JP6444753B2 - Obstacle detection device - Google Patents

Description

  The present invention relates to an obstacle detection device.

  When a train approaches a railroad crossing where the railroad track and the road intersect on the same plane, the electric railroad breakers installed at both ends of the railroad crossing close the railroad crossing. When there are obstacles such as cars or people in the crossing road, the warning and warning signals are issued to the train, and the train and ground traffic lights are operated to take measures to stop the train. There is a need for an obstacle detection device that detects the presence or absence of an obstacle in a railroad crossing.

  Conventionally, obstacle detection methods include (a) an optical detection method such as a combination of an infrared light emitting unit and a light receiving unit, a combination of a laser oscillation unit and a light receiving unit, and (b) a coil on the ground. Alternatively, there are an electromagnetic detection system in which a loop antenna is embedded and an automobile passing above is detected by electromagnetic coupling, and (c) an image detection system in which an obstacle is monitored by a photographing means such as a television camera or a CCD camera. Each of these three types of obstacle detection methods has the following problems (see Patent Document 1).

In the optical detection method, it is necessary to align the optical axis with the infrared light emitting section and its light receiving section or the laser oscillation section and its light receiving section facing each other with high accuracy, so installation and maintenance of the device is easy. Not. For example, if there is an earthquake, maintenance of optical axis alignment is necessary. Further, since only the position through which infrared rays or laser light passes can be detected, it is difficult to detect a small one with many blind spots. Therefore, it is necessary to install many optical detection devices in order to cover the obstacle detection area. Further, when a snow lump is formed in the obstacle detection area due to the snow rising during snowfall, the infrared or laser light is reflected by this, and it may be erroneously detected as if there is an obstacle. Further, particularly in a combination of an infrared light emitting portion and its light receiving portion, light from an automobile headlamp may be incident and erroneously detected.

  Further, in the electromagnetic detection method, it is not easy to embed a coil or loop antenna in the ground, and in order to cover an obstacle detection area, it is necessary to install a large number of coils or loop antennas as sensors. . Furthermore, in the electromagnetic detection system, a person in the obstacle detection area cannot be detected.

  In the image detection method, a lens unit such as a television camera or a CCD camera has to be wiped at appropriate intervals, and maintenance is not easy. Also, when the visibility is poor due to rain, snow, fog, etc., it may not be detected even if there is an obstacle.

  In Patent Document 2, a contrast pattern having a high contrast is formed on the background of the obstacle detection area by a difference in luminous intensity, a difference in reflectance, or spot light, and the obstacle detection area is photographed by a camera, and there is no obstacle. There is disclosed an image detection type obstacle detection device that compares a background image captured in advance with a currently captured current image, and determines the presence or absence of an obstacle from the presence or absence of the difference and the duration of the difference.

JP 2003-207562 A JP-A-8-55288

  The obstacle detection device described in Patent Document 2 is a danger detection type obstacle in which a difference is detected in a comparison between a background image and a current image and the obstacle is present when the detection continues for a certain period of time, that is, a danger detection type obstacle that is judged as dangerous. It is an object detection device. In the obstacle detection device of Patent Document 2, it is necessary to store the background image and the current image and to compare both images (collation processing), but the comparison between the background image and the current image is performed on all pixels of both images. Many burdens are placed on the arithmetic processing unit to perform the processing. In addition, since there is no self-check function when a camera breaks down or a bug occurs in the verification processing program, for example, a background image is detected due to a camera or arithmetic processing unit failure despite the presence of an obstacle on a railroad crossing. There is a possibility that no difference is detected in the comparison with the current image and there is no obstacle, that is, there is a possibility of being erroneously judged to be dangerous.

  Furthermore, for example, when the entry prohibition pattern is timed entry prohibition, such as a railroad crossing, it may be desired to measure the traffic flow during the time period when the entry prohibition is canceled (non-prohibited time period). Uses a separate measuring device that is independent of the obstacle detection device.

The first object of the present invention is to reduce the burden of the obstacle presence / absence determination processing in the image detection type obstacle detection device, and to ensure high safety by exhibiting a fail-safe function. Objective.
The second object of the present invention is to enable the means for achieving the above object to exhibit self-soundness.
A third object of the present invention is to enable measurement of traffic flow in a non-prohibited time zone in the obstacle detection area using the same device when the obstacle detection area has a non-prohibited time zone. There is to do.

  In order to achieve the first object, the obstacle detection device of the present invention includes an irradiation unit that irradiates a predetermined checker pattern on the obstacle detection region, and a checker pattern that is irradiated to the obstacle detection region by the irradiation unit. An image capturing unit that captures an image and outputs an image signal; and an arithmetic processing unit that processes the image signal output from the image capturing unit. The arithmetic processing unit detects a saddle point of the checker pattern from the image signal and determines the saddle point. When the absolute value of the difference between the count value and the reference value is below a predetermined threshold, it is determined that there is no obstacle in the obstacle detection area. When the absolute value of the difference between the count value and the reference value is larger than a predetermined threshold value, it is determined that there is an obstacle in the obstacle detection area.

  In another aspect of the present invention, the obstacle detection device shuts off the operation processing device from the electric level crossing breaker or the control circuit of the electric level crossing breaker existing in the vicinity of the obstacle detection region set in the level crossing. Railroad crossing obstacle detection device comprising a control unit that captures an image signal output from the photographing means between the time when a signal or a signal that detects the approach of a train is input and the time when a shut-off signal is input It is.

  In order to achieve the second object, the obstacle detection device of the present invention is characterized in that a reversing unit for temporarily or alternately reversing the lightness or darkness or hue of the checker pattern irradiated by the irradiation unit is added to the arithmetic processing unit. And

  In order to achieve the third object, the obstacle detection device of the present invention shoots the irradiation means for irradiating the obstacle detection area with the checker pattern, and the checker pattern irradiated to the obstacle detection area by the irradiation means. The image pickup means for outputting an image signal, and an arithmetic processing device for processing the image signal output from the image pickup means, the arithmetic processing device detects and counts saddle points from the picked-up image input from the image pickup means, Count the saddle points that are not detected based on the count value and the stored reference value in time series, and estimate the traffic flow in the obstacle detection area for a certain time from the time series count value. Features.

In another aspect of the present invention, the arithmetic processing unit recognizes a shape formed by a set of saddle points that are not detected or its contour by the image processing unit, and performs recognition processing of the shape and contour in time series, recognized shape and contour or al, automobile type of object to be detected, bicycles, wheelchairs, people, determines that a relatively large pet, by aggregating the calculation unit, the traffic at a given time of the obstacle detection area It measures the flow.

  In still another aspect of the present invention, the obstacle detection device photographs the surface of the electric railroad crossing breaker installed at both ends of the railroad crossing and the line side surface of the electric barrier crossing lowered to the interruption position. And a processing unit for processing an image signal output from the photographing means, and the interceptor is printed with a checker pattern on the surface on the track side or irradiated with the checker pattern by the irradiation means The arithmetic processing unit detects the saddle point of the checker pattern from the image signal, counts the saddle point, compares the count value with a pre-stored reference value, and calculates the absolute value of the difference between the count value and the reference value. Is less than or equal to a predetermined threshold, it is determined that there is no obstacle in the obstacle detection area. If the absolute value of the difference between the count value and the reference value is greater than the predetermined threshold, there is an obstacle in the obstacle detection area. And Characterized in that it is intended to.

  ADVANTAGE OF THE INVENTION According to this invention, while aiming at the burden reduction of an obstacle presence determination process, a fail safe function can be exhibited and high safety | security can be ensured. Moreover, self-soundness can be demonstrated. Furthermore, when the obstacle detection area has a non-prohibited time zone, the traffic flow in the non-prohibited time zone in the obstacle detection area can be estimated or measured using the same device.

It is a top view of a level crossing which shows a state where a checker pattern is irradiated to a level crossing in an embodiment of the invention. It is a block diagram which shows the structure of the obstruction detection apparatus which concerns on embodiment of this invention. It is a figure which shows the example of three types of checker patterns irradiated to a railroad crossing. It is a figure explaining the saddle point of a checker pattern. It is a flowchart explaining the effect | action of the obstruction detection apparatus of FIG. It is a time chart which shows the relationship of the timing of each operation | movement of the obstruction detection apparatus of FIG. It is a flowchart explaining the self-diagnosis function of the obstacle detection apparatus of FIG. It is a top view of a level crossing explaining an embodiment in which a checker pattern is printed on a breaker of an electric level crossing breaker. It is a figure which shows the barrier when applying to a general barrier, (a) is the front view seen from the road side, (b) is the back view seen from the crossing inside. It is a figure which shows the barrier when applying to a large aperture barrier, (a) is the front view seen from the road side, (b) is the back view seen from the level crossing inside. It is a top view of a level crossing explaining embodiment which irradiates the checker pattern to the barrier of the electric level crossing breaker.

<First Embodiment>
Next, a first embodiment in which the obstacle detection device of the present invention is applied to crossing obstacle detection will be described with reference to the drawings.

  The level crossing obstacle detection apparatus according to the present embodiment includes obstacles such as automobiles, bicycles, wheelchairs, and people in a level crossing road 3 where the road 1 in FIG. 1 and track tracks 2a and 2b intersect on the same plane. Is detected. FIG. 1 shows a railroad crossing 3 in which a road 1 and two tracks 2a and 2b intersect on the same plane, and an electric railroad crossing breaker 4a is provided on one side of one end of the railroad crossing 3 in addition to the railroad crossing 3. An example in which an electric railroad crossing breaker 4b is installed on the other side of the end is shown. However, this is only an example, and the length and width of the railroad crossing, the number of tracks, the installation position of the electric railroad breaker, etc. are arbitrary.

  As shown in FIG. 2, a crossing obstacle detection apparatus A according to the present embodiment includes a projector 10 as an illuminating unit, a video camera (hereinafter referred to as a camera) 20 as an imaging unit, and an arithmetic processing unit 30. Have

  As shown in FIG. 1, the projector 10 irradiates the obstacle detection area CA in the railroad crossing 3 with a checker pattern CP with near infrared light. When the checker pattern CP is irradiated with visible light, when the checker pattern CP is photographed by the camera 20 described later, the camera is affected by sunlight or the brightness of the checker pattern CP changes depending on the time zone. Since a problem arises in the image quality of 20, near infrared light is preferably used. However, for example, it may be better to irradiate the checker pattern CP with visible light for nighttime crime prevention purposes.

  The obstacle detection area CA is a rectangular portion surrounded by the side rails 3a, 3b on both sides of the railroad crossing 3 and the blocking rails 5a, 5b of the electric railroad crossing machines 4a, 4b installed at both ends of the railroad crossing 3 or from it. Set slightly inside.

  The checker pattern CP is a pattern in which a large number of square or rectangular squares are continuous in a matrix, and one of the neighboring squares is bright, the other is dark, or one and the other are colored in opposite colors. The checker pattern CP illustrated in FIG. 1 has eight squares in the longitudinal direction and the width direction of the railroad crossing 3, for a total of 64 squares. The number of squares in the longitudinal direction and the width direction naturally depends on the area of the obstacle detection area CA and the area of one square.

  The installation position of the projector 10 is preferably a position where the entire shape of the checker pattern CP irradiated to the entire area of the obstacle detection area CA or the inside thereof is not distorted as much as possible. The optimum installation position is an upper position in the center of the obstacle detection area CA, and it is preferable that the irradiation is performed downward from the position toward the obstacle detection area CA. Further, as shown in FIG. 1, when there are a plurality of tracks, it is preferable to irradiate from the upper position on the side of the railroad crossing 3 between the tracks 2a and 2b toward the center of the obstacle detection area CA. . However, when installation is difficult above the center of the obstacle detection area CA or between the lines 2a and 2b, or when the line is a single line, the projector 10 is outside the building limit line on one end side of the railroad crossing 3. For example, it can be installed on the side of the level crossing 3 inside the electric level crossing breaker 4a or 4b.

  The camera 20 takes an image of the checker pattern CP irradiated to the obstacle detection area CA by the projector 10. Therefore, a near infrared camera is used as the camera 20. However, when the projector 10 emits visible light as described above, a visible light camera is used.

  The installation position of the camera 20 may be the same as that of the projector 10 as long as the entire checker pattern CP irradiated to the obstacle detection area CA can be photographed with as little distortion as possible. As illustrated in FIG. Another position such as the side of the railroad crossing 3 opposite to the projector 10 may be used.

  The projector 10 and the camera 20 are electrically connected to the arithmetic processing unit 30 as shown in FIG. The arithmetic processing unit 30 may be provided in the vicinity of the railroad crossing or may be placed in a signal equipment room or the like. As will be described later in detail, the arithmetic processing unit 30 is connected to the electric level crossing breaker 4a or 4b of the railroad crossing 3 and is shut off from the electric level crossing breaker 4a (or 4b, hereinafter the same). Signal s1 and shut-off opening signal s2 are input. The arithmetic processing unit 30 is connected to the ground control device 40 and, as will be described later, issues an alarm and a warning signal to the train when an obstacle is detected in the obstacle detection area CA. A drive signal s3 for stopping the train by operating a traffic light or a traffic light on the ground is given to the ground control device 40.

  The breaker closing signal s1 and the breaker opening signal s2 may be input from a control circuit (not shown) of an electric level crossing breaker provided outside the electric level crossing breaker 4a instead of the electric level crossing breaker 4a. Moreover, you may use the signal (For example, signal which starts ringing of an alarm device) which detected the train approach to the railroad crossing 3 instead of the interruption | blocking cannula closing signal s1.

  The arithmetic processing device 30 includes a control unit 31, a storage unit 32, an image processing unit 33, a calculation unit 34, a determination unit 35, and an output unit 36. The storage unit 32 stores a program storage unit 321 in which various programs for realizing each function of the arithmetic processing device 30 are stored, and an irradiation image storage unit 322 in which a checker pattern CP irradiated from the projector 10 is stored. A captured image storage unit 323 that temporarily stores a captured image input from the camera 20, and the total number of saddle points of the checker pattern CP or the number of saddle points expected to be detected in the state where there is no obstacle in the obstacle detection area CA. A reference value storage unit 324 that stores the reference value and a count value storage unit 325 that temporarily stores the count value of saddle points detected from the captured image captured from the camera 20 are included.

  As illustrated in FIG. 3, the irradiation image storage unit 322 stores a first checker pattern CP <b> 1 and a second checker pattern CP <b> 2 obtained by inverting the brightness of the first checker pattern CP <b> 1. Further, a third checker pattern CP3 having no checker on the entire surface may be stored.

  The image processing unit 33 has a function of detecting a saddle point by analyzing a captured image captured from the camera 20 that captured the obstacle detection area CA and stored in the captured image storage unit 323. More specifically, the image processing unit 33 reads the image signal frame by frame from the captured image storage unit 323, and among the image signals of the one frame, the connection points of adjacent cells of the same brightness or the same color of the checker pattern CP. That is, the saddle point is detected.

  The saddle point of the checker pattern CP is a portion b where the corners of the adjacent low-lightness square a1 (or the corners of the high-lightness square a2) are adjacent as shown by circles in FIG. The saddle point b can be detected by a feature detection method using a known algorithm for edge detection or corner detection in an image. Here, a detailed description is omitted. In particular, detection of saddle points when the position of saddle points is known has a feature that the processing load is small. In addition, since the dependency between individual saddle points is sparse, it has a high affinity with high-speed techniques such as parallel processing.

  The calculation unit 34 counts saddle points b detected from the checker pattern CP irradiated on the entire obstacle detection area CA. The count value is stored in the count value storage unit 325.

  Depending on the irradiation direction of the checker pattern CP from the projector 10 to the obstacle detection area CA and / or the shooting direction of the camera 20 to the obstacle detection area CA, the checker pattern CP irradiated to the obstacle detection area CA is distorted. There is a case. Further, the railroad crossing 3 is not limited to a single flat surface, but may include a combination of a plurality of inclined surfaces, or one or both of a convex surface and a concave surface. In such a case, since it is difficult to accurately detect and count saddle points from the captured image, the image processing unit 33 stores the distorted captured image read from the captured image storage unit 323 in one plane. A function of performing a correction process for converting to an image is provided. Further, the image processing unit 33 may have a function of distorting the checker pattern CP output to the projector 10 by inverse transformation so that the distortion of the irradiated checker pattern CP is reduced.

  The determination unit 35 compares the count value Nc stored in the count value storage unit 325 with the reference value Ns stored in the reference value storage unit 324, and the difference D between the count value Nc and the reference value Ns is a predetermined threshold value. When it is T or less (D ≦ T), it is determined that no obstacle exists in the obstacle detection area CA, that is, it is safe.

  Here, the threshold value T is a saddle point detected in a captured image when an object exists in the obstacle detection area CA but the size of the object is such that there is no possibility of colliding with the train. The number of can be determined by reference. Extremely speaking, when no object is detected, that is, when the count value Nc is equal to the reference value Ns (when D is 0), it is surely safe, and obstacle detection that determines that this time is safe The device is a safety confirmation type obstacle detection device.

  However, even if an object is present in the obstacle detection area CA, it cannot be said that the object is an obstacle if the object is of a size that does not cause a collision with the train. The threshold T for safety confirmation is the number of saddle points that cannot be detected due to the size of the smallest object that needs to avoid the collision of the train (in other words, the size of the largest object that is sufficiently unlikely to collide with the train). Even in such a case, it is confirmed that there is no object that needs to avoid the collision of the train, so it is a safety confirmation type. Unlike the danger detection type described in Patent Document 2, the safety confirmation type can meet higher safety requirements. Objects that need to avoid a train collision include, for example, cars, bicycles, wheelchairs, people, livestock, relatively large pets. It can also be confirmed as safe when the count value is below a certain percentage of the total number of saddle points (reference value).

  The operation of the crossing obstacle detection device A having the above-described configuration will be described with reference to FIGS. First, the controller 31 monitors whether or not the shut-off closure signal s1 is input from the electric railroad crossing breaker 4a (step S101). When the shut-off closure signal s1 is input, the irradiation image storage section The first checker pattern CP1 is read from 322 and applied to the projector 10, and the obstacle check area CA is irradiated with the first checker pattern CP1 as shown in FIG. 1 (step S102). At almost the same time, an image signal is received from the camera 20 that is photographing the obstacle detection area CA, and is temporarily stored in the photographed image storage unit 323 (step S103).

  In this case, the irradiation of the first checker pattern CP1 by the projector 10 may be performed continuously as shown in (3) of FIG. 6 or, for example, several tens of ms as shown in (3 ′) of FIG. It may be performed intermittently every 1 second interval. Further, the photographing by the camera 20 may be performed continuously, or may be performed intermittently at intervals of several tens of ms to 1 second, for example, but the image signal from the camera 20 is captured by the projector 10. It is preferable to carry out intermittently at intervals of, for example, several tens of ms to 1 second in the time overlapping with the time. In this way, as will be described later, not only the presence of an obstacle can be detected, but also the movement of the obstacle can be detected.

  Next, the image processing unit 33 performs processing for correcting the distortion of the captured image read from the captured image storage unit 323 (step S104). Subsequently, the saddle point b in the first checker pattern CP1 of the corrected captured image is determined. The detected saddle point b is counted by the calculation unit 34 (step S105), and the count value is temporarily stored in the count value storage unit 325 (step S106).

  Then, the determination unit 35 compares the count value Nc temporarily stored in the count value storage unit 325 with the reference value Ns of the first checker pattern CP1 read from the reference value storage unit 324 (step S107). It is determined whether or not the absolute value | D | of the difference between the reference values Ns is equal to or smaller than a predetermined threshold T (| D | ≦ T) (step S108).

  If | D | ≦ T, it is determined that there is no obstacle in the obstacle detection area CA (step S109). Therefore, in this case, the arithmetic processing device 30 does not transmit the drive signal s3 to the ground control device 40.

  In S102, when the irradiation of the first checker pattern CP1 is not performed continuously, for example, when it is performed intermittently at intervals of several tens of ms to one second, a timer t1 of several tens of ms to one second is set simultaneously with the start of irradiation. In step S110, it is checked whether the timer t1 has expired. When the time is up, it is checked whether or not the shut-off opening signal s2 is input (step S111). If the shut-off opening signal s2 is not input, the process returns after step S101, and the first checker pattern CP1 at the next timing is detected. An irradiation process is performed.

  On the other hand, if the determination result in step S108 is not | D | ≦ T, the determination unit 35 determines that there is an obstacle in the obstacle detection area CA when the same determination continues for a certain time or more (step S108). S112). In response to this determination, the output unit 36 of the arithmetic processing device 30 transmits the drive signal s3 to the ground control device 40 (step S113). Therefore, the ground control device 40 issues a warning and a warning signal to the train, and operates the train traffic light or ground traffic light to stop the train.

  In the above description, the projector 10 and the camera 20 are driven and controlled from the time when the shut-off / closing signal s1 is input from the electric railroad crossing breaker 4a to the time when the shut-off / opening signal s2 is input, thereby obstructing the obstacle. However, the drive and control of the projector 10 and the camera 20 may be started when the alarm device that notifies the train approaching to the railroad crossing 3 starts ringing.

  However, when the presence or absence of an obstacle is determined by driving and controlling the projector 10 and the camera 20 between the time when the shut-off switch closing signal is input or the time when the alarm ringing is started and the time when the shut-off opening signal is input, When it arrives at the railroad crossing 3, the train itself is judged as an obstacle. In order to avoid this, when the train arrives at the railroad crossing 3, the detection operation of the railroad crossing obstacle detection device A may be interrupted. As a method for interrupting the detection operation, a method used in a conventional level crossing obstacle detection device can be applied. For example, the arrival of a train can be detected by a signal indicating that the train has approached the end point of a railroad crossing.

  In order to avoid the train itself being judged as an obstacle, an upper limit is set on the number of saddle points b that are not detected by a normal obstacle, and the absolute value of the difference between the count value Nc and the reference value Ns | D | May exceed the upper limit, the determination unit 35 may design the program so as not to determine that there is an obstacle.

  The first embodiment uses an image processing technique with a small processing load called saddle point detection, and further compares the count value Nc of the saddle point b of the checker pattern CP and the digital value of the reference value Ns (| D | ≦ T) Since the presence / absence of an obstacle is judged, the burden of the judgment processing of the arithmetic processing unit is reduced, and since it is a safety confirmation type that judges that there is no obstacle when | D | ≦ T, fail-safe is realized. The

<Second Embodiment>
In the second embodiment of the present invention, the level crossing obstacle detection device A uses the comparison between the reference value Ns and the count value Nc obtained by the irradiation of the checker pattern CP and the processing of the captured image. It also has a self-diagnosis function to check the soundness.

  That is, the control unit 31 performs the operation from the time when the shut-off control signal s1 is inputted from the electric railroad crossing breaker 4a to the time when the shut-off control signal s2 is inputted, preferably at the time of N in step S111. The process proceeds to step S114 shown in FIG. Then, instead of the first checker pattern CP1 irradiated so far, the second checker pattern CP2 is irradiated (step S114). That is, the contrast of the checker pattern CP is reversed. Subsequently, the processing of steps S103 to S107 described above is performed in step S115, and it is confirmed in S116 whether or not the checker pattern CP has been inverted. When the checker pattern CP is inverted, it is subsequently determined whether the count values Nc1, Nc2 of the saddle point b immediately before and after the inversion or the differences D1, D2 thereof are equal (Nc1 = Nc2 or D1 = D2) (step) S117). If the checker pattern CP is not reversed in S116, it is determined that there is an abnormality in any of the crossing obstacle detection devices A including the projector 10 and the camera 20 (Step S119, an abnormality notification signal is output to a maintenance area or the like) (Step S120) and the self-diagnosis process is terminated.

  Instead of comparing the difference values D1 and D2 of the saddle point b immediately before and after the inversion, it may be determined whether the count values Nc1 and Nc2 of the saddle point b immediately before and after the inversion are equal. . Then, when D1 = D2 or Nc1 = Nc2, the determination unit 35 determines that the entire system including the projector 10 and the camera 20 is normal (healthy) (step S118), and the process proceeds to S101 in FIG. . On the other hand, if D1 = D2 or Nc1 = Nc2 is not satisfied in step S116, it is determined that there is an abnormality in any of the level crossing obstacle detection devices A including the projector 10 and the camera 20 (step S119), and a maintenance zone, etc. An abnormality notification signal is output (step S120), and the self-diagnosis process is terminated.

  It should be noted that the state of the obstacle detection area CA changes during a minute time during which the checker pattern CP is reversed, and there is a difference between the count values Nc1 and Nc2 of the saddle point b even though there is no abnormality in the crossing obstacle detection device A. It can happen. For this reason, in step S117, it may be determined whether or not the difference between D1 and D2 or Nc1 and Nc2 is equal to or less than a predetermined threshold value.

  As a method for confirming that the checker pattern CP is inverted based on a command from the control unit 31, there is the following method. Whether the image before reversal (here, I image) and the image after reversal (here, J image) are acquired and reversed using the absolute value of the difference between the I image and J image To check. In the image for which the absolute value of the difference is obtained, the entire area should be white (the pixel value of the 8-bit grayscale image is 255) (the pixel value is bright, such as 200, but overflows depending on the degree of imaging) Sometimes not). Therefore, it is checked whether or not there is a pixel having a threshold value (for example, 200) or less in all pixels. When there is no pixel below the threshold, inversion is detected. If not reversed, a black region exists in the image for which the absolute value of the difference has been obtained, and can be detected. Similarly, when the checker pattern CP uses hue reversal instead of light and dark, the entire area of the image for which the absolute value of the difference is obtained is white (a bright color with low saturation). In the case where the image is not inverted, it can be detected because there is a dark and highly saturated color in the image for which the absolute value of the difference is obtained.

  In the above processing flow, the self-health check is executed once for the two captured image captures of the first checker pattern CP1 and the second checker pattern CP2, but at the time when the shut-off closure signal s1 is input. The self-soundness check may be executed every time for the second and subsequent captured image capture.

  In the second embodiment described above, in the crossing obstacle detection device that determines the presence or absence of an obstacle by comparing the count value Nc of the number of saddle points of the checker pattern CP with the reference value Ns, the lightness / darkness or hue of the checker pattern CP is determined. By reversing, since the soundness is checked based on the presence or absence of the difference, a self-diagnosis function can be realized by adding a simple program without adversely affecting obstacle detection.

<Third Embodiment>
As still another embodiment of the present invention, the crossing obstacle detection device A utilizes the point of detecting the presence or absence of an obstacle by irradiating the checker pattern CP from the projector 10 and processing the image signal from the camera 20. Thus, it also has a function of measuring the traffic flow while the railroad crossing 3 is open (non-prohibited time zone).

  That is, the level crossing obstacle detection device A performs the obstacle detection in the obstacle detection area CA described above during the non-prohibited time period and measures the traffic flow on the level crossing road 3.

More specifically, the railroad crossing obstacle detection device A continues to irradiate the first checker pattern CP1 when the breaking rail opening signal s2 is input from the electric railroad crossing breaker 4a. Then, the saddle point b is detected and counted from the captured image input from the camera 20, and the saddle point b that is not detected is counted based on the count value Nc and the reference value Ns. Such counting of saddle points b that are not detected is performed in time series, that is, at predetermined time intervals, for example, at intervals of several tens of ms to 1 second. From a time-series count value of the saddle point b that is not detected to a certain time of the railroad crossing 3 (obstacle detection area CA) or until the next shut-off closing signal s1 is input from the electric railroad crossing breaker 4a (non-prohibited time) Can be estimated. Alternatively, the image processing unit 33 recognizes the shape formed by the set of saddle points b that are not detected or the contour thereof, and recognizes the shape and the contour in time series, that is, at predetermined time intervals, for example, several tens of ms to 1 Perform at intervals of seconds. Recognized shape and contour or al, automobile type of object to be detected, bicycles, wheelchairs, people, determines that a relatively large pet, by aggregating the calculation unit 34, crossing roads 3 (obstacle detection area It is possible to measure the traffic flow during a certain time or non-prohibited time period of CA.

  When measuring the traffic flow, instead of irradiating the first checker pattern CP1, the non-checker pattern CP3 as shown in FIG. 3C is irradiated, and the captured image from the camera 20 is captured by the image processing unit 33. It is also possible to perform known shape recognition processing and measure traffic flow by type such as automobiles, bicycles, and people.

  In the third embodiment, the saddle point b is detected and counted in time series from the captured image of the checker pattern CP irradiated to the obstacle detection area CA, and the shape is determined from the change in the count value Nc or from the captured image. Since traffic flow in the obstacle detection area CA is estimated or measured from time-series changes of objects obtained by recognition, a part of the irradiation means, photographing means, and arithmetic processing device of the crossing obstacle detection device is used in common. Obstacle detection and traffic flow estimation or measurement in the obstacle detection area CA can be performed.

<Fourth embodiment>
In FIG. 8, the checker pattern CP is photographed with a camera, and whether the count value of the saddle point b in the photographed image is equal to the reference value or whether the difference between the count value and the reference value does not exceed the threshold value. Another example of a crossing obstacle detection device that determines whether or not an obstacle exists is shown.

  This level crossing obstacle detection device is configured such that the barrier cans 5c and 5d of the electric railroad crossing machines 4c and 4d installed at both ends of the railroad crossing 3 and the back surfaces of the barriers 5c and 5d lowered to the blocking position, that is, The camera 20A and 20B which image | photograph the surface by the side of the track | line 2c, and arithmetic processing unit 30A (not shown) are comprised.

  Since the shut-off cans 5c and 5d have the same configuration, one shut-off can 5c will be described in detail. FIGS. 9A and 9B are a front view seen from the road side and a rear view seen from the inside of the railroad crossing when the barrier can 5c is a general interrupter. FIGS. 10A and 10B are a front view seen from the road side and a rear view seen from the inside of the railroad crossing when the shut-off can 5c is a large-diameter shut-off can. Thus, the checker pattern CP is printed on the back surface of the blocking can 5c. This checker pattern CP corresponds to a part of the first checker pattern CP1 in FIG.

  The cameras 20A and 20B in FIG. 8 are installed on or near the electric level crossing breaker on the side opposite to the electric level crossing breaker having the breaking rod to be photographed on the lines 2c and 2d. That is, the camera 20A is installed at or near the electric level crossing breaker 4c on the opposite side of the electric level crossing breaker 4d having the blocking canopy 5d to be photographed. Further, the camera 20B is installed at or near the electric level crossing breaker 4d opposite to the electric level crossing breaker 4c having the shut-off can 5c to be photographed. Therefore, the range of the triangle T1 connecting the camera 20A and both ends of the shut-off can 5d at the lowered position of the electric railroad crossing breaker 4d is one obstacle detection area, and the lower position of the camera 20B and the electric railroad crossing breaker 4c is at the lower position. A range of a triangle T2 connecting both ends of a certain shut-off can 5c is another obstacle detection area.

  The image signals output from the cameras 20A and 20B are respectively subjected to saddle point detection / counting and comparison between the count value and the reference value in the arithmetic processing unit 30A, as in the previous embodiment. Based on this, it is determined whether or not an obstacle exists in each obstacle detection area.

  Each of the cameras 20A and 20B can be attached to the front end of a shut-off canopy facing the shut-off canopy that is the subject of photography or in the vicinity thereof. In the example shown in FIG. 11, a camera 20A that captures the interrupting can 5d is attached to the tip of the interrupter 5c, and a camera 20B that captures the interrupter 5c is attached to the distal end of the interrupter 5d.

  The checker pattern CP on the back surface of the blocking can 5c may be irradiated with the checker pattern CP from a projector instead of printing. In the example shown in FIG. 11, projectors 10 </ b> A and 10 </ b> B are attached to the front ends of the respective shut-off cans 5 a and 5 b or in the vicinity thereof.

  Since this embodiment prints or irradiates the checker pattern CP without interruption, unlike the case of irradiating a railroad crossing, there is an advantage that it is not easily affected by snow accumulation or heavy rain, so there is a large amount of snowfall or rainfall. It is suitable as a crossing obstacle detection device installed in the area.

  The fourth embodiment has an advantage that an obstacle can be detected without being affected by snowfall, rain, flooding, or the like, since the shutter can be photographed at the blocking position.

<Other embodiments>
Although the present invention has been described above with respect to a crossing obstacle detection device applied to a crossing in which electric crossing breakers exist at both ends of the obstacle detection area CA, the present invention is not limited to the presence or absence of a non-prohibited time zone. It can be applied to detect an obstacle in an obstacle detection area other than a level crossing.

DESCRIPTION OF SYMBOLS 1 Road 2a, 2b, 2c, 2d, 2e Line 3 Railroad crossing 4a, 4b, 4c, 4d Electric railroad crossing-off machine 5a, 5b, 5c, 5d Barrier 10, 10, A, 10B Projector (irradiation means)
20, 20A, 20B Camera (photographing means)
30 arithmetic processing unit CP checker pattern a1 square with low brightness (black square)
a2 High brightness square (white square)
b Saddle point CP1 First checker pattern CP2 Second checker pattern CP3 Third checker pattern

Claims (6)

Irradiation means for irradiating the obstacle detection area with a predetermined checker pattern, imaging means for imaging the checker pattern irradiated to the obstacle detection area by the irradiation means, and outputting an image signal, and output from the imaging means An arithmetic processing unit for processing an image signal;
The arithmetic processing unit detects saddle points of the checker pattern from the image signal, counts the saddle points, compares the count value with a reference value stored in advance, and calculates a difference between the count value and the reference value. When the absolute value is less than or equal to a predetermined threshold, it is determined that there is no obstacle in the obstacle detection area, and when the absolute value of the difference between the count value and the reference value is greater than a predetermined threshold, the obstacle detection An obstacle detection device for determining that an obstacle exists in an area.   The obstacle detection device according to claim 1, wherein the arithmetic processing unit is connected to an electric crossing breaker or a control circuit of the electric crossing breaker existing in the vicinity of an obstacle detection region set on a railroad crossing or A fault comprising a control unit that captures the image signal output from the imaging unit between the time when a signal that detects the approach of a train is input and the time when a shut-off signal is input Object detection device.   3. The obstacle detection device according to claim 1, wherein an inversion unit that temporarily or alternately inverts the brightness or darkness or hue of the checker pattern irradiated by the irradiation unit is added to the arithmetic processing unit. Obstacle detection device.   Irradiation means for irradiating the obstacle detection area with the checker pattern, imaging means for imaging the checker pattern irradiated to the obstacle detection area by the irradiation means and outputting an image signal, and image signal output from the imaging means The arithmetic processing unit detects and counts saddle points from the photographed image input from the photographing means, and detects saddle points that are not detected based on the counted value and the stored reference value. An obstacle detection device characterized in that the counting is performed in time series, and the traffic flow in the obstacle detection area in a certain time is estimated from the time series count values. 5. The obstacle detection device according to claim 4, wherein the arithmetic processing unit recognizes a shape formed by a set of saddle points that are not detected or an outline thereof by an image processing unit, and performs recognition processing of the shape and the outline in time series. to perform, recognized shape and contour or al, automobile type of object to be detected, bicycles, determines wheelchairs, people, relatively large pet, by aggregating the calculation unit, the obstacle detection region An obstacle detection device characterized by measuring traffic flow in a certain time.   Shutters of electric railroad crossing breakers installed at both ends of the railroad crossing, photographing means for photographing the back surface of the barrier that is lowered to the shutoff position, and calculation for processing image signals output from the photographing means A checker pattern printed on the inner surface of the railroad crossing or irradiated by irradiation means, and the arithmetic processing unit detects a saddle point of the checker pattern from the image signal. The saddle point is counted, and the count value is compared with a reference value stored in advance. When the absolute value of the difference between the count value and the reference value is equal to or less than a predetermined threshold value, the obstacle detection area is displayed. It is determined that there is no obstacle, and when the absolute value of the difference between the count value and the reference value is greater than the threshold, it is determined that there is an obstacle on the railroad crossing. Dress .

Images