JPH0574880B2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention measures the number of moving objects passing through a monitoring surface, for example, the number of people entering and exiting a building, etc. using a storage type photodetector such as a CCD sensor. The present invention relates to a moving body amount measuring device that displays a moving body amount.

(Prior art) Conventionally, devices that automatically measure the number of people entering and exiting a building, etc. have been, for example, a photoelectric switch device that uses a light beam, a device that uses video from a TV camera, or a switch built into a floor mat. Devices using so-called pine sensors are known.

(Problems to be solved by the invention) However, in the case of the light beam method, it is not possible to distinguish when multiple people pass side by side, and in the case of the television camera method, it is impossible to distinguish when multiple people pass by from the camera image. The technology for pattern recognition of human images is complicated and the equipment is expensive, and the matt sensor has problems with durability because detection is based on mechanical switch operation.

(Means for Solving the Problems) The present invention has been made in view of the above-mentioned conventional problems, and it is possible to accurately determine the number of passing moving objects, for example, the number of people, and the device configuration is simple and cost-effective. It is an object of the present invention to provide a moving object amount measuring device which is inexpensive, and is capable of obtaining sufficient luminance changes as the moving object passes, and is capable of high-speed processing.

In order to achieve this object, the present invention includes a plurality of light-receiving pixels that monitor changes in the brightness of a monitoring surface that monitors a moving object passing over the monitoring surface from the top or bottom in the moving direction of the moving object. A fluorescent lamp is configured to detect by storage type photodetectors arranged in a straight line, determine the number of moving objects based on light reception information detected by the storage type photodetectors, and further form a monitoring line on the monitoring surface. A light source such as the above is buried, and the light from the light source that enters the storage type photodetector is blocked at a portion through which the moving body passes, thereby obtaining a sufficient change in brightness.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the present invention.

First, to explain the configuration, 1 and 2 are charge-coupled devices (hereinafter referred to as "CCD sensors") as storage type photodetectors, and as shown in FIG. 2, a plurality of light receiving pixels 3a to 3n are linearly arranged. With this structure, when light hits each light-receiving pixel, it is possible to obtain an accumulated charge proportional to the integral amount of incident light over a certain accumulation time (exposure time) determined by the readout timing.

Regarding the CCD sensors 1 and 2 having such a structure, in the present invention, all received light data is not used as data used for calculation processing of traffic volume based on received light data, and is shown in the shaded area in Fig. 2. For example, every fourth received light data A1, A5,...
An-1, B2, B6, . . . Bn are read out from the memory storing the received light data by specifying the jumping address and arithmetic processing is executed.

To explain one example more specifically, for example,
As the CCD sensors 1 and 2, 2048 light-receiving pixels arranged in a straight line are used, and the light-receiving data of every 16 pixels is read out and data processing is performed. Here, CCD
Assuming that the monitoring length on the monitoring surface for each light-receiving pixel connected to the sensor is 0.25 cm, reading the light-receiving data every 16 pixels means reading the light-receiving data every 4 cm from the memory on the monitoring surface. It means to perform data processing.

By data processing the light reception data read out at regular intervals among the light reception data obtained from the light reception pixels arranged in a straight line in this way, high-speed data processing is possible even when the number of pixels per line is large. becomes.

Note that the density of the light reception data used for data processing corresponding to the light reception pixels is determined within the range where the passing of a person can be detected, for example, when the moving object is a person, and is variable within the range where the person can be detected. I can do it.

The structure of the monitoring surface and the detection optical system of the present invention using these two CCD sensors are shown in FIG.

In FIG. 3, for example, if a person is passing by as a moving object to be detected, there are two light source installation grooves 102 in the floor surface (monitoring surface) 4 of the entrance of a building, etc.
a, 102b are formed, and this light source installation groove 102
A, 102b are provided with a plurality of fluorescent lamps 106 arranged vertically facing upward, and a protective cover 1 made of glass or the like is placed in the groove opening at the top of the fluorescent lamps 106.
08a and 108b are provided.

As shown in FIG. 4b with the protective cover removed, the light source installation grooves 102a and 102b incorporate a plurality of fluorescent lamps 106 arranged in the longitudinal direction as light sources, and the protective cover 10 provided in the groove opening
8a and 108b are equipped with transparent or milky-white transmission windows 110a and 110b that form monitoring lines A and B by means of shaded areas shown in FIG. 4a, and the lines defined by the transmission windows 110a and 110b forming a light source. Referring again to FIG. 3, the images of the two monitoring lines A and B formed by the line light source provided on the floor surface 4 are reflected from the upper surface by the reflecting mirror 5 and collected in the direction of movement in which the person passes. The image of the monitoring line A enters the optical lens 6, and is reflected from the condensing lens 6 by the half mirror 7 and sent to the CCD.
An image is formed on the sensor 1, while an image of the monitoring line B is formed on the CCD sensor 2 through a condenser lens 6 and a half mirror 7. With such an optical system, the monitoring lines A and B are connected to the CCD sensor 1,
For each light-receiving pixel in 2, a 0.25 cm image is reduced and formed on a line, for example.

On the other hand, in this embodiment, as an example, monitoring lines A and B consisting of line light sources shown in plan in FIG. 4 are installed with monitoring line A installed outside the building and monitoring line B installed inside the building. There is. Here, the distance D between the parallel monitoring lines A and B is determined depending on the size of the moving object, and for example, when the moving object is a person, D=about 10 cm. Further, the width W of the monitoring lines A and B is determined by the size of the entrance/exit.

The reason why the distance D between the monitoring lines A and B is set to about 10 cm when the moving object is a person is as follows.

First, the shoulder width H1 of the person to be measured is statistically about 40 cm, as shown by the two-dot chain line, and the variance thereof is 20 to 60 cm. Also, the thickness of the chest
H2 is distributed in the range of H2 = 10 to 24 cm. Therefore, in order to determine the amount of traffic by detecting the change in brightness when passing through the two monitoring lines A and B, it is necessary to estimate a measurable probability close to 100% based on the chest thickness H2. Calculating the resulting line spacing D = 9.64cm
Therefore, in practice, it is sufficient to set D = around 10 cm.

Next, the accumulation time of the CCD sensors 1 and 2 shown in FIG.
Since the speed is 2.19 m/s, the line passing time is 44 ms when the line spacing D is 9.64 cm. Therefore, if data is sampled twice during this passing time, it will be 22ms, and if this amount of accumulation time can be secured, sufficient photoelectric charge will be accumulated due to changes in the brightness of the monitoring line even when a person passes by. It is possible to detect people passing by using a CCD sensor as a storage type photodetector.

Furthermore, since monitoring lines A and B are equipped with light sources such as fluorescent lights, the line images incident on the CCD sensor are at maximum brightness under normal monitoring conditions with no people passing by, and the line images entering the CCD sensor are at maximum brightness when a person passes by. The brightness decreases significantly, and the passage of a person is determined from the change in received light data accompanying this decrease in brightness.

Referring to Figure 1 again, CCD sensors 1 and 2
receives a transfer clock from the CCD drive circuit 8 and outputs a light reception signal at fixed accumulation time intervals, and the transfer clock outputs a light reception signal at fixed accumulation time intervals, and when the interval D between the monitoring lines is 10 cm, the transfer clock is output within the time period during which a person passes between the monitoring lines. Due to the conditions of sampling data at least twice and synchronizing with the light emission waveform of the light source, the data is accumulated in the light receiving pixel of the CCD sensor every 20 ms in the 50 Hz region and 16.7 ms in the 60 Hz region. The accumulated charge is then read out as a light reception signal.

The outputs of the CCD sensors 1 and 2 are sent to the A/D converter 9.
The signals a and 9b are converted into digital signals according to the received light level, and are provided to the multiplexer 10.

Here, since the CCD sensors 1 and 2 are equipped with, for example, 2048 light-receiving pixels, the readout time will be longer if the CCD sensors 1 and 2 are read out one after another.
Two systems of D converters 9a and 9b are provided, and the received light signals are read out in parallel to enable high-speed processing.

Following the multiplexer 10, there is a background processing circuit 11 for performing background processing on the received light data.
RAMs 12a and 12b are provided.

The light reception data of the A line and the B line obtained in a steady state with no people passing on the monitoring lines A and B are written in the RAM 12a, and this written light reception data is used for background processing by the background processing circuit 11. Used as background reference data for conversion. The background processing circuit 11 subtracts the received light data obtained in real time via the multiplexer 10 from the background reference data stored in the RAM 12a to create received light data that is not affected by the background. This background processing of the light reception data is performed in real time for each light reception data corresponding to the light reception pixel.

On the other hand, light reception data is written in real time to the other RAM 12b, which is not used for storing background reference data, via the multiplexer 10, and background reference data can be written by switching from RAM 12b to RAM 12a as necessary. We are making it possible to refresh.

The received light data subjected to background processing by the background processing circuit 11 is written to the buffer memory 14a or 14b via the gate circuit 13. The buffer memories 14a and 14b are connected via a gate circuit 15 to an arithmetic processing unit 16 that executes arithmetic processing for measuring the amount of traffic.

The reason why two buffer memories 14a and 14b are provided here is that when writing received light data into one buffer memory, for example, the buffer memory 14a, via the gate circuit 13, the other buffer memory 14b is written via the gate circuit 15. It is connected to the arithmetic processing unit 16, and the light reception data already stored in the buffer memory 14b is transferred to the arithmetic processing unit 16.
reads and processes the data.

In this way, by processing data in one buffer memory while writing data in the other buffer memory, timing alignment between the CCD sensor side and the arithmetic processing section 16 side becomes unnecessary, and the CCD sensor 1,
The transfer of the light reception data from 2 and the data processing by the arithmetic processing unit 16 can be made independent.

The arithmetic processing unit 16 is executed under program control by a CPU, for example, and reads the received light data stored in either one of the buffer memories 14a or 14b selected by the gate circuit 15, and performs arithmetic processing for measuring the amount of traffic. Do this. The arithmetic processing by the arithmetic processing unit 16 has the processing contents shown in the general flow in FIG.

First, data is read from the buffer memory 14a or 14b, as shown in FIG. 2, by reading received light data at preset regular intervals and performing data processing. Such data processing can be realized by specifying a jumping address for the buffer memory.

Based on the received light data from the buffer memory 14a or 14b input via the gate circuit 15, the arithmetic processing section 16 first reads the data on the A line as shown in block 18, and detects changes in the data. Subsequently, in block 19, a switch is made to reading data on the B line, and a change in the data is detected. Regarding the detection of data changes in blocks 18 and 19, the line position where the data change occurred is also detected at the same time.

Subsequently, in block 20, a pattern when a person passes through monitoring lines A and B shown in FIG. 6 is determined from changes in data on lines A and B.

In FIG. 6, time changes are shown for monitoring lines A and B when a person enters from the monitoring line A side located outside the entrance/exit, divided into times t1 to t3.
Here, pattern 1 is the state where the monitor line A at time t1 is reached, and the monitor lines A and B at time t2
Pattern 2 is a state in which the line is approaching both of the above, and pattern 3 is a state in which the line passes through monitoring line A and reaches monitoring line B. Of course, when leaving, a change from time t3 to t1 is obtained.

Therefore, the pattern determination in block 20 of FIG. 5 is to determine which pattern in FIG. 6 is based on the A and B line data.

In the next block 21, the direction of movement is determined from the pattern determined in block 20 and the previous pattern that has already been registered. For example, pattern 1
is determined, the previous pattern is pattern 1
If the pattern is not one of patterns 1 to 3, the player is admitted, and if pattern 3 is determined and the previous pattern is not one of patterns 1 to 3, the player is allowed to leave. Of course, when pattern 2 is determined, if the previous pattern was pattern 1, the player enters the room, and if pattern 3, the player exits.

Once the direction of movement has been determined, the number of entries and exits is counted in block 22. That is, if the user is entering, the counter is counted up, and if the user is leaving, the counter is counted down. When such measurement processing is completed, finally in block 23, among the passing positions of the person on the line and the patterns at the passing positions obtained by the data processing so far, the passing positions and patterns that are not used for determining the moving direction are determined. The pattern is registered and the process returns to block 18 again.
Then, the direction of movement is determined based on how the pattern of the unused passing positions has changed from the previous time.

Referring again to FIG. 1, the number of people entering and leaving the building at that time, calculated by the arithmetic processing unit 16, is given to the display 17, which displays the number of people currently remaining in the building or the number of people entering the building.

Note that the arithmetic processing unit 16 includes a multiplexer 10,
It also controls gate circuits 13 and 15.

7A, 7B, and 7C are flowcharts showing the control process of FIG. 1 in more detail.

First, FIG. 7A shows data processing of the A line.

When the apparatus is started, background reference data is set in block 50, and the next block 52
The data change number counter L, the position counter N, and the loop counter I for monitoring the number of data processed per line are each initialized to zero.

Next, in block 54, the first A line data set by loop counter I is input, and in judgment block 56, it is checked whether or not it is above a threshold value in order to determine that it is a person. If the received light data is less than the threshold value, the loop counter I is incremented in block 58, and the judgment block 60 checks whether the loop counter I has reached the final value I=128, and detects a data change that exceeds the threshold value. The processing of blocks 54 to 60 is repeated until detection is made.

When the received light data exceeds the threshold value, the process proceeds from the judgment block 56 to block 62, where the length of the Nth line from the end of the monitoring line specified by the position counter N at this time, that is, the number of data changes L, is incremented.
A decision block 64 checks whether the threshold value Lo has been reached. If the threshold value Lo has not been reached, a loop counter I is incremented in a block 66, and a decision block 68 checks whether the final value I=128 has been reached. After that, the next A line data is input in block 70, and if it is equal to or greater than the threshold value, the process returns to block 62 again from judgment block 72, and the data change number L is incremented.
Repeat this loop until =Lo is reached.

In addition, if received light data that does not exceed the threshold value is obtained before the data change number L reaches the threshold value Lo,
It is determined that the received light data is free, and the process proceeds from judgment block 72 to judgment block 74, where the free number counter for counting the number of free light received data is incremented, and when the count value of the free number counter exceeds a predetermined value, block 76 is executed. Proceed to the number of data changes L
is reset to zero and the process returns to block 54.

In the judgment block 64, the number of data changes L is the threshold value Lo.
If it is determined that the loop counter I has reached the final value, the process proceeds to block 78, where the person's position N is registered, and then the position counter N is incremented at block 80, and if the loop counter I has not reached the final value, the process returns to block 54. return.

By processing the A line data in this way, if there is a change in brightness due to a person passing on the monitoring line A,
Based on the value of the position counter N, the passing position on the line and the number of people passing can be obtained as registered data.

When the final value of loop counter I is determined in decision block 60 or 68, the process proceeds to block 82, where counters I, N, and L are reset to zero, and step 7B
The process proceeds to the B line data shown in the figure.

The processing of the B line data shown in FIG. 7B also
Since the processing is exactly the same as the processing of the A line data in FIG. 7A, each processing block is indicated by the same number with the addition numeral b.

When the data processing for the A line and the B line is completed in this way, the process proceeds to the pattern and movement direction determination process shown in FIG. 7C.

First, in the determination block 84, it is determined which of the patterns 1 to 3 shown in FIG. 6 is, and the process proceeds to any of the blocks 86a to 86c. Note that this pattern discrimination is performed in correspondence with the position counter N.

Next, in a determination block 88, the pattern is compared with the previously registered pattern to determine the direction of entry into block 90a or exit from block 90b. When determining the entrance to block 90a, the number of entrances and exits counter M is incremented by the number of people given by the position counter N. On the other hand, when leaving the block 90b, the number of entrances and exits counter M is given by the position counter N. Subtract by the number of people. Of course, since there may be a case where there is a mixture of persons entering and leaving the room, there may be cases where addition and subtraction of the entry/exit counter M are performed at the same time.

When the counting of the entry/exit number counter M is completed in this way, the blocks 86a to 86 are processed in block 92.
The pattern determined in step c is registered together with its position as the previous pattern, and the process is again performed at block 54 in FIG. 7A.
Return to

In the data processing in Figures 7A and 7B, when the number of data changes L reaches the threshold Lo, it is determined that the data change is due to a person passing by. is set in multiple stages, for example, in two stages, and the threshold value Lo for data processing of A and B lines is set to, for example, Lo=5 (20
cm), the pattern may be determined first, and after this pattern determination, the entry/exit counting operation may be performed after a data change reaching the threshold Lo=15 (60 cm on the line) is obtained.

Furthermore, although the above-mentioned embodiment takes as an example the passing of a person as a moving body, the present invention is not limited to this, and can also detect the amount of passing vehicles or luggage. The interval between the two monitoring lines A and B may be set to an appropriate value based on the body size and moving speed.

Furthermore, in the case of moving objects such as vehicles that move in only one direction, the number of moving objects can be counted by simply providing one line of storage type photodetectors.

Further, in this embodiment, the moving body is configured to be monitored from above in the direction of movement, but the line plateau may be installed on a ceiling surface or the like and monitored from below.

(Effects of the Invention) As described above, according to the present invention, a plurality of monitoring surfaces that monitor a moving object passing over the monitoring surface from above or below in the moving direction of the monitoring surface monitor brightness changes. The number of moving objects is determined based on the light reception information detected by the storage type photodetector, and a light source is embedded in the monitoring surface. , the light entering the storage type photodetector from the light source located on the monitoring surface is blocked by the passing part of the moving object to obtain a change in brightness, so when multiple people pass side by side at the same time, Even if exiting persons intersect, the number of people passing through and the direction of movement can be accurately determined.
Traffic volume can be detected with extremely high accuracy.

In addition, a light source is embedded in the monitoring surface so that the passing of a moving object blocks the light from entering the storage type photodetector, so that the change in brightness due to the passing of a moving object can be sufficiently large, and the change in brightness due to the passing of a moving object can be sufficiently large. Even if the accumulation time is short, a sufficient signal change can be obtained, and high-speed processing becomes possible. Furthermore, since the light source on the monitoring surface is blocked by the passing of a moving object to obtain a change in brightness, there is no effect of the color of clothing, for example, and the moving object can be seen as a backlight, further improving resolution. .

In addition, since a CCD sensor is used as a storage type photodetector, the detection optical system and data processing for determining traffic volume are simpler than with the conventional TV camera method, making it possible to achieve high detection accuracy at low cost. .

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is an explanatory diagram showing the structure of a CCD sensor used in the present invention, and Fig. 3 shows the detection optical system and line light source of the present invention. 4 is an explanatory diagram showing the structure of a line light source provided on the monitoring surface, FIG. 5 is a general flowchart showing an overview of data processing by the arithmetic processing unit in FIG. 1, and FIG. Explanatory diagram showing the state of a person passing through the monitoring line used in determining the movement direction of the invention, Nos. 7A, 7B, 7
Figure C is a flowchart showing the traffic volume measurement process of the present invention under program control. 1, 2...CCD sensor, 4...floor surface, 5...
Reflection mirror, 6... Condensing lens, 7... Half mirror, 8... CCD drive circuit, 9a, 9b...
A/D converter, 10...Multiplexer, 12
a, 12b...RAM, 11...background processing circuit,
13, 15...gate circuit, 14a, 14b...
Buffer memory, 16... Arithmetic processing unit, 17...
Indicator, 102a, 102b... Light source installation groove, 1
06... Fluorescent lamp, 108a, 108b... Protective cover, 110a, 110b... Transparent window.

Claims (1)

[Scope of Claims] 1. A storage type photodetector comprising a plurality of light receiving pixels arranged in a straight line for monitoring changes in brightness of a monitoring surface from the top or bottom in the direction of movement of a moving body passing over the monitoring surface. , in a moving object amount measuring device equipped with a determination means for determining the number of moving objects based on light reception information on the monitoring surface detected by the storage type photodetector, a light source is embedded in the monitoring surface; A moving body amount measuring device characterized in that light entering the storage type photodetector from a light source is blocked at a passage portion of the moving body to obtain a change in brightness.