Intrusion detection system

Abstract

This invention provides an intrusion detection system that prevents the system from mistakenly identifying the entry of workers or other personnel into restricted areas as an intrusion. [Solution] The intrusion detection system 1 determines that an intrusion into the restricted area R has occurred when a human Z is detected within the restricted area R. However, if multiple human Zs are detected within the restricted area R, and the colors of the same body part detected for each of the multiple human Zs fall within a predetermined color range, then it is not determined that an intrusion into the restricted area R has occurred.

Description

Technical Field

[0001] The present invention relates to an intrusion determination system for suppressing the determination that the entry of workers or the like into a prohibited intrusion area is an intrusion.

Background Art

[0002] Conventionally, a technique capable of accurately determining the entry of a moving object into a prohibited entry area shown in an image has been known (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above technique, for example, by setting a line area shown in an image as a prohibited intrusion area, it is possible to determine that the entry of a person into the line is an intrusion.

[0005] However, since line maintenance workers frequently enter the line for line maintenance, if the above technique is used as it is, every time a line maintenance worker enters the line for line maintenance, it is determined as an intrusion, resulting in a problem of impairing the effectiveness of intrusion determination.

[0006] Therefore, an object of the present invention is to provide an intrusion determination system for suppressing the determination that the entry of workers or the like into a prohibited intrusion area is an intrusion.

Means for Solving the Problems

[0007] The present invention provides an intrusion detection system comprising: an acquisition unit that acquires an image captured by a shooting means installed to photograph a predetermined range; a storage unit that stores a human intrusion prohibited area set within the predetermined range and detection criteria for detecting at least a part of a person captured in the image; a first detection unit that can detect a person in the image based on the feature quantities captured in the image and the detection criteria; a second detection unit that can detect the color of the detected person; and a determination unit that determines that an intrusion into the prohibited area has occurred when a person is detected within the prohibited area, wherein the determination unit does not determine that an intrusion into the prohibited area has occurred when multiple people are detected in the prohibited area and the color of the same part detected for each of the multiple people is included in a predetermined color range.

[0008] With this configuration, if multiple people wearing identical uniforms of conspicuous colors are detected, for example, if multiple workers enter a restricted area for track maintenance, it will not be determined as "entry into a restricted area," thus preventing the effectiveness of intrusion detection from being compromised.

[0009] Furthermore, according to another aspect of the present invention, an intrusion detection program and an intrusion detection method corresponding to the above-mentioned intrusion detection system are provided. [Effects of the Invention]

[0010] According to the intrusion detection system of the present invention, it is possible to suppress the misinterpretation of workers entering restricted areas as intrusions. [Brief explanation of the drawing]

[0011] [Figure 1] Diagram illustrating human intrusion into a restricted area according to an embodiment of the present invention. [Figure 2] Block diagram of an intrusion detection system according to an embodiment of the present invention. [Figure 3] Diagram illustrating an intrusion-restricted area according to an embodiment of the present invention. [Figure 4]Diagram illustrating multiple intrusion-restricted areas according to embodiments of the present invention. [Modes for carrying out the invention]

[0012] The intrusion detection system 1 according to an embodiment of the present invention will be described below with reference to Figures 1-4.

[0013] As shown in Figure 1, the intrusion detection system 1 is for determining whether a person Z has entered a restricted area R as captured in an image Y (frames constituting the video in Figure 1) taken by the imaging means X. In this embodiment, for ease of understanding, the person Z is simply represented by their skeleton.

[0014] As shown in Figure 2, the intrusion detection system 1 comprises an acquisition unit 2, a storage unit 3, a first detection unit 4, a second detection unit 5, a determination unit 6, and a notification unit 7.

[0015] The acquisition unit 2 acquires an image Y captured by the shooting means X, which is installed to photograph a predetermined range. In this embodiment, the frames constituting the video captured by the shooting means X are used as image Y. In this embodiment, the entire shooting range of the shooting means X is set as the predetermined range.

[0016] The memory unit 3 stores a restricted area R set within the predetermined range for human Z, and detection criteria for detecting at least a portion of human Z as seen in the image Y.

[0017] In this embodiment, the restricted area R is to be memorized by the user. Figure 3 shows an example where the area near the train tracks is memorized as the restricted area R.

[0018] Furthermore, in this embodiment, the following detection criteria are stored: “joint identification criteria,” “behavioral body identification criteria,” and “behavioral identification criteria.”

[0019] The "joint identification criteria" are for identifying a plurality of joints (neck, right elbow, left elbow, waist, right knee, left knee, etc.) of human Z, and for each joint, they show feature amounts such as shape, direction, size, etc. for identifying each joint.

[0020] The "behavior body identification criteria" show the "basic posture", "range of motion of each joint", "distance between each joint" in a single human Z, etc. for various variations ("walking", "standing upright", etc.) of human Z.

[0021] The "behavior identification criteria" show the movement of each joint when human Z performs a predetermined behavior. In this embodiment, the "behavior identification criteria" for various types of behaviors are stored.

[0022] The first detection unit 4 can detect human Z in the image Y based on the feature amounts reflected in the image Y and the above detection criteria.

[0023] In this embodiment, when detecting human Z, first, human Z reflected in the image Y is identified.

[0024] Specifically, after detecting the feature amounts corresponding to the "joint identification criteria" stored in the storage unit 3, referring to the "behavior body identification criteria", a plurality of joints included in a single human Z are identified. In the example of FIG. 1, all joints are identified as the joints included in a single human Z, and it is identified that there is a single human Z.

[0025] The second detection unit 5 can detect the color of the detected human Z. This means detecting the clothing, etc. of the detected human Z.

[0026] For the color, it is conceivable to detect the numerical values of R, G, and B for each pixel on the image Y respectively.

[0027] For example, if person Z is wearing an orange-colored jacket, R, G, and B values ​​for orange will be detected almost entirely from the shoulders to the waist. Similarly, if person Z is wearing a yellow-colored helmet, R, G, and B values ​​for yellow will be detected all over the head. Furthermore, if person Z is wearing a blue-colored safety strap across both shoulders, R, G, and B values ​​for blue will be detected along the strap from both shoulders to the waist.

[0028] As shown in Figure 1, the determination unit 6 determines that "intrusion into the restricted area R" has occurred when a person Z is detected within the restricted area R. In this embodiment, "intrusion into the restricted area R" is determined when the feet of person Z overlap with the restricted area R.

[0029] However, for example, since maintenance workers frequently enter the railway tracks for track maintenance purposes, using the above technology as is would result in an intrusion being detected every time a maintenance worker enters the tracks for maintenance, thus undermining the effectiveness of the intrusion detection.

[0030] Therefore, in this embodiment, as shown in Figure 3(a), when multiple people Z1 are detected in the restricted area R, the determination unit 6 does not determine that "intrusion into the restricted area R" has occurred if the colors of the same body part detected for each of the multiple people Z1 fall within a predetermined color range.

[0031] This observation focuses on the fact that, for example, track maintenance is often carried out by multiple maintenance workers wearing identical uniforms in conspicuous colors.

[0032] To exclude multiple track maintenance workers wearing orange-colored uniforms from the intrusion criteria, one possible approach is to set the designated color range to "hues (R, G, B) that correspond to orange."

[0033] Furthermore, in order to exclude multiple track maintenance workers wearing brightly colored uniforms (including fluorescent colors) from the scope of intrusion, it is conceivable to set the specified color range to "above a specified level of brightness (R, G, B)" or "above a specified level of saturation (R, G, B)".

[0034] Furthermore, to achieve more accurate determination, a predetermined color range may be set for a specific part of human Z1. This predetermined part can be detected, for example, based on the "joint identification criteria" described above.

[0035] For example, if a user sets a predetermined body part to "upper body" and a predetermined color range to "brightness and saturation (R, G, B) above a predetermined level" in the settings unit (not shown), then when multiple people Z1 wearing brightly colored uniforms are detected, it will be considered that multiple maintenance workers have entered the restricted area R for track maintenance purposes, and it will not be determined as "intrusion into the restricted area R".

[0036] On the other hand, in this case, if multiple students wearing uniforms with brightness and saturation below a specified level enter the restricted area R, it will be determined as "entry into the restricted area R."

[0037] Incidentally, while track maintenance may involve multiple maintenance workers, it is also possible for supervisors or other personnel not wearing uniforms to be present. In such cases, it is preferable that even if the supervisor or other personnel enter the restricted area R, it is not judged as "entering the restricted area R."

[0038] Therefore, in this embodiment, as shown in Figure 3(b), if the determination unit 6 detects multiple people Z1 whose body parts are colored within a predetermined color range, and one or more other people Z2 whose body parts are colored within the same range, the determination unit 6 makes a determination according to the ratio of the multiple people Z1 to one or more other people Z2.

[0039] In other words, since the number of supervisors not wearing uniforms is generally less than the number of track maintenance workers, for example, by setting it so that "if the number of people Z2 is less than the number of people Z1, it will not be determined as an 'intrusion into restricted area R'," even if only one supervisor not wearing a uniform is detected in restricted area R, it will not be determined as an 'intrusion into restricted area R'.

[0040] The ratio of the multiple individuals to the one or more other individuals can be set by the user as appropriate, depending on the desired accuracy of the intrusion detection.

[0041] Furthermore, by making judgments according to the above proportions, it becomes possible to make accurate judgments even in cases where color detection of some human Z1s fails.

[0042] Furthermore, as shown in Figure 4(a), for example, when multiple people Z1 (track maintenance workers) are performing track maintenance on one track, a third party Z3 may intrude onto another track, or as shown in Figure 4(b), even on the same track, a third party Z3 may intrude into a location far away from where track maintenance is being performed. In such cases, it is preferable that the intrusion be judged as "intrusion into a restricted area R".

[0043] Therefore, in this embodiment, if there are multiple restricted areas R1 and R2 within a predetermined range captured by the imaging means X, the determination unit 6 determines whether an intrusion has occurred "into the restricted areas R1 and R2" for each restricted area R1 and R2.

[0044] Note that the multiple restricted access zones R1 and R2 may be separated by a gap, as shown in Figure 4(a), or they may be adjacent, as shown in Figure 4(b). In the case of Figure 4(b), one restricted access zone R is divided into multiple restricted access zones R1 and R2.

[0045] The notification unit 7 issues a notification when the determination unit 6 determines that "intrusion into the restricted area R" has occurred. For example, an alert may be issued to a person Z using an audio output device installed near the restricted area R, or the notification of "intrusion into the restricted area R" may be sent to an information terminal of a facility manager or other person.

[0046] As explained above, in the intrusion detection system 1 according to this embodiment, if multiple people Z1 are detected in the restricted area R, and the color of the same body part detected for each of the multiple people Z1 falls within a predetermined color range, then it is not determined that "intrusion into the restricted area R" has occurred. Furthermore, it is sufficient that the color of the same body part falls within the same color range, and it is not necessary to set the body part or color in detail, so it can accommodate various uniforms, helmets, etc.

[0047] With this configuration, if multiple people Z wearing identical uniforms of conspicuous colors are detected, for example, if multiple workers enter a restricted area R for track maintenance, it will not be determined as "entry into restricted area R," thus preventing the effectiveness of intrusion detection from being compromised.

[0048] Furthermore, in the intrusion detection system 1 according to this embodiment, if multiple people Z1 whose body parts are colored within a predetermined color range and one or more other people Z2 whose body parts are colored within the same range are detected in the intrusion-restricted area R, a determination is made according to the ratio of the multiple people Z1 to one or more other people Z2.

[0049] With this configuration, even when multiple workers are present, such as when a supervisor or other person not wearing a uniform is present, the system is less likely to mistakenly determine that an entry has occurred in the restricted area R. Furthermore, it is possible to make accurate determinations even if the color detection of some individuals Z1 fails.

[0050] Furthermore, in the intrusion detection system 1 according to this embodiment, if there are multiple intrusion-restricted areas R1 and R2 within a predetermined range captured by the imaging means X, the system determines "intrusion into intrusion-restricted areas R1 and R2" for each of the intrusion-restricted areas R1 and R2.

[0051] With this configuration, it becomes possible to prevent intrusion by third parties even when workers are performing their duties.

[0052] Furthermore, the intrusion detection system of the present invention is not limited to the embodiments described above, and various modifications and improvements are possible within the scope described in the claims.

[0053] For example, in the above embodiment, the explanation was given using an example where the predetermined body part was set to the "upper body," but in order to further improve the accuracy of the determination, multiple predetermined body parts may be set, and a predetermined color range may be set for each of these multiple predetermined body parts.

[0054] Furthermore, in the above embodiment, the explanation was given using an example in which a predetermined color range was set as "a predetermined or greater color" and "a predetermined or greater brightness and saturation (R, G, B)," but a single color, multiple colors, a predetermined range of colors, etc., may also be set. In addition, multiple predetermined color ranges may be set, for example, "hues corresponding to orange (R, G, B)" as the first predetermined color range, and "hues corresponding to red (R, G, B)" as the second predetermined color range. In this case, it is possible to configure the system so that "when multiple people Z1 are detected in the restricted area R, if the colors of the same body part detected for each of the multiple people Z1 fall within the same color range, it is not determined to be an 'intrusion into the restricted area R'," making it possible to perform intrusion determination more accurately.

[0055] Furthermore, while the first detection unit 4 detected the joints of human Z, it may also detect other key points such as both eyes. In this case, the memory unit 3 will store the criteria for detecting the key points as detection criteria.

[0056] Furthermore, although the above embodiment was explained using an example in which track maintenance workers are excluded from the scope of intrusion, it goes without saying that the present invention can also be used to exclude workers and the like at other work sites.

[0057] Furthermore, the present invention can also be applied to programs and methods corresponding to the processing performed by each component acting as a controller, as well as to recording media that store such programs. In the case of recording media, the program will be installed on a computer or the like. Here, the recording media that stores the program may be a non-transient recording media. Examples of non-transient recording media include CD-ROMs, but the invention is not limited to them. [Explanation of Symbols]

[0058] 1. Intrusion Detection System 2 Acquisition part 3 Storage section 4. First detection unit 5. Second detection unit 6 Judgment section 7 Hochi Department R No entry area R1, R2 No entry area X Photography Method Y Image Z Human

Claims

1. An acquisition unit that acquires an image captured by a shooting means installed to photograph a predetermined range, A storage unit that stores a human intrusion-restricted area set within the predetermined range and detection criteria for detecting at least a portion of a person captured in the image, A first detection unit capable of detecting a person in the image based on the features captured in the image and the detection criteria, A second detection unit capable of detecting the colors of the detected human, A determination unit that determines that an intrusion into the restricted area has occurred when a person is detected within the restricted area, Equipped with, The intrusion determination system is characterized in that, when multiple people are detected in the restricted area, if the color of the same body part detected for each of the multiple people falls within a predetermined color range, the determination unit does not determine that an intrusion into the restricted area has occurred. However, when multiple people whose color of the same body part falls within a predetermined color range and one or more other people whose color of the same body part does not fall within the predetermined color range are detected in the restricted area, the determination unit determines whether or not an intrusion into the restricted area has occurred according to the ratio of the multiple people and the one or more other people.

2. Multiple predetermined color ranges can be set. The intrusion detection system according to claim 1, characterized in that when multiple persons are detected in the restricted area, the determination unit does not determine that an intrusion into the restricted area has occurred if the colors of the same body part detected for each of the multiple persons are included in the same predetermined color range.

3. The intrusion determination system according to claim 1, characterized in that, if there are multiple restricted areas with space between them within a predetermined range captured by the aforementioned imaging means, the determination unit determines intrusion into each of the restricted areas.

4. A program executed on a computer that stores a human-restricted area set within a predetermined range and detection criteria for detecting at least a portion of a person captured in an image, The steps include acquiring an image captured by a shooting means installed to capture the predetermined range, A step of detecting a person in the image based on the features captured in the image and the detection criteria, The step of detecting the color of the human being detected, The steps include determining that an intrusion into the restricted area has occurred if a person is detected within the restricted area, Equipped with, In the step of making the determination, if multiple people are detected in the restricted area, and the colors of the same body part detected for each of the multiple people are included in a predetermined color range, the program does not determine that an intrusion into the restricted area has occurred. However, if multiple people whose body parts are included in a predetermined color range and one or more other people whose body parts are not included in the predetermined color range are detected in the restricted area, the program determines whether an intrusion into the restricted area has occurred according to the ratio of the multiple people and the one or more other people.

5. A method performed on a computer that stores a human-restricted area set within a predetermined range and detection criteria for detecting at least a portion of a person in an image, The steps include acquiring an image captured by a shooting means installed to capture the predetermined range, A step of detecting a person in the image based on the features captured in the image and the detection criteria, The step of detecting the color of the human being detected, The steps include determining that an intrusion into the restricted area has occurred if a person is detected within the restricted area, Equipped with, In the step of making the determination, if multiple people are detected in the restricted area, and the colors of the same body parts detected for each of the multiple people are included in a predetermined color range, it is not determined that an intrusion into the restricted area has occurred. However, if multiple people whose body parts are included in a predetermined color range and one or more other people whose body parts are not included in the predetermined color range are detected in the restricted area, it is determined whether or not an intrusion into the restricted area has occurred according to the ratio of the multiple people and the one or more other people.

Images