Inspection device and inspection method

Abstract

To reduce noise in an inspection device using terahertz waves.SOLUTION: An inspection device according to an embodiment captures visible light images and terahertz images of the same subject, distinguishes between the subject's body and the background using the visible light images, and utilizes images obtained by removing areas identified as background in the visible light images from the terahertz images. According to the embodiment, the use of a noise-reduced composite terahertz image for possession inspection enhances inspection reliability. Furthermore, the time and effort required for inspection are reduced, improving inspection efficiency.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an inspection device and an inspection method for the possessions of a person to be inspected. [Background technology]

[0002] Inspection devices using terahertz waves are known. For example, Patent Document 1 describes an imaging device that irradiates an object with terahertz waves to generate image signals and combines the multiple image signals to generate a composite image. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2023-047487 Summary of the Invention [Problem to be solved by the invention]

[0004] When possessions are hidden under clothing, the electromagnetic waves emitted from the human body are blocked, creating a shadow, and this shadow, reflected in the terahertz image, can be detected using image processing or AI (Artificial Intelligence) to inspect the possessions. However, with conventional technology, noise (e.g., sunlight or lighting) is reflected in areas where the human body is not visible, which can interfere with the inspection.

[0005] The present invention provides a technique for reducing noise in an inspection device that uses terahertz waves. [Means for solving the problem]

[0006] An inspection device according to one embodiment captures a visible light image and a terahertz image of the same object, uses the visible light image to distinguish between the subject's body and the background, and uses an image obtained by removing from the terahertz image the areas identified as the background in the visible light image.

[0007] Furthermore, the background is identified from the difference between an image obtained by cutting out and enlarging a portion of the visible light image and the visible light image itself.

[0008] The part includes the part that was photographed the earliest.

[0009] Furthermore, a striped pattern is applied to a member located at a position corresponding to the background of the sensor that captures the terahertz image.

[0010] Furthermore, the clothing of the person to be inspected is detected before the person enters the inspection device, and a color or pattern corresponding to the clothing is generated on a member located in a position corresponding to the background of the sensor that captures the terahertz image.

[0011] Furthermore, there are a plurality of pillars each provided with a sensor for capturing the terahertz image, and the position corresponding to the background is the surface of the other pillars.

[0012] The visible light image is an image obtained by cutting out one line of images corresponding to the scan range of a sensor that captures the terahertz image from a video captured by a camera and connecting them in chronological order.

[0013] An inspection method according to one embodiment involves taking a visible light image and a terahertz image of the same object, using the visible light image to distinguish between the subject's body and the background, and using an image obtained by removing from the terahertz image the area identified as the background in the visible light image. [Effects of the Invention]

[0014] According to the present invention, noise can be reduced in an inspection device using terahertz waves. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a diagram showing the overall configuration of an inspection apparatus according to an embodiment. [Figure 2] FIG. 2 is a diagram showing the functional configuration of the inspection device. [Figure 3]FIG. 2 is a diagram showing the hardware configuration of the inspection device. [Figure 4] FIG. [Figure 5] 4 is a flow chart showing an image generation process performed by an inspection device. [Figure 6] FIG. 3 is a diagram showing an example of a visible light image captured by a first image capturing means. [Figure 7] FIG. 4 is a diagram showing an example of a terahertz image captured by a second imaging means. [Figure 8] FIG. 10 is a diagram showing an example of a synthetic visible light image. [Figure 9] FIG. 10 is a diagram showing an example of a background image. [Figure 10] FIG. 10 is a diagram showing an example of a synthetic terahertz image that has been processed to extract a human body region. [Figure 11] FIG. 10 is a diagram showing an example of a synthetic terahertz image. [Figure 12] FIG. 10 is a diagram showing an example of a synthetic terahertz image after filtering. [Figure 13] FIG. 10 is a diagram showing the overall configuration of an inspection device according to a modified example. [Figure 14] FIG. 10 is a diagram showing an example of a background image according to a modified example. [Figure 15] FIG. 10 is a diagram showing the overall configuration of an inspection device according to a modified example. [Figure 16] FIG. 10 is a diagram showing the overall configuration of an inspection device according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0016] 1. Configuration 1 is a diagram showing the overall configuration of an inspection device 1 according to one embodiment. The inspection device 1 is a device for inspecting belongings of an inspection subject 2 (hereinafter referred to as "possession inspection"). The inspection device 1 is installed indoors or outdoors in places where safety and security must be ensured, such as airports, train stations, public facilities, schools, concert halls, theaters, exhibitions, convention centers, commercial facilities, and the like.

[0017] The inspection device 1 has a plurality of photographing units 11a1, 11a2, 11b1, and 11b2 and inspection units 12a and 12b (hereinafter, when the photographing units 11 and the inspection units 12 are not particularly distinguished from each other, they will be simply referred to as the "photographing units 11" and the "inspection units 12," respectively). The photographing units 11 are components for photographing the subject 2. The inspection units 12 are components for processing the images photographed by the photographing units 11. The photographing units 11 and the inspection units 12 are connected wirelessly or by wire. In the example of FIG. 1, the number of photographing units 11 is four and the number of inspection units 12 is two, but this number is not limited to these.

[0018] Each of the photographing units 11 and the inspection units 12 is housed in, for example, a housing case. A housing case is a hard exterior or enclosure that houses the device or its components and protects them from causes of malfunction such as physical damage, dust, water droplets, or electromagnetic interference. The housing case is manufactured from a material (e.g., metal or tempered glass) that is durable and reflects the surrounding environment in which it is installed. Furthermore, the multiple photographing units 11 and inspection units 12 (i.e., housing cases) form a passage through which the subject 2 passes to carry out a personal belongings inspection. Note that in the example of FIG. 1, the overall shape of the housing case is a concave shape formed by combining multiple rectangular housing cases, but this shape is not limited to this. For example, a single rectangular housing case may be used as a whole.

[0019] The inspection unit 12 also functions as a computer that provides software (hereinafter referred to as "possession inspection software") related to the inspection device 1 accessed by the user (administrator). The possession inspection software is software that inspects whether the person 2 to be inspected is carrying any dangerous or prohibited items (hereinafter referred to as "dangerous items, etc."). Dangerous items are items that may pose a risk to human health or safety. Examples of dangerous items include weapons (such as handguns or knives) and explosives (such as gunpowder or gasoline). Prohibited items are items that are prohibited from being brought into certain places by laws, regulations, or other rules. Examples of prohibited items include (in addition to dangerous items) food and drink, photography equipment (such as cameras or videos), and communication devices (such as smartphones, tablets, or PCs (Personal Computers)).

[0020] The person to be inspected 2 is a person who passes through the installed inspection device 1 (between the housing cases) and attempts to enter a specific location. In the example of FIG. 1, the person to be inspected 2 passes through the inspection device 1 from left to right along the traveling direction A. Note that the number and arrangement of the photographing units 11, inspection units 12, and housing cases are not limited to this.

[0021] 2 is a diagram showing the functional configuration of the inspection device 1. The inspection device 1 has a plurality of photographing units 11 and an inspection unit 12. The photographing unit 11 has a first photographing means 101 and a second photographing means 102. The inspection unit 12 has a communication means 103, a storage means 104, an image generation means 105, an identification means 106, and a filter processing means 107.

[0022] The first image capturing means 101 captures a moving image of a visible light image showing the subject 2. A visible light image is an image acquired within the range of light that can be directly recognized by the human eye (380 to 740 nm (nanometers)).

[0023] The second imaging means 102 captures a terahertz image of the subject 2. A terahertz image is an image captured using terahertz waves (THz waves). Terahertz waves are electromagnetic waves in a frequency range between light and radio waves, with a frequency of around 1 THz, for example, in the range of approximately 0.1 THz to 10 THz. The human body emits terahertz waves due to blackbody radiation caused by body heat. Terahertz waves can pass through certain materials (especially clothing, paper, or plastic), but not through water or metal. Utilizing this property, terahertz waves are used for non-destructive testing, security checks, and the like.

[0024] The communication unit 103 communicates with other devices. The storage unit 104 stores various information and programs. The storage unit 104 stores image data such as visible light images and terahertz images, for example.

[0025] The image generating means 105 generates various images. For example, the image generating means 105 cuts out a trimmed image (an example of an "image for one line") corresponding to the imaging range (an example of a "scan range") of the second imaging means 102 from the visible light image, and synthesizes the multiple trimmed images in chronological order to generate a composite visible light image. The image generating means 105 also cuts out and stretches a portion of the visible light image to generate a background image. The image generating means 105 also synthesizes terahertz images in chronological order to generate a composite terahertz image.

[0026] The identification means 106 identifies the attributes of what is shown in the image. The attributes are, for example, a human body and a background. The identification means 106 distinguishes the human body of the person being inspected 2 from the background, for example, using a visible light image. The filter processing means 107 generates an image to be used for the possession inspection. The filter processing means 107 performs, for example, a process (hereinafter referred to as "filter processing") to remove areas identified as the background in the visible light image from the terahertz image of the person being inspected 2.

[0027] FIG. 3 is a diagram showing the hardware configuration of the inspection device 1. The photographing unit 11 is a photographing device including a camera 154 and a sensor 155. The camera 154 is, for example, a visible light camera (digital camera). The visible light camera photographs video at a frame rate of, for example, 30 to 60 fps (frames per second). The sensor 155 receives terahertz waves emitted from the human body. The sensor 155 converts the captured terahertz waves into electrical signals based on characteristics such as their intensity and / or phase to generate a terahertz image (digital image). That is, characteristics such as the intensity and / or phase of the electrical signal correspond to the brightness and contrast of the terahertz image. The sensor 155 photographs the subject 2 by, for example, periodically scanning up and down. The sensor 155 photographs the terahertz image at a scanning speed (scanning frequency) of, for example, 10 Hz (Hertz).

[0028] In this example, the camera 154 and the sensor 155 are housed in a columnar housing (not shown in FIG. 3). This housing is an example of a column provided with a sensor that captures terahertz images.

[0029] The inspection unit 12 is a computer including a processor 151, a communication IF (Interface) 152, and a memory 153. The processor 151 controls the operation of other components of the inspection device 1 by executing a predetermined program stored in the memory 153. The communication IF 152 communicates with a communication partner device wirelessly or via a cable. The memory 153 stores a program (hereinafter referred to as an "inspection program") for causing the computer to function as the inspection device 1.

[0030] When the processor 151 is executing the inspection program, the camera 154 is an example of the first photographing means 101, the sensor 155 is an example of the second photographing means 102, the communication IF 152 is an example of the communication means 103, the processor 151 is an example of the image generating means 105, the identification means 106, and the filter processing means 107, and the memory 153 is an example of the storage means 104.

[0031] FIG. 4 is a diagram showing a top view of the inspection device 1. FIG. 4 shows the photographing direction and photographing range R1 of the first photographing means 101 of the photographing unit 11a1, and the photographing direction and photographing range R2 of the second photographing means 102. To simplify the drawing, the photographing directions and photographing ranges of the photographing means of the other photographing units 11a2, 11b1, and 11b2 are not shown. The first photographing means 101 is a camera 154 (e.g., a digital camera), and therefore the viewing angle of its photographing range R1 is, for example, 45 to 60°. The second photographing means 102 is a sensor 155, and therefore the viewing angle of its photographing range R2 is, for example, 1 to 10°.

[0032] In the example of FIG. 4, similar to the example of FIG. 1, the person under inspection 2 passes through the inspection device 1 along the traveling direction A. At this time, the photographing unit 11a1 installed on the right side of the person under inspection 2 in the traveling direction A mainly photographs the right rear side (in the traveling direction A) of the person under inspection 2, and the photographing unit 11a2 mainly photographs the right front side (in the traveling direction A). In addition, the photographing unit 11b1 installed on the left side of the person under inspection 2 in the traveling direction A mainly photographs the left rear side (in the traveling direction A) of the person under inspection 2, and the photographing unit 11b2 mainly photographs the left front side (in the traveling direction A). The position corresponding to the background of the photographed image is the surface of the other photographing units 11. For example, the surface of the photographing unit 11b2 is reflected in the image photographed by the photographing unit 11a1.

[0033] 2. Operation FIG. 5 is a flow diagram showing an image generation process performed by the inspection device 1. The process shown in FIG. 5 is initiated, for example, when a human presence sensor (not shown) installed near the entrance side (upstream side in the traveling direction A) of the inspection device 1 detects that an inspection subject 2 has approached the inspection device 1. In the following, functional elements such as the first photographing means 101 are described as the subject of processing, which means that hardware elements such as the processor 151 executing a program such as an inspection program execute the processing in cooperation with other hardware elements. Note that in the following example, only the processing performed by the photographing unit 11a1 will be described, and descriptions of the processing performed by the other photographing units 11a2, 11b1, and 11b2 will be omitted.

[0034] First, the inspection device 1 captures an image of the person under inspection 2 (step S101). The first imaging unit 101 captures, for example, a visible light image. The second imaging unit 102 captures, for example, a terahertz image. That is, the first imaging unit 101 and the second imaging unit 102 each capture an image of the same person under inspection 2. In this example, the first imaging unit 101 and the second imaging unit 102 continue to capture an image of the person under inspection 2 while the person under inspection 2 passes through the inspection device 1. That is, the first imaging unit 101 and the second imaging unit 102 respectively acquire a visible light image and a terahertz image independently of the processing from step S102 onward. However, for simplicity of the drawing, this is illustrated as one step in a series of steps. Note that, for example, when a human presence sensor (not shown) installed near the exit side (downstream in the traveling direction A) of the inspection device 1 detects that the person under inspection 2 has passed through the inspection device 1, the first imaging unit 101 and the second imaging unit 102 stop capturing images.

[0035] FIG. 6 is a diagram showing an example of a visible light image captured by the first photographing means 101. The visible light image in FIG. 6 is an example of a visible light image captured by the first photographing means 101 of the photographing unit 11a1. In other words, it is an example of a frame (still image) of a moving image captured by the first photographing means 101. Visible light images a1 to f1 in FIG. 6 show the test subject 2 and the photographing unit 11b2 installed diagonally from the subject's device (photographing unit 11a1). The first photographing means 101 captures visible light images a1 to f1. The visible light images a1 to f1 are captured in order of time t.

[0036] Generally, the frame rate of a moving image captured by the first imaging means 101 (camera 154) is faster than the scanning speed of the second imaging means 102 (sensor 155). Therefore, the image generating means 105 extracts frames that match the scanning speed of the second imaging means 102 from among a plurality of frames (still images) in the moving image captured by the first imaging means 101, for example. For example, the number of images captured per second is 30 for the first imaging means 101 (camera 154) that captures moving images at 30 fps, and 10 for the second imaging means 102 (sensor 155) that captures terahertz images at 10 Hz. Therefore, in this example, the number of images captured per second by the first imaging means 101 (camera 154) is three times the number of images captured by the second imaging means 102 (sensor 155). Therefore, in this example, the image generating means 105 extracts one visible light image every three frames from the video captured by the first imaging means 101. Moreover, area C surrounded by a dotted line in Fig. 6 indicates an example of an area cut out by the image generating means 105. The image generating means 105 generates, for example, trimmed images Ca1, Cb1, Cc1, Cd1, Ce1, and Cf1 by cutting out parts of the visible light images a1, b1, c1, d1, e1, and f2, respectively (details will be described later). Note that, although six visible light images are shown here for simplicity, this number is not limited to six.

[0037] FIG. 7 is a diagram showing an example of a terahertz image captured by the second imaging means 102. The terahertz image in FIG. 7 is an example of a terahertz image captured by the second imaging means 102 of the imaging unit 11a1. In the terahertz images a2 to f2 in FIG. 7, the subject 2 and his / her surroundings appear white, and the background appears black. The second imaging means 102 captures the terahertz images a2 to f2. The terahertz images a2 to f2 are captured in order of time t. Note that, although six terahertz images are shown here for simplicity, the number is not limited to six.

[0038] Next, the inspection device 1 generates a composite visible light image (step S102). The image generation means 105 generates a trimmed image by, for example, cutting out a portion of the visible light image. The portion here refers to, for example, a vertical region having a width of a predetermined number of pixels that corresponds to the imaging range of the second imaging means 102 in each visible light image. Area C surrounded by a dotted line in FIG. 6 shows an example of a region trimmed out by the image generation means 105. This width of the number of pixels is determined, for example, taking into consideration the difference between the resolution and angle of view of the visible light image and the resolution of the terahertz image. Furthermore, the image generation means 105 generates a composite visible light image by, for example, combining these trimmed images in order of time t.

[0039] FIG. 8 shows an example of a composite visible light image. The image shown in FIG. 8(a) is an example of trimmed images Ca1 to Cf1 obtained by cutting out portions of a visible light image by the image generating means 105. The image shown in FIG. 8(b) is an example of a composite visible light image generated by the image generating means 105 by combining the trimmed images Ca1 to Cf1. The order in which the image generating means 105 stitches together multiple trimmed images is determined by the position of the photographing unit 11 and its relative position with respect to the subject 2. The visible light images shown in FIGS. 6 and 8 are examples of visible light images captured by the photographing unit 11a1 on the front right in the direction A of travel of the subject 2, so in the visible light image, the subject 2 moves from left to right. Therefore, the image generating means 105 arranges the trimmed images in chronological order of time t, starting from the right end. Similarly, the trimmed images of visible light images captured by the photographing unit 11a2 on the far right in the direction A of travel of the subject 2 are arranged in chronological order of time t, starting from the right end.

[0040] Specifically, the image generating means 105 composites the trimmed images Ca1 to Cf1 in order from the right end at time t. That is, trimmed image Ca1 cut out from visible light image a1 constitutes the right end portion of the composite visible light image, trimmed image Cb1 cut out from visible light image b1 constitutes the portion adjacent to the right end portion of the composite visible light image, and trimmed image Cf1 cut out from visible light image f1 constitutes the left end portion of the composite visible light image (trimmed images Cc1 to Ce1 are not shown).

[0041] On the other hand, although not shown, in the visible light images captured by the photographing units 11b1 and 11b2 on the left side (front left and back left) in the traveling direction A of the person under inspection 2, the person under inspection 2 moves from right to left. Therefore, the image generating means 105 arranges the trimmed images from the left end in chronological order of time t.

[0042] Next, the inspection device 1 generates a background image (step S103). The image generation means 105 generates the background image, for example, by cutting out and stretching a portion of the visible light image. The portion here includes the portion captured earliest. The portion captured earliest is, for example, the visible light image captured first by the first image capture means 101 of one of the inspection subject 2. Specifically, it is visible light image a1 out of visible light images a1 to f1. Furthermore, "stretching" here refers to, for example, duplicating a portion of the visible light image until it has the same length as the width of the composite visible light image.

[0043] FIG. 9 is a diagram showing an example of a background image. The image shown in FIG. 9(a) is an example of an image obtained by image generation means 105 cropping a portion of visible light image a1. The area c surrounded by a dotted line in FIG. 9(a) indicates an example of an area cropped by image generation means 105. The image shown in FIG. 9(b) is an example of a background image generated by image generation means 105. Specifically, image generation means 105 further crops the right edge portion of the cropped visible light image and stretches this cropped image to generate a background image. Alternatively, image generation means 105 may generate a background image by simply stretching a cropped portion of the visible light image. The cropped visible light image shows the housing case of photographing unit 11b2. This housing case reflects the surrounding environment in which it is installed, and therefore the background image changes depending on the surrounding environment in which inspection device 1 is installed. To simplify the drawing, the surrounding environment reflected by the housing case is not shown.

[0044] The inspection device 1 uses the synthetic visible light image to distinguish between the human body of the inspection subject 2 and the background (step S104). The identification means 106, for example, extracts the difference between the synthetic visible light image and the background image, and identifies areas where the difference is large as human body area X, and areas where the difference is small as background area B. To extract this difference, a technique such as background subtraction is used. Specifically, the identification means 106 subtracts pixel values ​​of the background image from pixel values ​​(RGB) of the synthetic visible light image, and identifies pixel areas where the subtracted pixel values ​​are equal to or greater than a predetermined threshold as human body area X, and pixel areas where the subtracted pixel values ​​are less than the predetermined threshold as background area B. Note that the method of extracting the difference is not limited to this.

[0045] 10 is a diagram showing an example of a synthetic visible light image that has been processed to extract a human body region X. In FIG. 10, the blackened area indicates an example of the human body region X, and the shaded area indicates an example of the background region B.

[0046] Next, the inspection device 1 generates a composite terahertz image (step S105). The image generating means 105 generates a composite terahertz image by, for example, combining the terahertz images a2 to f2 in order of time t.

[0047] FIG. 11 is a diagram showing an example of a composite terahertz image. The image shown in FIG. 11(a) is an example of a terahertz image captured by the second imaging unit 102. The image shown in FIG. 11(b) is an example of a composite visible light image generated by the image generating unit 105. The order in which the image generating unit 105 stitches together multiple terahertz images is determined by the position of the imaging unit 11 and its relative position with respect to the subject 2. The terahertz images shown in FIGS. 7 and 11 are examples of terahertz images captured by the imaging unit 11a1 on the front right in the direction A of travel of the subject 2, so in the terahertz images, the subject 2 moves from left to right. Therefore, the image generating unit 105 arranges the terahertz images from the right end in chronological order of time t. Similarly, the terahertz images of visible light images captured by the imaging unit 11a2 on the far right in the direction A of travel of the subject 2 are arranged from the right end in chronological order of time t.

[0048] Specifically, the image generating means 105 composites the terahertz images from the right in the order of time t. That is, terahertz image a2 constitutes the right end portion of the composite terahertz image, terahertz image b2 constitutes the portion adjacent to the right end portion of the composite terahertz image, and terahertz image f2 constitutes the left end portion of the composite terahertz image (description of terahertz images c2 to e2 is omitted).

[0049] On the other hand, although not shown, in the terahertz images captured by the imaging units 11b1 and 11b2 on the left side (front left and back left) in the traveling direction A of the person under examination 2, the person under examination 2 moves from right to left. Therefore, the image generating means 105 arranges the terahertz images from the left end in chronological order of time t.

[0050] Next, the inspection device 1 performs a filtering process (step S106). The filtering means 107 rewrites, for example, pixel values ​​in a region identified as background region B in the synthetic visible light image from the synthetic terahertz image to a predetermined value (for example, a value corresponding to black).

[0051] 12 is a diagram showing an example of a composite terahertz image after filtering. Specifically, filtering means 107 performs processing on the composite terahertz image to emphasize the human body region X identified in step S104, and also performs processing to rewrite pixel values ​​in the region identified as the background region B to a predetermined value (e.g., a value corresponding to black). The emphasis processing is, for example, processing to emphasize (expand) brightness differences.

[0052] The image shown in FIG. 12(a) is an example of an image in which the person being inspected 2 is not carrying any belongings. Specifically, this is an image in which the person being inspected 2 has nothing (regardless of whether it is a dangerous object or the like) in the pockets or other storage spaces of his / her clothes. As shown in the image of FIG. 12(a), in this example, the entire human body region X is displayed in white. On the other hand, the image shown in FIG. 12(b) is an example of an image in which the person being inspected 2 is carrying any belongings. Specifically, this is an image in which the person being inspected 2 has something (for example, a knife K, which is a dangerous object) in the pockets or other storage spaces of his / her clothes. As shown in the image of FIG. 12(b), in this example, the part of the human body region X that is displayed in white and that contains the belongings is displayed in black.

[0053] When the process of step S106 is completed, the inspection device 1 performs an inspection of the belongings using the filtered composite terahertz image. For simplicity of the drawing, the flow is shown as ending at step S106.

[0054] According to this embodiment, the noise-removed synthetic terahertz image is used for the inspection of personal belongings, thereby improving the reliability of the inspection. In addition, the time and labor required for the inspection are reduced, thereby improving the efficiency of the inspection.

[0055] 3. Variations The present invention is not limited to the above-described embodiment, and various modifications are possible. Two or more of the features described in the following modifications may be combined.

[0056] (1) Photo shoot unit 11 In the above embodiment, the image cut out from a portion of the visible light image shows the housing case of the photographing unit 11, and the background image is an image that changes depending on the surrounding environment in which the inspection device 1 is installed, but this is not limiting. For example, (another) photographing unit 11 (an example of a "component") located in a position corresponding to the background of the image photographed by the photographing unit 11 may be given a pattern that is highly distinguishable from the subject's clothing. Examples of highly distinguishable patterns include stripes, borders, checks, lattices, herringbone, and other designs.

[0057] Fig. 13 is a diagram showing the overall configuration of an inspection device 1 (photographing unit 11) according to a modified example. A pattern is applied to the surface of the photographing unit 11 in Fig. 13. Fig. 14 is a diagram showing an example of a background image according to a modified example. The pattern appears in the background image generated by the image generating means 105, making it easier for the identifying means 106 to extract the difference between the clothing of the person being inspected 2 and the background image.

[0058] (2) Color changes depending on the subject's clothing The inspection device 1 may further include clothing detection means 108. The clothing detection means 108 detects the clothing of the inspection subject 2 before the inspection subject 2 enters the inspection device 1, for example. The inspection device 1 may further include display means 109. The display means 109 is a display installed in the photographing unit 11 (housing case). The display means 109 is installed, for example, on the surface of the photographing unit 11 inside the passage formed by the housing case (i.e., the surface photographed by the photographing unit 11). The display means 109, for example, causes a color, a pattern, or both corresponding to the clothing detected by the clothing detection means 108 to appear on (another) photographing unit 11 located in a position corresponding to the background of the image photographed by the photographing unit 11 (an example of a "component located in a position corresponding to the background of a sensor that photographs a terahertz image").

[0059] A color corresponding to the clothing is, for example, a complementary color of the color of the detected clothing. Complementary colors are colors that are opposite each other on the color wheel. For example, the complementary color of "red" is "green," the complementary color of "blue" is "orange," and the complementary color of "yellow" is "purple." Furthermore, a pattern corresponding to the clothing is, for example, a pattern that contrasts with the pattern of the detected clothing. A contrasting pattern is, for example, a combination of "vertical stripes" and "horizontal stripes," or "solid color" and "checked."

[0060] FIG. 15 is a diagram showing the overall configuration of an inspection device 1 according to a modified example. In the example of FIG. 15, the clothing detection means 108 is installed in the photographing units 11a1 and 11b1 on the upstream side of the traveling direction A, adjacent to and further upstream from these photographing units 11. Note that the location at which the clothing detection means 108 is installed is not limited to this. For example, the clothing detection means 108 may be built into the photographing units 11, or the first photographing means 101 may also have the function of the clothing detection means 108, or the clothing detection means 108 may be installed independently of the photographing units 11. In this modified example, the image generation means 105 generates a background image that displays a color, pattern, or both corresponding to the clothing detected by the clothing detection means 108. Therefore, a strong contrast is created by combining complementary colors, contrasting patterns, or both between the clothing (i.e., the person) and the background, which makes it easier for the identification means 106 to extract differences between the clothing of the subject 2 and the background image.

[0061] (3) Changes in lighting depending on the subject's clothing The inspection device 1 may include an illumination means 110 instead of the display means 109. The illumination means 110 is a lighting fixture that can freely change the hue of light. The illumination means 110 is installed, for example, on the surface of the inspection device 1. For example, when the illumination means 110 emits light, the surface of (another) photographing unit 11 located in a position corresponding to the background of the image photographed by the photographing unit 11 reflects the light of the illumination means 110, thereby generating a color corresponding to the clothing detected by the clothing detection means 108.

[0062] FIG. 16 is a diagram showing the overall configuration of an inspection device 1 according to a modified example. In the example of FIG. 16, a total of four lighting units 110 are installed on the top surface of the inspection device 1 (two on the top surface of each inspection unit 12). Note that the number of lighting units 110 and the locations at which the lighting units 110 are installed are not limited to this. Furthermore, the lighting units 110 may be installed anywhere, such as on the ground, a wall, a ceiling, or another surface of the inspection device 1, as long as the surface of the photographing unit 11 reflects the light from the lighting units 110. In this modified example, the image generating unit 105 generates a background image in which a color corresponding to the clothing detected by the clothing detection unit 108 is displayed. Therefore, a strong contrast is created by combining complementary colors between the clothing (i.e., the person) and the background, which makes it easier for the identification unit 106 to extract the difference between the clothing of the inspection subject 2 and the background image.

[0063] (4) Detection of approach of subject 2 In the above embodiment, the human presence sensor (not shown) detects that the subject 2 approaches the inspection device 1, but this is not limiting. For example, the clothing detection means 108 may detect that the subject 2 approaches the inspection device 1.

[0064] (5) Synchronization of the first imaging means 101 and the second imaging means 102 In the above embodiment, the image generating means 105 extracts frames that match the scanning speed of the second imaging means 102 from among a plurality of frames (visible light images) in a moving image captured by the first imaging means 101, but this is not limiting. For example, the first imaging means 101 may capture still images. Furthermore, the first imaging means 101 and the second imaging means 102 may be synchronized to capture visible light images and terahertz images, respectively.

[0065] (6) Emphasis on possessions In the above embodiment, when the subject 2 is carrying an item, the part of the human body region X that is displayed in white is displayed in black, but this is not limiting. For example, the filter processing means 107 may display the part of the item in a color (e.g., red) different from the color of the human body region X and the background region B in the composite terahertz image after filtering.

[0066] (7) Identification of possessions The inspection device 1 may further include a determination means (not shown). The determination means is an AI related to image recognition, such as a trained AI. The determination means, for example, determines whether or not a possession shown in the composite terahertz image after filtering is a dangerous object or the like. If the determination means determines that the possession is a dangerous object or the like, the filter processing means 107 may display the possession in a color (e.g., red) different from the color of the human body region X and the background region B. On the other hand, if the determination means determines that the possession is not a dangerous object or the like, the filter processing means 107 may display the possession in a color (e.g., blue) different from the color of the human body region X, the background region B, and the dangerous object or the like.

[0067] (8) 3D images In the above embodiment, an image (two-dimensional image) such as a composite terahertz image is generated using the image captured by the photographing unit 11, but the present invention is not limited to this. For example, the image generating means 105 may use a plurality of images of the subject 2 captured by the photographing unit 11 to generate an image (3D modeling) such as a composite terahertz image of a three-dimensional image representing the subject 2 in all directions (360° directions). For example, techniques such as stereo vision or photogrammetry may be used to generate the three-dimensional image.

[0068] (9) Startup timing Some of the devices (for example, the inspection unit 12) of the inspection device 1 may be always activated, or may be activated automatically when a human sensor or clothing detection means 108 detects that the person under inspection 2 has approached the inspection device 1. Alternatively, the photographing unit 11 and the inspection unit 12 may be activated automatically when the clothing detection means 108 detects that the person under inspection 2 has approached the inspection device 1.

[0069] (10) Stop timing Some of the devices (e.g., the inspection unit 12) of the inspection device 1 may be constantly running, or may automatically stop (transition to a sleep state) if the human presence sensor or clothing detection means 108 does not detect that the inspection subject 2 has approached the inspection device 1 for a predetermined time (e.g., 5 minutes). Alternatively, the photographing unit 11 and the inspection unit 12 may automatically stop (transition to a sleep state) if the clothing detection means 108 does not detect that the inspection subject 2 has approached the inspection device 1 for a predetermined time (e.g., 5 minutes).

[0070] (11) Other The correspondence between the functions and hardware in the inspection device 1 is not limited to that exemplified in the embodiment. For example, multiple physical devices may cooperate to have the functions of the inspection device 1. Furthermore, some of the functions of the inspection device 1 exemplified in the embodiment may be omitted.

[0071] The hardware configuration of the inspection device 1 is not limited to that exemplified in the embodiment. For example, the shape of the housing of the photographing unit 11 is not limited to a cylindrical shape and may be any shape, such as a spherical or polyhedral shape. Furthermore, the hardware configuration of the inspection unit 12 may differ from that exemplified in the embodiment. For example, the correspondence between the photographing units 11 and the inspection units 12 is not limited to that exemplified in the embodiment. In the embodiment, one inspection unit 12 processes images from two photographing units 11, but one inspection unit 12 may process images from all four photographing units 11. Alternatively, the inspection units 12 may correspond one-to-one to the photographing units 11 and be housed in the same housing. Alternatively, the inspection units 12 may be a server on a computer network. This server may be a physical server or a virtual server (so-called cloud).

[0072] The program executed by processor 151 may be provided in a state recorded on a computer-readable recording medium such as a DVD-ROM, or may be provided by downloading via a network such as the Internet. [Explanation of symbols]

[0073] 1...inspection device, 11...photographing unit, 12...inspection unit, 101...first photographing means, 102...second photographing means, 103...communication means, 104...storage means, 105...image generation means, 106...identification means, 107...filter processing means, 108...clothing detection means, 109...display means, 110...illumination means, 2...inspection subject

Claims

1. An inspection device that takes a visible light image and a terahertz image of the same object, uses the visible light image to distinguish between the subject's body and the background, and uses an image obtained by removing from the terahertz image the areas identified as the background in the visible light image.

2. The background is identified based on the difference between a part of the visible light image that has been cut out and enlarged and the visible light image. The inspection device according to claim 1 .

3. The part includes the part that was photographed the earliest. The inspection device according to claim 2 .

4. A striped pattern is applied to a member located at a position corresponding to the background of the sensor that captures the terahertz image. The inspection device according to claim 1 .

5. The clothing of the person to be inspected is detected before the person enters the inspection device, and a color or pattern corresponding to the clothing is generated on a member located at a position corresponding to the background of a sensor that captures the terahertz image. The inspection device according to claim 1 .

6. The system has a plurality of pillars each provided with a sensor for capturing the terahertz image, and the position corresponding to the background is the surface of another pillar.

6. The inspection device according to claim 4 or 5.

7. The visible light image is an image obtained by cutting out one line of an image corresponding to the scan range of a sensor that captures the terahertz image from a video captured by a camera and connecting them in chronological order. The inspection device according to claim 1 .

8. An inspection method in which a visible light image and a terahertz image are taken of the same object, the visible light image is used to distinguish between the subject's body and the background, and the area identified as the background in the visible light image is removed from the terahertz image to use an image.

Images