JP2019101510A - Visually impaired person detection program and visually impaired person entry monitoring device - Google Patents

Abstract

To surely detect a visually impaired person.SOLUTION: A program for detecting the visually impaired person causes a computer to execute: movement area detection processing of detecting a movement region of a moving body on the basis of a plurality of images; white region detection processing of detecting a white region within the movement region; linear portion extraction processing of extracting a linear portion from the white region; calculation processing of calculating a length and angle of the linear portion; and detection signal output processing of outputting a detection signal indicating that a visually impaired person has been detected, on the basis of the length and angle of the linear portion.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a blind person detection program and a blind person entrance surveillance device.

  Conventionally, a system for detecting a white cane user based on three-dimensional distance information obtained from a stereo camera is known (for example, Patent Document 1).

Patent No. 4691701

  However, the system described in Patent Document 1 is a system using a stereo camera and is a complex and expensive system.

(1) The visually impaired person detection program according to the first aspect of the present invention includes moving area detection processing for detecting a moving area of a moving object based on a plurality of images, and white for detecting a white area in the moving area. Region detection processing, straight portion extraction processing for extracting a straight portion from the white region, calculation processing for calculating the length and angle of the straight portion, and the visually impaired person based on the length and angle of the straight portion And causing the computer to execute detection signal output processing that outputs a detection signal representing that the sensor has been detected.
(2) The visually impaired person detection program according to the second aspect of the present invention is the visually impaired person detection program according to the first aspect, wherein in the detection signal output processing, the length of the linear portion is longer than a predetermined length And when it determines with the angle of the said linear part being a predetermined angle range, the said detection signal is output.
(3) The visually impaired person detection program according to the third aspect of the present invention is the visually impaired person detection program according to the second aspect, wherein the visual object to be detected in the detection target area for detecting the visually impaired person is It further has flow line direction setting processing for setting a flow line direction of a person with a disability, and the lower limit value and the upper limit value of the predetermined angle range are defined based on the set flow line direction.
(4) The visually impaired person detection program according to the fourth aspect of the present invention is the visually impaired person detection program according to the third aspect, wherein the detection signal output process comprises an image in a first image and a second image Assuming that a lateral direction is a reference direction, an angle formed by the flow line direction clockwise from the reference direction is θref, and an angle formed by the linear portions clockwise from the reference direction is θdet, then θdet is (θref / 2 The detection signal is output when the condition of ≦ θ det ≦ (90 + θ ref / 2) is satisfied.
(5) The visually impaired person detection program according to the fifth aspect of the present invention is the visually impaired person detection program according to the third aspect, wherein a detection target area for inputting and setting a detection target area for detecting a visually impaired person The system further includes a setting process and a length setting process for inputting and setting the predetermined length, and when a plurality of detection target areas are set, the flow line direction and the predetermined length are set for each detection target area. It is set.
(6) A device for monitoring attendance of visually impaired according to a sixth aspect of the present invention is a device for detecting a visually impaired implementing the detection program for visually impaired according to any of the first to fifth aspects, and an entrance gate of a facility. And an informing device for receiving the detection signal to notify the entrance of a visually impaired person.
(7) The visually impaired person attendance monitoring apparatus according to the seventh aspect of the present invention is the sixth vision impaired person entrance surveillance apparatus, wherein the imaging device is installed to image a plurality of the entrance gates. ,
A detection target area, a flow line direction, a predetermined length, and a predetermined angle range are set for each of the entrance gates in the frame image captured by the imaging device.
(8) The apparatus for monitoring visually impaired entry according to the eighth aspect of the present invention is the entrance blind monitoring apparatus for a visually impaired according to the sixth aspect, wherein the imaging apparatus comprises a plurality of the entrance gates in the object field. Only one is installed to include.
(9) In the surveillance apparatus for persons with visual impairment according to the ninth aspect of the present invention, in the visually impaired entrance monitoring apparatus according to any one of the sixth to eighth aspects, the facility is a station and the entrance gate is It is a ticket gate, and the imaging device is installed at a position where it captures an image of an visitor who passes through the ticket gate and enters the yard of the station, and the notification device further detects a voice for a visually impaired person Present information prompting passengers.

  According to the present invention, since the white cane can be detected by a simplified algorithm, an inexpensive blind person detection program and blind person entrance surveillance can be provided.

BRIEF DESCRIPTION OF THE DRAWINGS It is a block diagram explaining the visual impaired person entrance monitoring apparatus of one Embodiment. It is a figure which shows an example of a 1st frame image. It is a figure which shows an example of a 2nd frame image. It is a figure which shows the binarized image of the detected white stick. It is a figure explaining the angle range in which a white stick is included, when a person with visual impairment walks in the direction from the left to the right of the imaging screen. It is a figure explaining the angle range in which a white cane is included, when a person with visual impairment walks in the direction from the right to the left of the imaging screen. It is a figure explaining the angle range in which a white cane is included, when a person with visual impairment walks in a direction from the back side to the front side of the imaging screen. It is a flowchart explaining the process performed by a visual impaired person detection program. It is a figure which shows the mode of provision of a vision impaired person detection program.

  A visual impaired person's entrance surveillance device (hereinafter, also simply described as a surveillance device) according to an embodiment will be described with reference to the drawings. The visually impaired person entrance surveillance device according to the present embodiment has, for example, input image data of an image obtained by imaging the surveillance area with a surveillance camera installed near the automatic ticket gate device of the station, and holding a white wand. It is used to detect that a visually impaired person (hereinafter, also simply referred to as a visually impaired person) passes through the automatic ticket gate.

  FIG. 1 is a view showing an example of the configuration of a blind person entrance surveillance device provided with a blind person detection device according to the present embodiment. The visually impaired person entrance monitoring device 100 comprises an imaging device (hereinafter referred to as a surveillance camera) 120 for imaging a surveillance area, a visually impaired person detection device 130 for detecting a visually impaired person walking with a white cane, And a notification device 150 for notifying that a has been detected.

  The visually impaired person detection device 130 receives the moving image of the monitoring area (hereinafter also referred to as a detection area) obtained by the monitoring camera 120 via the interface 125, and the visually impaired person walks with the white cane based on this image. Specifically, it detects a white cane that moves. The visually impaired person detection device 130 includes an operation input unit 131 such as a keyboard and a touch panel, an image processing unit 132 that performs various processes on an image of a monitoring area obtained by the monitoring camera 120, and a storage unit (memory 133, an operation unit 134 that performs white cane determination calculation and the like based on image data after image processing, a condition setting unit 135 that stores setting parameters such as determination conditions used for the white cane operation, and the like, visually impaired person detection The signal processing unit 136 generates a signal.

  The surveillance camera 120 is a camera capable of shooting moving pictures at a frame rate of 60 frames / second, for example. The visually impaired person detection device 130 is a processing device equipped with a processor, and is a computer installed in a facility management center. The operation input unit 131 inputs a determination condition and the like necessary for the arithmetic processing. When the visually impaired person detection device 130 detects a visually impaired person, the signal generation unit 136 generates a signal (hereinafter also referred to as a detection signal). The signal is output to the notification device 150 via the interface (I / F) 155. The notification device 150 is a display device or a speaker of the facility management center, and notifies the entrance of a visually impaired person as text information and voice information from these devices.

  The image processing unit 132 performs processing such as generation of image data, compression, and binarization. The storage unit 133 stores image data and other various data. The calculation unit 134 performs calculation of luminance information, a difference between a plurality of image data, and calculation of the length and angle of a white area recognized as a straight line part. The operation input unit 131 inputs the size and position of the detection target area, the sensitivity of white area detection, and the like, and the condition setting unit 136 stores these setting conditions (data). Only one detection target area may be set within the range of the frame image input from the surveillance camera 120, or a plurality of detection target areas may be set.

  A series of processes of the visually impaired person detection by the visually impaired person detection device 130 will be described. These processes are performed by the disabled person detection device 130 constituting the monitoring device 100 executing the visual disabled person detection program stored in the storage unit 133 in advance.

(Image acquisition / compression operation)
The moving image captured by the monitoring camera 120 is input to the blind person detection device 130. The moving image signal received by the visually impaired person detection device 130 is stored in the original image storage area of the storage unit 133. The visually impaired person detection apparatus 130 according to the embodiment selects a plurality of frame images (still images) arranged in time series in real time from the moving image signals stored in the original image storage area, and selects these plurality of frame images. Detect visually impaired persons based on it.
In the following description, two old and new frame images are read out from the storage unit 133 as the first and second frame images, and the visually impaired person detection process is executed.

  FIG. 2 shows a first frame image. The first frame image 200 shows, as an example, a state in which the visually impaired person 202 who walks using the white stick 201 is included. The image data of the first frame image 200 is compressed by the image processing unit 133 and stored in the buffer area of the storage unit 133. The buffer area of the storage unit 133 can store the capacity of compressed image data for two frames in a queue (first-in first-out configuration).

  A detection direction (flow line direction) D is set in the image 200 based on the setting data of the condition setting unit 135 based on the direction in which the movement of the visually impaired person is to be detected. In the image 200, a detection target area 203, which is an area for detecting a visually impaired person, can be set in advance by setting data of the condition setting unit 135. Although the detection target area 203 is the left half area in the image in FIG. 2, the entire image area may be set as the detection target area 203. The process of setting the detection target area can be omitted. The detection direction D is set based on the set position of the detection target area 203 and the range thereof.

When the detection target area 203 is set in the image 200, the image processing unit 132 may generate compressed image data by compressing only the image data of the area corresponding to the detection target area 203. The number of pixels of the image to be compressed is, for example, 480 × 270 pixels when the aspect ratio of the image to be compressed is 16: 9, and 320 × 240 pixels when the aspect ratio is 4: 3. Compressed (transcoded).
The subsequent processing may be performed without compressing the image of the detection target area 203.

A second frame image is selected following the first frame image.
The processing of the second frame image will be described. Here, the second frame may be a frame immediately after the first frame, or may be a frame after several frames. In the present embodiment, the moving object region is extracted by performing differential processing on two new and old image data, so the second frame image is a new frame image for performing differential operation on the previous image data.

  FIG. 3 shows a second frame image 300. The image 300 shows, as an example, a situation in which a visually impaired person walking using a white cane is approaching to the front according to the lapse of time from the time when the first frame image 200 is captured. Also in the image 300, a detection direction D common to the detection direction set in the detection target area 203 is set in the detection target area 303. The detection direction D is set as shown in FIG. 7 described later. Similar to the first frame image, the image data of the second frame image 300 is also data compressed by the image processing unit 133, and is stored in the buffer area of the storage unit 133.

  The first and second compressed image data generated from the moving image captured by the monitoring camera 120 as described above is stored in the storage unit 133. Hereinafter, in the buffer area of the storage unit 133, first compressed image data obtained by compressing image data corresponding to the area of the detection target area 203 set in the image 200 and the area of the detection target area 303 set in the image 300 The white stick detection process will be described by way of an example in which the second compressed image data obtained by compressing the corresponding image data is stored.

(Brightness calculation)
The calculation unit 134 calculates luminance data for each pixel for each of the first compressed image data and the second compressed image data stored in the buffer area of the storage unit 133. Luminance data can be obtained using known discrete cosine transforms. Thereby, the first luminance data is calculated from the first compressed image data, and the second luminance data is calculated from the second compressed image data.

(Difference operation, movement area operation)
The calculation unit 134 calculates the difference between the first luminance data and the second luminance data for each of the corresponding pixels of the first compressed image data and the second compressed image data. The operation unit 134 extracts pixels where the calculated difference in luminance data is larger than a predetermined value, and the number of pixels in an area (hereinafter, difference area) formed by those images continuously is larger than the predetermined number of pixels. Determine if it is. That is, it is determined whether the area where the change in luminance is larger than the predetermined area is larger than the predetermined area. If it is determined that the number of pixels in the difference area in which the difference in luminance data is larger than a predetermined value is larger than the predetermined number of pixels, the computing unit 135 determines that the subject has moved. Specifically, the determination as to whether the subject has moved is performed, for example, by the procedure described below.

The luminance data (luminance value) of each pixel of the first compressed image data and the second compressed image data is represented by quantization numerical values of 256 gradations of 0 to 255. The condition setting unit 135 sets the sensitivity for determining that the luminance data has changed for the corresponding pixels of the first compressed image data and the second compressed image data, that is, the quantization numerical value corresponding to the difference of the luminance data. It is set as (for example, 50). The computing unit 135 determines that a region in which the difference pixels larger than the set value continue more than a predetermined number is the movement region.
Note that the condition setting unit 135 is an environment (for example, brightness, color temperature, etc.) in which the monitoring device 100 detects a person with visual impairment (sensitivity, numerical value of difference, etc.) for determining whether the luminance data has changed. It can be set from 1 to 255 based on.

(White extraction operation)
A white area extraction process described below is performed on the moving area determined by the calculation unit 134. The calculation unit 134 extracts a white area corresponding to white from the image of the moving area determined as the area where the luminance change in the second compressed image data is larger than a predetermined area. The specific procedure for extracting a white area is described next.
The calculation unit 134 selects pixels in a predetermined range of quantization values of luminance data of R (red), G (green), and B (blue) for each pixel of the difference area in the second compressed image data. , These are determined as white pixels. For example, the operation unit 134 determines that the quantization value of the luminance data of each of R, G, and B of the pixels in the difference area determined to have a luminance change is a white reference threshold 200 or more based on 24-bit RGB and 24 If the allowable range of each quantization value is within 0 on the basis of bit RGB, it is determined that the pixel is a white pixel.

  The condition setting unit 135 can appropriately set the condition of the luminance data for determining a white pixel based on the condition of the use environment (for example, brightness, color temperature, and the like) of the monitoring apparatus 100.

(Binarization operation)
The image processing unit 132 performs binarization processing on the pixels determined to be white and the other pixels in the second compressed image data to generate a binarized image. The generated binarized image is shown in FIG. Furthermore, the calculation unit 134 obtains a straight portion 301 which is a portion extending linearly in the white region of the binarized image, and calculates the length and the inclination (angle) thereof. The operation of obtaining a straight portion and obtaining its length and inclination can be performed using a known Hough transform.

(White wand detection operation)
The calculation unit 134 determines whether the length of the acquired linear portion is longer than a predetermined length. In addition, the calculation unit 134 determines whether the inclination of the linear portion is within a predetermined angle range. When the length of the linear portion is longer than the predetermined length and the inclination thereof is within the predetermined angle range, the computing unit 134 corresponds to the white wand used by the visually impaired person in the binarized image. It is determined that the That is, when the length and the angle of the straight line portion of the white area in the binarized image satisfy the predetermined condition, the computing unit 134 determines that the white area in the binarized image is a white wand used by a visually impaired person. It determines that it corresponds.

  Here, the predetermined length is a length serving as a reference for determining whether the length of the linear portion of the white region in the binarized image matches the length corresponding to the white cane used by the visually impaired. It is. The predetermined length can be set variously by the condition setting unit 135 based on the environment in which a visually impaired person is detected, information on new product development of the white stick, and the like. The determination of the length and the inclination of the straight line portion may be performed first, or may be performed simultaneously.

As described above, the predetermined length is set based on the length of a commercially available white cane, but is appropriately set according to the mounting angle of the monitoring camera 120, the optical performance, and the like. When optical performances such as the length of the white cane, the direction of the optical axis of the surveillance camera with respect to the surveillance area, the magnification of the surveillance camera, and the focal length are input, usability is improved by setting the predetermined length automatically.
The predetermined length is defined as the number of pixels in the image.

  If the calculation unit 134 determines that the white area in the binarized image corresponds to the white wand that the visually impaired has, the calculation unit 134 counts the number of times of detection as the white wand is determined (count). I do. As a result of repeatedly executing the above-described series of processing, when the number of white canes detected by the operation unit 134 reaches a predetermined number, the signal generation unit 136 generates a visually impaired person detection signal, and the notification device 150 is generated via the interface 155. Output.

  The predetermined number can be set to 5, for example. That is, when the white cane is detected in the image for five frames, the visual impaired person detection device 130 outputs a visual impaired person detection signal. The condition setting unit 135 can variously set the above-mentioned predetermined number of times based on the environment in which the monitoring apparatus 100 is used. The above process is repeatedly performed until the user inputs an operation end to the operation input unit 131.

  In the white cane determination processing by the operation unit 134, when the white cane is not detected from the image data to be processed, the first compressed image data is discarded from the buffer of the storage unit 133, and the first compressed image data is processed in the buffer. The second compressed image data is transferred as the first compressed image data to the area where the second image data is stored. In the buffer, compressed image data corresponding to the third frame image is newly stored as new second compressed image data in the area where the second compressed image data is stored.

(Slope judgment of straight line part)
An example of a specific procedure for the calculation unit 134 to determine whether the inclination of the linear portion is within the predetermined angle range will be described.
The determination as to whether or not the inclination of the linear portion is within the predetermined angle range is performed by the calculation unit 134 based on the angle range calculated based on the detection direction D set in advance. With the direction parallel to the upper and lower lines in the second image, that is, the horizontal direction of the image as a reference direction, the angle formed by the detection direction D clockwise from this reference direction is θref, and the inclination of the linear portion clockwise from this reference direction When θdet is an angle formed, if θdet satisfies Expression (1), it is determined that the white area in the binarized image corresponds to the white cane used by the visually impaired.
(Θref / 2) ≦ θdet ≦ (90 + θref / 2) (1)

(Detection direction and predetermined angle range)
Specifically, the determination processing of the above-mentioned white cane will be described with reference to FIGS. In addition, although this determination is performed based on a binarized image and FIGS. 5-7 originally show a binarized image, in order to make an understanding easy, a visually impaired person is also displayed in FIGS. 5-7. ing.

The case where the length of the straight portion is longer than a predetermined length will be described.
FIG. 5 shows a case where the detection direction D is set in the direction from the left to the right of the screen in the image 200 shown in FIG. In this case, θref, which is an angle formed by the detection direction D clockwise from the reference direction, is θref = 0 °. Therefore, θ det is the angle formed by the slope of the linear portion in the clockwise direction from the reference direction, and 0 ° ≦ θ det ≦ 90 ° according to equation (1).
If the above condition is satisfied, that is, if the slope of the linear portion is within an angle range of 0 ° to 90 °, it is determined that the white area in the binarized image is a white stick.

FIG. 6 shows the case where the detection direction D is set in the direction from the right to the left of the screen in the image 200 shown in FIG. In this case, θref, which is an angle formed by the detection direction D clockwise from the reference direction, is θref = 180 °. Therefore, θ det is the angle formed by the inclination of the linear portion in the clockwise direction from the reference direction, 90 ° ≦ θ det ≦ 180 ° by equation (1)
If the above condition is satisfied, that is, if the slope of the linear portion is within an angle range of 90 ° or more and 180 ° or less, it is determined that the white area in the binarized image is white cane.

FIG. 7 shows a case where the detection direction D is set in a direction from the back to the front of the screen in the image 200 shown in FIG. In this case, θref, which is an angle formed by the detection direction D clockwise from the reference direction, is θref = 90 °. Therefore, θ det is the angle formed by the slope of the linear portion in the clockwise direction from the reference direction, according to equation (1): 45 ° ≦ θ det ≦ 135 °
If the above condition is satisfied, that is, if the slope of the linear portion is in the angle range of 45 ° to 135 °, it is determined that the white area in the binarized image is white cane.

  5 and 6 show a case where a surveillance camera is installed to monitor the ticket gate from the side, and FIG. 7 shows a case where a surveillance camera is installed to monitor the ticket gate from the inside of the station yard.

It is also possible to set a plurality of detection target areas in one image. When a plurality of detection target areas are set in one image, the detection direction can be set to an appropriate direction in each detection target area. In addition, the predetermined length and the predetermined angle range for determining the length and the inclination of the linear portion can be set to values appropriate for each detection target area.
The detection target area is set to include an area until the visually impaired person passes through the entrance gate of the facility, for example, the ticket gate of the station.

(Initial setting process)
When using the visually impaired person detection device 130 according to the embodiment, an initialization process is required. In this initial setting process, a detection target area setting process, a flow line direction setting process, and various threshold value setting processes such as a predetermined length and a luminance difference are performed.

  The detection target area setting process is a process of inputting setting data of a detection target area for detecting movement of a visually impaired person from the operation input unit 131 and setting the setting data in the condition setting unit 135. For example, a rectangular frame of a size including the entrance and the exit of the ticket gate of the image of the surveillance camera 120 is defined and set on the screen.

  The flow line direction setting process inputs setting data of the flow line direction of the visually impaired person to be detected from the operation input unit 131 in the rectangular detection target area for detecting the visually impaired person, and sets the data in the condition setting unit 135. It is a process. Also in this case, a straight line connecting the entrance and the exit of the ticket gate of the image of the surveillance camera 120 is defined and set on the display screen.

  In the embodiment, the area of the pixel group where the difference between the luminance data of the two images to be compared is larger than the predetermined threshold is the movement area, but the threshold of the luminance difference is also input from the operation input unit 131 in the initial setting process. Are set in the condition setting unit 135. A quantization numerical value range at the time of detecting a white area from within the moving area is also input from the operation input unit 131 in the initial setting process and set in the condition setting unit 135. The quantization numerical range is defined by the above-described white reference threshold and white tolerance.

In the embodiment, a straight line portion is extracted from the white area in the movement area, and when the length of the straight line portion exceeds a predetermined threshold, it is recognized as a straight line portion corresponding to a white stick. The threshold value of this length is also input from the operation input unit 131 in the initial setting process and set in the condition setting unit 135. The threshold of this predetermined length is set, for example, in 500 steps of 1 to 500, for example, the number of pixels of the image.
In the embodiment, when the number of times the white wand is detected exceeds a predetermined threshold, that is, when the white wand is correctly detected a plurality of times, the signal generation unit 136 generates and outputs a visually impaired person detection signal. Do. This frequency threshold is also input from the operation input unit 131 in the initial setting process and set in the condition setting unit 135. The number threshold is set, for example, in ten steps of 1 to 10.

  The above threshold value and the allowable range can be input by, for example, displaying a slide bar switch for each setting data and moving this bar on the screen.

(Visually impaired person detection processing procedure)
The process for detecting visually impaired persons will be described with reference to the flowchart shown in FIG. The process shown in FIG. 8 is performed by executing a vision impaired person detection program stored in advance in the storage unit 133.

  In step S1, the first frame image captured by the camera is captured by the storage unit 133, and the process proceeds to step S2. In step S2, based on the setting information set in the condition setting unit 135, a detection target area in which movement of a visually impaired person is to be detected is set, and the process proceeds to step S3.

In step S3, the image processing unit 132 compresses the image data corresponding to the set detection target area to generate first compressed image data, and the process proceeds to step S4. In step S4, the first compressed image data is stored in the buffer, and the process proceeds to step S5. In step S5, the second frame image captured by the camera is captured by the storage unit 133, and the process proceeds to step S6. In step S6, based on the setting information set in the condition setting unit 135, a detection target area in which movement of a visually impaired person is to be detected is set, and the process proceeds to step S7. In step S7, the image processing unit 132 compresses the image data corresponding to the set detection target area to generate second compressed image data, and the process proceeds to step S8. In step S8, the second compressed image data is stored in the buffer, and the process proceeds to step S9.
Note that the procedure for storing the first and second compressed image data in the storage unit 133 is an example.

  In step S9, luminance data for each pixel is calculated by the calculation unit 134 for each of the first compressed image data and the second compressed image data, and the process proceeds to step S10. In step S10, the calculation unit 134 calculates the difference between the first luminance data and the second luminance data for each corresponding pixel of the first compressed image data and the second compressed image data, and the process proceeds to step S11. In step S11, it is determined by the operation unit 134 whether the number of pixels in the difference area, in which the calculated difference in luminance data is larger than the predetermined value, is greater than the predetermined number. If the number of pixels in the difference area where the calculated difference in luminance data is larger than the predetermined value is larger than the predetermined number of pixels, the determination in step S11 is affirmative and the process proceeds to step S12. If the number of pixels whose calculated difference in luminance data is larger than the predetermined value is not larger than the predetermined number of pixels, the determination in step S11 is negative and the process returns to step S1.

  In step S12, the arithmetic unit 134 extracts a white area corresponding to white from the movement area in the second compressed image data, and the process proceeds to step S13. In step S13, the image processing unit 132 generates a binarized image of the second compressed image data, and the process proceeds to step S14. In step S14, the calculation unit 134 acquires a straight line portion which is a portion extending linearly in the white area of the binarized image, and the process proceeds to step S15. In step S15, it is determined by the calculation unit 134 whether the length of the acquired linear portion is longer than a predetermined length. If it is determined that the length of the acquired linear portion is longer than the predetermined length, an affirmative determination is made in step S15, and the process proceeds to step S16. If it is determined that the length of the acquired linear portion is not longer than the predetermined length, a negative determination is made in step S15, and after performing buffer data exchange processing in step S22, the process returns to step S1. In step S16, it is determined by the calculation unit 134 whether the inclination of the acquired linear portion is within a predetermined angle range. If it is determined that the inclination of the acquired linear portion is within the predetermined angle range, an affirmative determination is made in step S16, and the process proceeds to step S17. If it is determined that the inclination of the acquired linear portion is not within the predetermined angle range, a negative determination is made in step S16, and after performing buffer data exchange processing in step S22, the process returns to step S1.

  In step S17, the calculating unit 134 determines that the white area in the binarized image corresponds to the white wand used by the visually impaired person, performs counting, and proceeds to step S18. In step S18, the calculation unit 134 determines whether the counting result determined to correspond to the white stick has reached a predetermined number. If it is determined that the counting result has reached the predetermined number, step S18 is affirmed and the process proceeds to step S19. If it is determined that the count result has not reached the predetermined number, step S18 is negative, and after the data exchange process of the buffer is performed in step S22, the process returns to step S1. In step S19, the signal generation unit 136 generates a disabled person detection signal, and the disabled person detection signal is output to the notification device 150 via the interface 155. Then, the process proceeds to step S20. In step S20, the data in the buffer area is replaced, and the process proceeds to step S21. In step S21, it is determined by the operation unit 134 whether or not an input of operation termination has been performed on the operation input unit 131. If it is determined that the operation end input has been performed, an affirmative determination is made in step S21, and the process ends. If it is determined that the operation end input has not been performed, an affirmative determination is made in step S21, and the process returns to step S1.

  The program for detecting the visually impaired person in FIG. 8 detects moving area of a moving object based on a plurality of images (steps S10 and S11) and white area detecting a white area in the moving area. Detection process (step S12), straight line part extraction process (step S14) for extracting a straight line part from the white region, calculation process (steps S15 and S16) for calculating the length and angle of the straight line part, straight line part The computer is caused to execute detection signal output processing (steps S17 to S19) for outputting a detection signal representing that a visually impaired person has been detected based on the length and the angle.

In the detection signal output process of the visually impaired person detection program, it is determined that the length of the linear portion is longer than the predetermined length (YES in step S15) and the angle of the linear portion is within the predetermined angle range (YES in step S16) And the detection signal is output when the satisfaction of the two conditions in which the length of the linear portion is longer than the predetermined length and the angle of the linear portion is within the predetermined angle range is detected more than the predetermined number of rotations (YES in step S18) (Step S19).
The detection signal may be output by determining that visual impaired person detection is performed when two conditions in which the length of the linear portion is longer than the predetermined length and the angle of the linear portion is within the predetermined angle range are satisfied once.

The visually impaired person detection program further includes flow line direction setting processing for setting the flow line direction of the visually impaired person who is the detection target in the detection target area for detecting the visually impaired person, and based on the set flow line direction. The lower limit value and the upper limit value of the angle range are defined.
The detection signal output process of the visually impaired person detection program uses the lateral direction of the image in the first image and the second image as a reference direction, and forms an angle θref in the clockwise direction from the reference direction and a straight line clockwise from the reference direction. Assuming that the angle formed by the portions is θdet, a detection signal is output when θdet satisfies (θref / 2) ≦ θdet ≦ (90 + θref / 2).

  The visually impaired person detection program further includes detection target area setting processing for inputting and setting a detection target area for detecting a visually impaired person, and length setting processing for inputting and setting a predetermined length, and the detection target area When a plurality of are set, the flow line direction and the predetermined length are set for each detection target area.

  The program for realizing detection of a visually impaired person of the present embodiment is recorded in a computer readable recording medium, and the above-mentioned program concerning detection of a visually impaired person recorded in the recording medium is a computer system It may be read and executed. Here, the “computer system” includes an OS (Operating System) and hardware of peripheral devices. The term "computer-readable recording medium" refers to a portable recording medium such as a flexible disk, a magneto-optical disk, an optical disk, a memory card, or a storage device such as a hard disk incorporated in a computer system. Furthermore, “computer-readable recording medium” dynamically holds a program for a short time, like a communication line in the case of transmitting a program via a network such as the Internet or a communication line such as a telephone line. It may include one that holds a program for a certain period of time, such as volatile memory in a computer system that becomes a server or a client in that case. Further, the above program may be for realizing a part of the above-described functions, and may be a combination of the above-described functions and a program already recorded in a computer system. .

  When applied to a personal computer (hereinafter referred to as a PC) or the like, the program relating to the control described above can be provided through a recording medium such as a CD-ROM or a data signal such as the Internet. FIG. 9 is a diagram showing the situation. The PC 950 receives the provision of the program via the CD-ROM 953. Also, the PC 950 has a connection function with the communication line 951. The computer 952 is a server computer that provides the above program, and stores the program in a recording medium such as a hard disk. The communication line 951 is a communication line such as the Internet, PC communication, or a dedicated communication line. The computer 952 reads the program using the hard disk, and transmits the program to the PC 950 via the communication line 951. That is, the program is carried by a carrier wave as a data signal and transmitted through the communication line 951. Thus, the program can be supplied as various forms of computer readable computer program products such as a recording medium and a carrier wave.

According to the first embodiment described above, the following effects can be obtained.
(1) The blind person detection program can detect the white cane with a simplified algorithm, so it can provide an inexpensive blind person detection program and blind person entrance surveillance.
(2) The visually impaired person detection program detects the moving area of the moving body based on the plurality of acquired images, detects the white area from the moving area, and the length of the straight line portion of the white area is greater than a predetermined length If it is determined that the tilt is long and the inclination is within the predetermined angle range, the computer is caused to output a detection signal indicating that a visually impaired person has been detected. As a result, it is possible to reliably detect the white cane of the visually impaired and output a signal for performing appropriate measures such as vocalization to the visually impaired.

(3) The visually impaired person detection program can set the detection direction (flow line direction) for detecting the movement of the visually impaired person in the visually impaired person detection target image. Based on the set detection direction, it is possible to more appropriately set a predetermined length and a predetermined angle range for determining that the straight portion of the white area is a white wand.

(4) The visually impaired person detection program can set a plurality of detection target areas in the visually impaired person detection target image, and when a plurality of detection target areas are set, each detection target area is set. A suitable detection direction and a predetermined length can be set. This makes it possible to more reliably detect visually impaired persons.

(5) The visually impaired person entrance monitoring device provided with the visually impaired person detection device implemented with the visually impaired person detection program of the first embodiment is such that the visitors passing through the entrance gate of the facility are included in the imaging screen Use the image captured in For this reason, it is possible to reliably detect a visually impaired person who is about to enter from the entrance gate of the facility, and to take appropriate measures such as vocalization. In particular, in facilities where there is a risk that a visually impaired person who has entered may accidentally approach a dangerous place, such as at a ticket gate of a railway station, appropriate measures such as vocalization can be performed.

(Modification 1)
In the above embodiment, the second compressed image is used to extract and binarize the white region, and it is determined from the length and inclination of the linear white portion whether it is a white cane or not, Counting was performed when it was determined (steps S12 to S17). However, the series of processes may be performed using the first compressed image, or may be performed using both the second compressed image and the first compressed image.

(Modification 2)
In the above embodiment, it is determined based on the equation (1) whether or not the angle of the linear portion recognized as the white cane is included in the predetermined angle range. However, the flow line direction and the predetermined angle range may be set in advance as the LUT (lookup table) in the condition setting unit 135, and the predetermined angle range may be determined by referring to the LUT from the flow line direction determined in the initial setting process. .

(Modification 3)
The visually impaired person detection device 130 is a computer in a management office room in a station yard connected to the monitoring camera 120 by a wired or wireless network, or a computer in a railway central management room of a railway operator. Alternatively, it may be a mobile terminal of a station clerk connected to the surveillance camera 120 via a wireless network. In the former case, the visually impaired person detection program is implemented in a computer, the operation input unit 131 is a computer keyboard, and the notification device 150 is a computer monitor or speaker. When the computer in the management office is notified of detection of entrance for the visually impaired, the person who manages and operates the computer or the station clerk notified by that person voices the person with visual impairment at the ticket gate. When detection of entrance to the visually impaired is notified by the computer in the railway central management room of the railway company, a notification signal of entrance to the visually impaired is transmitted to a computer or mobile terminal in the station yard previously registered in the computer.
Alternatively, the visually impaired person's entrance may be notified from a broadcasting facility or an electric light bulletin board installed near the ticket gate. It is further preferable to present information for prompting station staff or surrounding passengers to speak for the visually impaired along with the entrance information for the visually impaired.

In the latter case, the visually impaired person detection program is implemented on a mobile terminal owned by a station staff, the operation input unit 131 is a touch panel of the mobile terminal screen, and the notification device 150 is a mobile terminal screen or a built-in speaker. If a blind person detection program is installed on mobile terminals of multiple station staff, any station clerk who has received notification of the entrance of a visually impaired person may be able to call out visually impaired persons at the ticket gate. Become.
It is further preferable to present information on the station clerk's mobile terminal to prompt the station staff to speak to the visually impaired along with the entrance information of the visually impaired.

(Modification 4)
In the example mentioned above, it demonstrated as an apparatus which detects a visually impaired person who passes the ticket gate of a station. However, the visual impaired person detection device and the detection program according to the present invention can also be used for visual impaired person detection for entering and leaving various facilities such as airports, halls, theaters, accommodation facilities and commercial facilities other than stations.

  The present invention is not limited to the above embodiment as long as the features of the present invention are not impaired, and other embodiments considered within the scope of the technical idea of the present invention are also included in the scope of the present invention. .

100: Visually impaired person entrance surveillance device 120: Imaging device 130: Vision impaired person detection device 131: Operation input unit 132: Image processing unit 133: Storage unit 134: Arithmetic unit 135: Condition setting unit 136: Signal generation unit 150: Notification Device 200, 300 ... Image 201, 301 ... White cane

Claims (9)

It is a program that detects visually impaired people.
Moving area detection processing for detecting a moving area of a moving object based on a plurality of images;
White area detection processing for detecting a white area in the movement area;
Straight line portion extraction processing for extracting a straight line portion from the white area;
Calculation processing for calculating the length and angle of the straight line portion;
A detection program for a visually impaired person that causes a computer to execute detection signal output processing that outputs a detection signal representing that a visually impaired person has been detected based on the length and angle of the linear portion. The visually impaired person detection program according to claim 1, wherein
The visually impaired person detection program, wherein the detection signal output processing outputs the detection signal when it is determined that the length of the linear portion is longer than a predetermined length and the angle of the linear portion is within a predetermined angle range. The visually impaired person detection program according to claim 2, wherein
In the detection target area for detecting the visually impaired person, the method further includes flow line direction setting processing for setting a flow line direction of the visually impaired person as the detection target,
The visually impaired person detection program by which the lower limit and upper limit of the said predetermined angle range are prescribed | regulated based on the set said flow line direction. The visually impaired person detection program according to claim 3, wherein
In the detection signal output processing, an angle formed by the flow direction in the clockwise direction from the reference direction is θref, with the image lateral direction in the first image and the second image as a reference direction, and the angle in the clockwise direction from the reference direction is θref. The visually impaired person detection program which outputs the said detection signal, when (theta) det satisfies Formula (1) when the angle which a linear part makes is set to (theta) det.
(Θref / 2) ≦ θdet ≦ (90 + θref / 2) (1) The visually impaired person detection program according to claim 3, wherein
Detection target area setting processing of inputting and setting a detection target area for detecting a visually impaired person;
And further including a length setting process of inputting and setting the predetermined length,
The visually impaired person detection program, wherein when a plurality of detection target areas are set, the flow line direction and the predetermined length are set for each detection target area. A visually impaired person detection device for implementing the visually impaired person detection program according to any one of claims 1 to 5;
An imaging device installed so as to capture an image of an visitor passing through an entrance gate of the facility;
And a notification device that receives the detection signal and reports the entrance of a visually impaired person. It is a visual impaired person entrance surveillance device according to claim 6,
The imaging device is installed to image a plurality of the entrance gates,
The visually impaired person's entrance monitoring device, wherein a detection target area, a flow line direction, a predetermined length, and a predetermined angle range are set for each of the entrance gates in a frame image captured by the imaging device. It is a visual impaired person entrance surveillance device according to claim 6,
The device for monitoring the entrance of a visually impaired person, wherein only one imaging device is installed to include a plurality of the entrance gates in the field. In the visually impaired person entrance surveillance device according to any one of claims 6 to 8,
The facility is a station, and the entrance gate is a ticket gate,
The imaging device is installed at a position where an image of an entrance person who passes through the ticket gate and enters the yard of the station is imaged.
The notification apparatus further presents information for prompting a station staff or a passenger to make a voice call to a detected visually impaired person.

Images