vehicle, the signal S / N is good and the detection accuracy for the lighting vehicle itself is good. Therefore, as shown in FIG. 13B, an appropriate single detection level may be set, and if there is luminance data exceeding this detection level, it may be determined that the vehicle is present.

It should be noted that the above-described image-based vehicle detection (for night use) is difficult to detect with respect to the detection of a parked vehicle on the roadside zone because the parked vehicle usually turns off the headlights. However, by setting the threshold level low (higher sensitivity) and detecting the lighting of the small light, it is possible to detect even the vehicle with the small light on even if the headlight is off It is possible. However, in order to completely detect night parking (absolute stop), it is considered that a parked vehicle detection device using, for example, infrared rays or the like other than the above-mentioned method is effective. The related art of the present application has been described above.

[0046]

As described in detail above, the method of the present invention is:
An image signal corresponding to the area including the road to be sensed is obtained from the imaging means, and the image signal corresponding to a plurality of measurement points located on the detection line assumed in the cross direction of the road on the screen in time respectively. Are sequentially obtained, a luminance data group corresponding to a case where no vehicle is present on the detection line is extracted, and an update process is performed at a predetermined frequency with the reference road surface level data (Dr). If the difference between the brightness data group (Dv) based on the brightness data group (Dt) sequentially obtained and the reference road surface level data (Dr) is larger than a predetermined value, a vehicle is present on the measurement line. An image-based vehicle sensing method for sending a sensing signal, in particular,
When the state of sensing the vehicle continues for a predetermined time (T1) or more, the vehicle shifts to a vehicle stop mode in which only the update processing of the reference road surface level data (Dr) is stopped, and a luminance data group sequentially obtained during the stop mode. When the difference between the luminance data group (Dv) based on (Dt) and the reference road surface level data (Dr) no longer exceeds a predetermined value, the transmission of the sensing signal is stopped, and the luminance data group is quickly changed. Since the above-mentioned vehicle stop mode is released by performing the updating process with the reference road surface level data (Dr), the sensing output is appropriately transmitted to the vehicle that has been stopped for a long time. As a result, the inconvenience of being unable to sense the presence of a vehicle (semi-presence operation) even when the vehicle actually exists is solved. In addition, even when the vehicle comes out of a long stop, the reference road surface level corresponding to the current situation is promptly prepared, so that there is no problem in detecting the next vehicle. If the measurement line is scanned and used as in the embodiment, a wide range of stopped vehicles can be monitored, and illegal parking in the roadside zone can be monitored.

[Brief description of the drawings]

FIG. 1 is a flowchart showing an embodiment of an image-based stopped vehicle detection method according to the present invention, in conjunction with FIG.

FIG. 2 is a series of flowcharts showing FIG. 1 showing one embodiment of the present invention.

FIG. 3 is an explanatory diagram showing an example of a measurement line and a road luminance pattern and a vehicle luminance pattern obtained along the measurement line according to the present invention.

FIG. 4 is an explanatory diagram relating to a road luminance change rate pattern according to the present invention.

FIG. 5 is an explanatory diagram showing a setting example of a measurement line according to the present invention.

FIG. 6 is an explanatory diagram of a process of obtaining reference road surface level data according to the present invention.

FIG. 7 is an explanatory diagram of a vehicle detection principle according to the present invention.

FIG. 8 is an explanatory diagram of a process of obtaining reference road surface level data immediately after leaving a stopped vehicle according to the present invention.

FIG. 9 is an explanatory diagram showing an example of setting a plurality of measurement lines in the embodiment.

FIG. 10 is a timing chart showing how image data is taken in from each measurement line in the embodiment.

FIG. 11 is a diagram illustrating a process of detecting a parked vehicle in a monitoring area in the embodiment.

FIG. 12 is an explanatory diagram showing an example of an image and a measurement line at nighttime according to the present invention.

FIG. 13 is an explanatory diagram of an example of a road luminance pattern and a vehicle luminance pattern obtained along a measurement line at night and a vehicle detection principle according to the present invention.

FIG. 14 is a diagram showing an installation example of the image-type vehicle sensing device according to the present invention.

FIG. 15 is a block diagram showing an example of the configuration of an image-type vehicle sensing device suitable for use in carrying out the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Imaging means, 20 ... Luminance data conversion part, 30 ... Operation processing part, 40 ... Control part.

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G01V 8/10 G06T 1/00 G06T 7/20

Claims (1)

(57) [Claims] 1. An image signal corresponding to an area including a road to be detected is obtained from an image pickup means, and a plurality of measurement points located on a detection line assumed in a cross direction of the road on a screen are temporally respectively provided. The luminance data group (Dt) at the position in the corresponding video signal is sequentially obtained (S1, S2), and the luminance data group corresponding to the case where no vehicle is present on the detection line is extracted and the reference road surface level is determined at a predetermined frequency. Data (D
r) (S10), and compares the luminance data group (Dv) based on the sequentially obtained luminance data group (Dt) with the reference road surface level data (Dr) (S3, S4). When a difference is recognized, a sensing signal indicating that a vehicle is present on the measurement line is transmitted (S5). In the image-based vehicle sensing method, if the state of sensing the vehicle continues for a predetermined time (T1) or more, A shift is made to a vehicle stop mode in which only the updating process of the reference road surface level data (Dr) is stopped (S7, S8). When the difference between the road level data (Dr) and the predetermined level is no longer recognized, the transmission of the sensing signal is stopped (S4, S6).
The luminance data group is quickly converted to reference road surface level data (Dr)
(S13), and the vehicle stop mode is released (S12).