Training information generating device, learning device, detecting device, and training information generating method

Abstract

Provided is a training information generating device that automatically collects training information and enables the load of training information acquisition to be lightened. A training information generating device according to the present disclosure comprises: an image receiving unit (1) that receives monitoring images of a monitored area; a signal acquiring unit (2) that acquires a baseline signal indicating that the monitored area is in a baseline state; and an image acquiring unit (3) that, when the signal acquiring unit has acquired the baseline signal, acquires a monitoring image as a training image, said monitoring image being an image from when the signal acquiring unit acquired the baseline signal among the monitoring images received by the image receiving unit.

Description

Learning information generating device, learning device, detection device, and learning information generating method

[0001] The present disclosure relates to a training information generation device, a learning device, a detection device, and a training information generation method.

[0002] Patent Document 1 discloses an intruder detection system that can detect whether an abnormal state has occurred even when the background image changes.

[0003] Japanese Patent Application Laid-Open No. 2023-168733

[0004] The intruder detection system described in Patent Document 1 has multiple trained models corresponding to small areas, seasons, and time periods into which surveillance video is divided, and therefore has the problem of requiring a huge amount of training data.

[0005] Therefore, an object of the present disclosure is to reduce the burden of acquiring learning information.

[0006] The learning information generation device of the present disclosure includes an image receiving unit that receives surveillance images of a monitored area, a signal acquiring unit that acquires a reference signal indicating that the monitored area is in a reference state, and an image acquiring unit that, upon the signal acquiring unit acquiring the reference signal, acquires, as a learning image, from the surveillance images received by the image receiving unit, the surveillance image at the time when the signal acquiring unit acquired the reference signal.

[0007] According to the present disclosure, the load of acquiring learning information can be reduced by automatically collecting learning information using a reference signal indicating that the monitoring area is in a reference state as a trigger for acquiring learning information.

[0008] FIG. 1 is a functional configuration diagram of a training information generation device and a learning device according to embodiment 1. FIG. 2 is an example of a monitoring area according to embodiment 1. FIG. 3 is an example of input and output of a learning model including an autoencoder according to embodiment 1. FIG. 4 is a functional configuration diagram of a detection device according to embodiment 1. FIG. 5 is a hardware configuration diagram of a training information generation device, a learning device, and a detection device according to embodiment 1. FIG. 6 is a flowchart showing an example of processing performed by the training information generation device and the learning device according to embodiment 1. FIG. 7 is a flowchart showing an example of processing performed by the detection device according to embodiment 1. FIG. 8 is a modified example of a training information generation device and a learning device according to embodiment 1. FIG. 9 is a modified example of a detection device according to embodiment 1. FIG. 10 is a modified example of a monitoring area and an opening / closing device according to embodiment 1. FIG. 11 is a functional configuration diagram of a training information generation device and a learning device according to embodiment 2. FIG. 12 is a hardware configuration diagram of a training information generation device, a learning device, and a detection device according to embodiment 2. FIG. 13 is a flowchart showing an example of processing performed by the training information generation device and the learning device according to embodiment 2. FIG. 14 is a flowchart showing an example of processing performed by the detection device according to embodiment 2. FIG. 15 is a modified example of a mask area according to embodiment 2. Fig. 1 is a functional configuration diagram of a learning information generation device and a learning device according to embodiment 3. Fig. 2 is an example of a mask region according to embodiment 3. Fig. 3 is a hardware configuration diagram of a learning information generation device, a learning device, and a detection device according to embodiment 3. Fig. 4 is a flowchart showing an example of processing performed by the learning information generation device and the learning device according to embodiment 3.

[0009] Embodiment 1. A learning information generation device 100, a learning device 101, and a detection device 102 will be described with reference to FIGS. 1 to 10. In the embodiments and drawings, the same or corresponding elements are designated by the same reference numerals. Descriptions of elements designated by the same reference numerals as previously described elements will be omitted or simplified as appropriate. Arrows in the figures primarily indicate the flow of data or the flow of processing.

[0010] ***Outline of First Embodiment*** <<<During Learning>>> <<Functions of the Learning Information Generation Device 100>> First, a functional overview of the learning information generation device 100 of this embodiment will be described. The learning information generation device 100 has a function of acquiring a reference signal indicating that the monitoring area 40 is in a reference state in order to acquire an image of the monitoring area 40 in a reference state as the learning image 51. When the monitoring area 40 is the area between the platform doors 30a installed on the station platform 41a and the tracks 41b, the learning information generation device 100 has a function of acquiring a signal including an open / closed state from the platform doors 30a in order to acquire an image of the platform doors 30a closed (a state in which no people or objects are present) as learning information. When the learning information generation device 100 receives a signal indicating that the platform doors 30a are closed (hereinafter referred to as the closed state), it acquires camera footage as the learning image 51. This makes it possible to collect images in a state in which no people or objects are present. It is also possible to reduce the burden of collecting learning information.

[0011] <<Configuration of Learning Information Generation Device 100>> In order to perform the above functions, the learning information generation device 100 includes an image receiving unit 1, a signal acquiring unit 2, and an image acquiring unit 3. The image receiving unit 1 acquires video output by a monitoring device 20 such as a camera. The signal acquiring unit 2 acquires a status signal 60 (opening / closing signal) output by an opening / closing device 30 such as platform doors 30a. When the opening / closing device 30 is in the reference state (closed state), the signal acquiring unit 2 outputs an acquisition signal 61 to the image acquiring unit 3. The image acquiring unit 3 acquires the video when the acquisition signal 61 is input from the signal acquiring unit 2 as a learning image 51.

[0012] <<Configuration of Learning Device 101>> The learning device 101 includes a learning information generation device 100 and a learning unit 4. The learning unit 4 uses learning images 51 to train a learning model 5 (obstacle detection artificial intelligence). Note that artificial intelligence will be referred to as AI hereinafter. The learning model 5 (obstacle detection AI) is an artificial intelligence that receives an input image and trains it so that the difference between the input image and the output image becomes small.

[0013] <<<During Analysis>>> <<Functions of Detection Device 102>> The detection device 102 of this embodiment will now be functionally outlined. During analysis, a surveillance image 50 is input to a learning model 5 (obstacle detection AI), the output image is compared with the original video by the analysis unit 11, and if the difference region is abnormal relative to the threshold, the difference region is detected as an obstacle 43. This disclosure does not mention how platform doors or trains are controlled based on the detection result 71.

[0014] <<Configuration of Detection Device 102>> To perform the above functions, the detection device 102 includes an inference unit 10 that uses a learning model 5 (obstacle detection AI) and an analysis unit 11. The inference unit 10 performs processing to detect an obstacle 43 using the learning model 5 (obstacle detection AI). The learning model 5 (obstacle detection AI) is an AI that receives an input image and is trained to minimize the difference between the input image and the output image. During learning, images that do not show the obstacle 43 are learned, so if the obstacle 43 is shown in the input image during analysis, a difference will occur between the input image and the output image. The analysis unit 11 calculates the difference between the input image and the output image, and if there is a difference, outputs a detection result 71 indicating an obstacle 43.

[0015] <<<Definition of Terms>>> <<Definition of Monitoring Area 40>> The monitoring area 40 refers to an area in which an unspecified number of moving objects move. The monitoring area 40 refers to an area in which people or moving objects come and go (pass or get on and off).

[0016] <<Definition of Opening / Closing Device 30>> The opening / closing device 30 is a movement prohibition device that prohibits a moving object from moving within the monitoring area 40. The opening / closing device 30 is a traffic prohibition device that prohibits people or moving objects from passing through.

[0017] <<Definition of Obstacle 43>> Obstacle 43 refers to an object that may impede the movement of moving objects moving through the monitoring area 40. Obstacle 43 includes lost items, forgotten items, suspicious objects, waste, and the like.

[0018] <<Definition of Reference State>> The reference state refers to a state in which no moving object is present in the monitoring area 40. The reference state refers to a state in which no moving object is moving in the monitoring area 40.

[0019] ***Description of Configuration*** The configuration of a learning device 101 equipped with the learning information generation device 100 will be described. <<<Functional Configuration of the Learning Information Generation Device 100>>> First, the learning information generation device 100 will be described based on FIG. 1. The learning information generation device 100 includes an image receiving unit 1, a signal acquisition unit 2, and an image acquisition unit 3. The learning information generation device 100 also includes a signal storage unit (not shown). The signal storage unit stores information for identifying a reference signal in advance as reference signal identification information. Furthermore, the signal storage unit may store reference signal identification information corresponding to a plurality of reference signals.

[0020] <<Image Receiving Unit 1>> The image receiving unit 1 receives a monitoring image 50 of the monitoring area 40. The image receiving unit 1 receives the monitoring image 50 from the monitoring device 20. The image receiving unit 1 receives, as the monitoring image 50 of the monitoring area 40, an image of the area around the opening and closing device 30 that blocks the movement of people or moving objects in the monitoring area 40. In the present embodiment, the image receiving unit 1 receives the monitoring image 50 of the monitoring area 40, which is an area including a station platform 41a on which platform doors 30a are installed.

[0021] The image receiving unit 1 receives the monitoring images 50 at a predetermined time interval. The predetermined time interval can be set arbitrarily. The image receiving unit 1 may receive the monitoring images 50 within a predetermined time. The predetermined time is a time period during which monitoring is determined to be necessary, and can be set arbitrarily.

[0022] The monitoring area 40 in this embodiment will be specifically described based on the arrangement of the platform 41a, platform doors 30a, monitoring device 20, and tracks 41b shown in Figure 2. The monitoring area 40 is the platform 41a and a three-dimensional area above the platform 41a, sandwiched between the platform doors 30a and the tracks 41b. The monitoring device 20 is installed at the platform doors 30a so that it can photograph the monitoring area 40.

[0023] The monitoring device 20 captures a monitoring image 50 that includes the monitoring area 40. The monitoring device 20 is a two-dimensional optical video camera that captures moving images. The monitoring device 20 is installed on the platform door 30a so that the angle of view and the shooting angle are fixed.

[0024] The monitoring image 50 is an image or video captured by the monitoring device 20. The monitoring image 50 is an image that includes the monitoring area 40. The monitoring image 50 may be an image that includes the monitoring area 40, as well as a background 42 including the tracks 41b and platform doors 30a. In this embodiment, the monitoring image 50 is assumed to be an image that includes the platform 41a, platform doors 30a, and background 42.

[0025] <<Signal Acquisition Unit 2>> As shown in Fig. 1 , the signal acquisition unit 2 acquires a reference signal indicating that the monitoring area 40 is in a reference state. The reference state refers to a state in which the monitoring area 40 is normal. In this embodiment, the reference state is a state in which no people or objects have entered the monitoring area 40. In this embodiment, the closed state of the platform doors 30a is regarded as the reference state, in which no people or objects have entered the monitoring area 40.

[0026] The signal acquisition unit 2 acquires, as a reference signal, a signal indicating the closed state of the opening and closing device 30. In this embodiment, the signal acquisition unit 2 acquires, as a reference signal, a signal indicating the closed state of the platform door 30a installed on the platform 41a. The signal acquisition unit 2 has a state signal acquisition unit 2a and a signal determination unit 2b.

[0027] The status signal acquisition unit 2a acquires a status signal 60 indicating the status of the monitoring area 40. The status signal 60 includes a reference signal indicating that the monitoring area 40 is in a reference state. The status signal 60 may be a signal indicating that a person or object has entered the monitoring area 40. A specific example in this embodiment is a signal indicating that the platform doors 30a are open (hereinafter referred to as the open state). The status signal 60 may also be a signal indicating an abnormality in the monitoring area 40. A specific example in this embodiment is a signal transmitted when an emergency button provided on the platform doors 30a is pressed.

[0028] The signal determination unit 2b determines whether the status signal 60 acquired by the status signal acquisition unit 2a is a reference signal. The signal determination unit 2b determines whether the status signal 60 acquired by the status signal acquisition unit 2a is a reference signal based on whether the status signal 60 includes reference signal identification information stored in the signal storage unit. If the signal determination unit 2b determines that the status signal 60 is a reference signal, it transmits an acquisition signal 61 to the image acquisition unit 3. The acquisition signal 61 is a signal that prompts the image acquisition unit 3 to acquire a learning image 51.

[0029] <<Image Acquisition Unit 3>> When the signal acquisition unit 2 acquires the reference signal, the image acquisition unit 3 acquires the learning image 51. Specifically, the image acquisition unit 3 receives an acquisition signal 61 from the signal determination unit 2b. When the image acquisition unit 3 receives the acquisition signal 61, the image acquisition unit 3 acquires, as the learning image 51, the monitoring image 50 received by the image receiving unit 1, which is the monitoring image 50 when the signal acquisition unit 2 acquired the reference signal. When the signal determination unit 2b determines that the signal acquisition unit 2 has acquired the reference signal, the image acquisition unit 3 acquires the learning image 51.

[0030] <<<Functional Configuration of Learning Device 101>>> Next, the configuration of the learning device 101 will be described with reference to Fig. 1. The learning device 101 includes a learning information generation device 100 and a learning unit 4.

[0031] <<Learning Unit 4>> The learning unit 4 trains the learning model 5 (obstacle detection AI) based on the learning images 51 acquired by the learning information generation device 100. That is, the learning unit 4 uses the learning images 51 to train the learning model 5 about the reference state of the monitoring area 40, and creates a trained learning model 5.

[0032] The learning model 5 is a model for inferring whether or not the monitoring area 40 is in a reference state. The learning model 5 is a model that includes an autoencoder 5a.

[0033] When an input image is input, the autoencoder 5a extracts features of the input image and outputs an output result 70 that reconstructs the input image based on the features. The autoencoder 5a learns so that the output result 70 approaches the input image.

[0034] An overview of the autoencoder 5a used in this embodiment will be described using (a) and (b) of Figure 3. The autoencoder 5a is preferably a convolutional autoencoder using a convolutional neural network (hereinafter referred to as CNN). As shown in (a) of Figure 3, the learning unit 4 trains the autoencoder 5a using multiple normal images (training images 51 without obstacles). The normal images are used as input and output for the autoencoder 5a. The autoencoder 5a includes an encoder and a decoder. The encoder extracts features by encoding the input image. The extracted features are provided to the decoder. The decoder decodes the extracted features to generate an output result 70, which is an output image. The autoencoder 5a trains to minimize the difference between the input image and the output result 70. As shown in (b) of Figure 3, in the autoencoder 5a trained with normal images, if an object (obstacle 43) is present in the input image, the object will disappear or change shape in the output result 70. The difference between the input image of the autoencoder 5a and the output result 70 becomes large, making it possible to detect the presence or absence of an obstacle.

[0035] The learning of the autoencoder 5a in the learning unit 4 will be described using (a) of Figure 3. The learning unit 4 inputs a learning image 51 to the encoder as an input image. The learning image 51 is assumed to be an image including a platform 41a, platform doors 30a, and a background 42. The encoder extracts features by encoding the learning image 51. The decoder decodes the features extracted by the encoder to output an output result 70 that restores the learning image 51. The autoencoder 5a learns the reference state in the monitoring image 50 by learning to reduce the difference between the learning image 51, which is the input image, and the output result 70.

[0036] <<<Functional Configuration of the Detection Device 102>>> Next, with reference to Fig. 4, a description will be given of the configuration of the detection device 102 that uses the learning model 5 trained by the learning device 101. The detection device 102 includes an image receiving unit 1, an inference unit 10, and an analysis unit 11. It is desirable for the detection device 102 to detect an obstacle 43 when the platform doors 30a are in a closed state. Therefore, the detection device 102 in this embodiment includes an image acquisition unit 3 and a signal acquisition unit 2. The detection device 102 also includes a signal storage unit (not shown). The signal storage unit corresponds to the signal storage unit included in the learning information generation device 100, and operates in the same manner as the signal storage unit included in the learning information generation device 100.

[0037] <<Image Receiving Unit 1>> The image receiving unit 1 corresponds to the image receiving unit 1 included in the learning information generation device 100. It is desirable that the monitoring images 50 received by the image receiving unit 1 are images captured in the same monitoring area 40 as the learning images 51 used to train the learning model 5. It is desirable that the monitoring images 50 received by the image receiving unit 1 are images captured with the same angle of view and shooting angle as the learning images 51 used to train the learning model 5.

[0038] <<Signal Acquisition Unit 2>> The signal acquisition unit 2 corresponds to the signal acquisition unit 2 provided in the learning information generation device 100, and operates in the same manner as the signal acquisition unit 2 provided in the learning information generation device 100. <<Image Acquisition Unit 3>> The image acquisition unit 3 receives an acquisition signal 61 from the signal acquisition unit 2 and acquires, from the monitoring images 50 received by the image receiving unit 1, the monitoring image 50 when the signal acquisition unit 2 acquired the reference signal.

[0039] <<Inference Unit 10 >> The inference unit 10 inputs the surveillance image 50 acquired by the image acquisition unit 3 to the learning model 5 including the autoencoder 5 a that has been trained by the learning device 101 , and outputs the output result 70 .

[0040] 3B, the input and output of the autoencoder 5a using the monitoring image 50 in the inference unit 10 will be described. Here, as an example, a monitoring image 50 including an obstacle 43 in the monitoring area 40 will be described. The inference unit 10 inputs the monitoring image 50 to the encoder as an input image. The encoder extracts features by encoding the monitoring image 50. The decoder decodes the features extracted by the encoder to output an output result 70 that reconstructs the monitoring image 50. However, because the encoder has learned to extract features necessary for reconstructing the training image 51, the extraction of features for reconstructing the monitoring image 50 is incomplete. In other words, the features of the obstacle 43 cannot be extracted. Furthermore, because the decoder reconstructs the monitoring image 50 from the incompletely extracted features, an output result 70 that does not reflect the features of the obstacle 43 is output. In other words, a difference occurs between the input image, the monitoring image 50, and the output image, the output result 70.

[0041] <<Analysis Unit 11>> The analysis unit 11 analyzes whether or not the monitoring area 40 is in a reference state, based on the output result 70 output by the inference unit 10. The analysis unit 11 detects a difference between the output result 70 output by the inference unit 10 and the monitoring image 50 input to the inference unit 10. The analysis unit 11 analyzes whether or not the monitoring area 40 is in a reference state, based on the detected difference. Specifically, if the detected difference is equal to or greater than a threshold, it is determined that an obstacle 43 is present in the monitoring area 40. The analysis unit 11 outputs the analysis result as a detection result 71.

[0042] <<<Hardware Configuration of the Training Information Generation Device 100, the Learning Device 101, and the Detection Device 102>>> The hardware configurations of the training information generation device 100, the learning device 101, and the detection device 102 will be described with reference to Figure 5. The training information generation device 100, the learning device 101, and the detection device 102 are connected to the monitoring device 20 and the opening and closing device 30. Each of the training information generation device 100, the learning device 101, and the detection device 102 includes a processor 90, a storage device 91, a communication device 92, and an input / output interface 93. The communication device 92 and the input / output interface 93 acquire a monitoring image 50 from the monitoring device 20 and a status signal 60 from the opening and closing device 30. The storage device 91 stores programs for realizing the functions of the image receiving unit 1, the signal acquiring unit 2, the image acquiring unit 3, the learning unit 4, the inference unit 10, and the analysis unit 11, as well as a learning model 5. The processor 90 is a CPU (Central Processing Unit) that executes arithmetic processing based on a program stored in the storage device 91 .

[0043] ***Explanation of Operation*** <<<Learning Information Generation Device 100, Learning Device 101>>> An example of processing performed by the learning information generation device 100 and the learning device 101 will be described with reference to the flowchart shown in FIG.

[0044] <<Step S101>> The image receiving unit 1 receives the monitoring image 50. The image receiving unit 1 receives the monitoring image 50 of the monitoring area 40 from the monitoring device 20.

[0045] <<Step S102>> The signal acquisition unit 2 acquires a status signal 60. Alternatively, the signal acquisition unit 2 acquires a reference signal indicating that the monitoring area 40 is in a reference state. Specifically, the status signal acquisition unit 2a acquires a status signal 60 indicating the state of the monitoring area 40 from the opening and closing device 30. <<Step S103>> The signal determination unit 2b then determines whether the status signal 60 acquired by the status signal acquisition unit 2a is a reference signal. In this embodiment, it determines whether the status signal 60 is a signal indicating that the platform doors 30a are in a closed state. If the signal determination unit 2b determines that the status signal 60 is a reference signal, it transmits an acquisition signal 61 to the image acquisition unit 3 and proceeds to step S104. If it is determined that the status signal 60 is not a reference signal, it returns to step S101.

[0046] <<Step S104>> The image acquisition unit 3 acquires the learning image 51. In response to the signal acquisition unit 2 acquiring the reference signal, the image acquisition unit 3 acquires, from among the monitoring images 50, the monitoring image 50 at the time the reference signal was acquired, as the learning image 51. The image acquisition unit 3 acquires the learning image 51 by receiving the acquisition signal 61 transmitted from the signal determination unit 2b.

[0047] <<Step S105>> The learning unit 4 trains the learning model 5 using the training images 51. The learning unit 4 trains the learning model for inferring whether or not the monitoring area 40 is in a reference state, based on the training images 51 acquired by the image acquisition unit 3 in step S104. Thereafter, the process returns to step S101.

[0048] <<<Detection Device 102 >>> An example of processing performed by the detection device 102 will be described with reference to the flowchart shown in FIG.

[0049] <<Steps S101 to S103>> In step S101, the image receiving unit 1 performs the same operation as in step S101 in the learning information generation device 100 shown in Fig. 6. In step S102, the signal acquiring unit 2 performs the same operation as in step S102 in the learning information generation device 100 shown in Fig. 6. In step S103, the signal determining unit 2b performs the same operation as in step S103 in the learning information generation device 100 shown in Fig. 6.

[0050] <<Step S104>> In step S104, the image acquisition unit 3 receives the acquisition signal 61 transmitted from the signal acquisition unit 2 in step S103. Also in step S104, the image acquisition unit 3 acquires the monitoring image 50 when the signal acquisition unit 2 acquired the reference signal, from among the monitoring images 50 received by the image receiving unit 1. After the processing in step S104, the process proceeds to step S111.

[0051] <<Step S111>> The inference unit 10 inputs the monitoring image 50 received by the image receiving unit 1 to the learning model 5 that the learning device 101 has trained based on the training image 51. The learning model 5 outputs an output result 70. The inference unit 10 outputs the output result 70 from the learning model 5 only in the closed state. The inference unit 10 does not operate in any state other than the closed state.

[0052] <<Step S112>> The analysis unit 11 performs analysis based on the output result 70. The analysis unit 11 calculates the difference between the monitoring image 50 input to the learning model 5 by the inference unit 10 and the output result 70 output by the learning model 5, and analyzes whether the monitoring area 40 is in a reference state.

[0053] <<Step S113>> The analysis unit 11 outputs a detection result 71, which is a result of analyzing whether or not the monitoring area 40 is in a reference state. The detection result 71 is output to a management system for the platform doors 30a, a management system for the train 44, a station management system, etc.

[0054] ***Features of Embodiment 1*** The training information generation device 100 of Embodiment 1 does not define the monitoring area 40 as a place where intrusion is prohibited. The training information generation device 100 of Embodiment 1 defines the monitoring area 40 as a place where the movement of moving objects such as people, robots, etc. is intermittent. The training information generation device 100 of Embodiment 1 does not constantly learn images of the monitoring area 40. The training information generation device 100 of Embodiment 1 learns only images of the monitoring area 40 where no moving objects are moving. The detection device 102 of Embodiment 1 does not detect intrusion of a suspicious person into a place where intrusion is prohibited. The detection device 102 of Embodiment 1 detects obstacles 43 that impede the movement of subsequent moving objects in the monitoring area 40 where the movement of moving objects is intermittent. The detection device 102 of Embodiment 1 does not constantly detect obstacles 43. When the moving object is moving, the detection device 102 of the first embodiment does not detect the obstacle 43. The detection device 102 of the first embodiment detects the obstacle 43 only when the moving object is not moving.

[0055] ***Effects of embodiment 1*** As described above, the learning information generation device 100 of embodiment 1 can reduce the load of acquiring learning information by automatically collecting learning information using a reference signal indicating that the monitoring area 40 is in a reference state as a trigger for acquiring learning information.

[0056] Furthermore, a three-dimensional sensor may be mounted on the platform door 30a to detect obstacles in the area on the platform 41a between the platform door 30a and the tracks 41b. Specifically, the image receiving unit 1 receives an image output by the three-dimensional sensor as a monitoring image 50. However, three-dimensional sensors are expensive. In the present disclosure, costs can be reduced by replacing the three-dimensional sensor with a camera (a two-dimensional optical video camera).

[0057] The learning device 101 of embodiment 1 can reduce the learning load on the learning model 5 because it is only necessary to have the learning model 5 learn the learning images 51 in which the monitoring area 40 is in the reference state.

[0058] *** Supplementary Note on First Embodiment *** A modification of the first embodiment will now be described.

[0059] <<<<Modifications Related to Monitoring Device 20>>> <<When Monitoring Device 20 is a Camera that Captures Still Images>> In the present embodiment, the monitoring device 20 is a video camera that continuously captures images, but the present invention is not limited to this. The monitoring device 20 may also be a camera that captures still images. In this case, as shown in FIG. 8 , the signal acquirer 2 of the learning information generation device 100 further includes a capture instruction unit 2c.

[0060] When the signal acquirer 2 acquires the reference signal, the photographing instruction unit 2c instructs the monitoring device 20, which captures the monitoring image 50, to capture the monitoring image 50. Specifically, the photographing instruction unit 2c receives from the signal determiner 2b a determination result that the status signal 60 is a reference signal. The photographing instruction unit 2c then transmits a photographing instruction signal 62 to the monitoring device 20 to instruct it to capture the monitoring image 50. By receiving the photographing instruction signal 62 from the photographing instruction unit 2c, the monitoring device 20 captures the monitoring image 50 when the monitoring area 40 is in the reference state.

[0061] The image receiving unit 1 receives a monitoring image 50 captured in response to an instruction from the image capturing instruction unit 2c. The image acquiring unit 3 acquires the monitoring image 50 captured in response to an instruction from the image capturing instruction unit 2c, which is received by the image receiving unit 1, as a learning image 51. This eliminates the need for the learning information generating device 100 to select a learning image 51 from the monitoring image 50, thereby further reducing the load of acquiring learning information. Note that the signal determining unit 2b may or may not perform a process of transmitting an acquisition signal 61 to the image acquiring unit 3.

[0062] Next, the learning device 101 and the detection device 102 will be described when the monitoring device 20 is a camera that captures still images. The learning unit 4 of the learning device 101 causes the learning model 5 to learn the learning images 51 captured by the image acquisition unit 3 in response to instructions from the image capture instruction unit 2c.

[0063] 9 , the signal acquisition unit 2 of the detection device 102 further includes a photographing instruction unit 2c. The photographing instruction unit 2c corresponds to the photographing instruction unit 2c included in the learning information generation device 100. When the signal acquisition unit 2 acquires the reference signal, the photographing instruction unit 2c transmits a photographing instruction signal 62 to the monitoring device 20 that captures the monitoring image 50, instructing the monitoring device 20 to capture the monitoring image 50. The image receiving unit 1 receives the monitoring image 50 captured in accordance with the instruction of the photographing instruction unit 2c. The image acquisition unit 3 acquires the monitoring image 50 captured in accordance with the instruction of the photographing instruction unit 2c from the image receiving unit 1.

[0064] The inference unit 10 inputs the monitoring image 50 captured at the instruction of the shooting instruction unit 2c into the learning model 5 that has been trained using the learning image 51 captured at the instruction of the shooting instruction unit 2c, and outputs the output result 70. The analysis unit 11 analyzes whether the monitoring area 40 is in a reference state based on the output result 70 output by the inference unit 10, and outputs the result as a detection result 71. This eliminates the need for the detection device 102 to select the monitoring image 50, thereby further reducing the load of acquiring the monitoring image 50.

[0065] <<When the monitoring device 20 is other than a 2D optical camera>> In the present embodiment, the monitoring device 20 is an optical camera, but this is not limiting. The monitoring device 20 may be a 3D sensor or an infrared camera. Furthermore, different types of monitoring devices 20 may be used in combination.

[0066] <<When the monitoring device 20 is installed on the platform 41a other than the platform door 30a>> In the present embodiment, the case where the monitoring device 20 is installed on the platform door 30a provided on the station platform 41a has been described, but this is not limited to this. The monitoring device 20 may be installed on equipment, a roof, a pillar, or the like provided on the platform 41a so as to be able to capture an image of the platform 41a and the three-dimensional area on the platform 41a sandwiched between the platform door 30a and the tracks 41b. The monitoring device 20 may be mounted on a moving object such as a vehicle, a robot, or a drone. It is desirable for the monitoring device 20 to capture images so that the angle of view and the imaging angle are constant, but the angle of view and the imaging angle may change. In this case, it is desirable to perform correction so that the angle of view and the imaging angle of the captured monitoring image 50 are consistent.

[0067] <<<<Modifications Related to Image Acquisition Unit 3>>> <<When Image Acquisition Unit 3 Determines Whether to Acquire Learning Images 51>> In this embodiment, in response to the signal acquisition unit 2 acquiring a reference signal, the image acquisition unit 3 acquires, as the learning images 51, the monitoring images 50 at the time when the signal acquisition unit 2 acquired the reference signal, but this is not limited to this. In response to the signal acquisition unit 2 acquiring the reference signal, the image acquisition unit 3 may perform a process of determining whether to acquire learning images 51. For example, a threshold value for how many learning images 51 are necessary may be set, and if the number of learning images 51 acquired by the image acquisition unit 3 exceeds the threshold value, the image acquisition unit 3 may determine not to acquire learning images 51.

[0068] <<<<Modifications Related to the Reference Signal Acquired by the Signal Acquirer 2>>> <<Cases Where the Reference Signal is Other than a Signal Indicating the Closed State of the Platform Doors 30a>> In the present embodiment, the case where the reference signal is a signal indicating the closed state of the platform doors 30a has been described, but the present invention is not limited to this. The signal acquirer 2 may acquire, as the reference signal, a signal indicating the state from when the train 44 that has entered the platform 41a stops until the doors of the train 44 open.

[0069] <<When the Monitoring Area 40 and the Opening / Closing Device 30 are Other than the Platform Doors 30a>> In the present embodiment, a case has been described in which the image receiving unit 1 receives a monitoring image 50 in which the monitoring area 40 is an area near the platform doors 30a. However, the present invention is not limited to this. The image receiving unit 1 may receive, as the monitoring image 50, an image captured in which the area around the opening / closing device 30 other than the platform doors 30a is the monitoring area 40. As shown in FIG. 10 , the opening / closing device 30 is, for example, a crossing gate, an elevator, a ticket gate, a security gate, an automatic door, an automatic revolving door, a vehicle, or the like. When the opening / closing device 30 is a crossing gate and the monitoring area 40 is a railroad crossing, the signal acquiring unit 2 acquires, as the reference signal, a signal indicating that the crossing gate is closed. When the opening / closing device 30 is an elevator and the monitoring area 40 is the inside of an elevator car, the signal acquiring unit 2 acquires, as the reference signal, a signal indicating that a button in the car has not been pressed for a predetermined period of time or a signal indicating that no person is in the car. When the monitoring area 40 is an elevator hall, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that a hall button has not been pressed for a predetermined period of time or a signal indicating that there is no person in the elevator hall. When the opening / closing device 30 is a ticket gate and the monitoring area 40 is an entrance to the ticket gate, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that the doors have been closed for a predetermined period of time. When the opening / closing device 30 is a security gate and the monitoring area 40 is an entrance to the security gate, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that the doors have been closed for a predetermined period of time. When the opening / closing device 30 is an automatic door and the monitoring area 40 is the vicinity of the automatic door, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that the doors have not been opened or closed for a predetermined period of time. When the opening / closing device 30 is an automatic revolving door and the monitoring area 40 is inside the automatic revolving door, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that the doors have not been rotated for a predetermined period of time. When the opening / closing device 30 is a vehicle and the monitoring area 40 is the interior of the vehicle, the signal acquisition unit 2 acquires, as the reference signal, a signal indicating that the boarding / alighting doors are closed.

[0070] <<<<Modifications Related to Learning Model 5>>> <<Model Used in Learning Model 5>> In this embodiment, the learning model 5 is a model including an autoencoder 5a using a CNN, but is not limited to this. The autoencoder 5a included in the learning model 5 does not have to be a convolutional autoencoder. The learning model 5 may use a model other than the autoencoder 5a, or another model may be used in combination with the autoencoder 5a. The learning model 5 may be a model that utilizes background subtraction, a technology that compares a prepared background image with an input image to find objects that are not in the background. The learning model 5 may be a model that utilizes a technology that detects objects based on optical flow that occurs when an object is moving in an image.

[0071] <<<<Modifications Related to the Detection Device 102>>> <<Monitoring Image 50 Acquired by the Inference Unit 10>> In the detection device 102 of the present embodiment, an obstacle 43 is not detected when a moving object such as a person or a robot is moving, but this is not limited to this. In the detection device 102 of the present embodiment, the image acquisition unit 3 acquires the monitoring image 50 in the reference state in which the platform doors 30a are closed from the image receiving unit 1, but this is not limited to this. In the detection device 102 of the present embodiment, the inference unit 10 inputs the monitoring image 50 acquired by the image acquisition unit 3 into the trained learning model 5, but this is not limited to this.

[0072] The detection device 102 may detect an obstacle 43 even when the mobile object is moving. The detection device 102 may not include the image acquisition unit 3. The inference unit 10 may input the monitoring image 50 received by the image receiving unit 1 to the learning model 5 even when the platform doors 30a are closed. In the processing performed by the detection device 102 shown in FIG. 7 , step S104, which is the processing performed by the image acquisition unit 3, may be omitted.

[0073] However, if the monitoring image 50 when the platform doors 30a are not in a closed state is input to the learning model 5, the accuracy of obstacle detection may decrease. This is because if the inference unit 10 inputs the monitoring image 50 when the platform doors 30a are in an open state to the learning model 5, the detection device 102 may detect a moving object entering or exiting the platform doors 30a as an obstacle 43.

[0074] Therefore, the learning model 5 in this modified example further includes a model that, when a moving object is included in the input monitoring image 50, infers whether the moving object is entering or exiting the platform doors 30a when the platform doors 30a are in the open state. In addition to learning to infer whether the monitoring area 40 is in the reference state, the learning model 5 is a model that has learned that a moving object entering or exiting the platform doors 30a when the platform doors 30a are in the open state is not an obstacle 43 (no abnormality). The signal storage unit pre-stores status signal identification information for identifying the status signal 60. Note that the reference signal identification information is included in the status signal identification information.

[0075] The detection device 102 in this modified example will now be described. The operations of the image receiving unit 1 and the status signal acquiring unit 2a are the same as those of the image receiving unit 1 and the status signal acquiring unit 2a of the detection device 102 described in the first embodiment. The signal determining unit 2b identifies the status signal 60 acquired by the status signal acquiring unit 2a based on the status signal identification information stored in the signal storage unit. When the signal determining unit 2b identifies the status signal 60 as a reference signal or as a signal indicating the open state of the platform doors 30a, the signal determining unit 2b transmits the identification result to the inference unit 10. When the signal acquiring unit 2b acquires the reference signal or the signal indicating the open state of the platform doors 30a, the inference unit 10 inputs the monitoring image 50 received by the image receiving unit 1 and the identification result by the signal determining unit 2b into the learning model 5. When a moving object is included in the input monitoring image 50, the learning model 5 infers whether the moving object is entering or exiting the platform doors 30a in an open state. When inferring whether the moving object is entering or exiting the platform doors 30a in an open state, the learning model 5 refers to the identification result by the signal determination unit 2b. When the learning model 5 infers that the moving object included in the monitoring image 50 is entering or exiting the platform doors 30a in an open state (no abnormality), the inference unit 10 outputs an inference result indicating no abnormality in addition to the output result 70. The analysis unit 11 detects a difference between the output result 70 output by the inference unit 10 and the monitoring image 50 input to the inference unit 10, and detects an obstacle 43 based on the difference. When the detected difference is equal to or greater than a threshold, and the learning model 5 infers that no abnormality exists, the analysis unit 11 determines that no obstacle exists.

[0076] In this modified example, the detection device 102 does not need to include the image acquisition unit 3 and the signal acquisition unit 2. When a moving object is included in the input surveillance image 50, the learning model 5 infers whether the moving object is entering or exiting the platform doors 30a in the open state, without referring to the identification result by the signal determination unit 2b. In the processing performed by the detection device 102 shown in Figure 7, steps S102, S103, and S104, which are the processing performed by the image acquisition unit 3 and the signal acquisition unit 2, may be omitted.

[0077] <<Detection Result 71 of Analysis Unit 11>> In the present embodiment, the analysis unit 11 outputs whether or not the monitoring area 40 is in a reference state (presence or absence of an obstacle 43) as the detection result 71, but this is not limiting. The analysis unit 11 may output, as the detection result 71, a difference region between the output result 70 output by the inference unit 10 and the monitoring image 50 used for the inference, as an index indicating the degree of abnormality. The analysis unit 11 may identify the true identity of the obstacle 43 and output the identification result as the detection result 71.

[0078] <<<<Modifications Regarding the Configurations of the Training Information Generation Device 100, the Learning Device 101, and the Detection Device 102>>> <<Configuration of the Training Information Generation Device 100>> In the present embodiment, a configuration has been described in which the training information generation device 100 is included in the learning device 101, but this is not limiting. The training information generation device 100 may not be included in the learning device 101. In this case, as shown in the example of processing performed by the training information generation device 100 and the learning device 101 in FIG. 6 , after the processing of acquiring the training image 51 in step S104, the process returns to the processing of receiving the monitoring image 50 in step S101. <<Configuration of the Detection Device 102>> In the present embodiment, a configuration has been described in which the detection device 102 does not include the training information generation device 100 and the learning device 101, but this is not limiting. The detection device 102 may include the training information generation device 100 and the learning device 101.

[0079] Embodiment 2. The training information generation device 100, the training device 101, and the detection device 102 will be described with reference to FIGS. 11 to 16. In the embodiments and drawings, the same or corresponding elements are designated by the same reference numerals. Descriptions of elements designated by the same reference numerals as previously described elements will be omitted or simplified as appropriate. Arrows in the figures primarily indicate the flow of data or the flow of processing.

[0080] ***Outline of Embodiment 2*** First, an outline of this embodiment will be described. Because many obstacles are often on the ground, movement in the background is ignored by setting the learning information generation device 100 to mask everything except the ground area (platform 41a) included in the monitoring image 50 and the learning image 51. Furthermore, because the learning model 5 only needs to learn the area excluding the masked area 80, the acquisition of learning information and the learning processing load can be reduced.

[0081] <<<During Learning>>> <<Functions of Learning Information Generation Device 100>> The following is a functional overview of the learning information generation device 100 of this embodiment. For video acquired from a monitoring device 20 such as a camera, a mask area setting unit 6 selects a home area, and sets the area outside the selected area as a mask area 80. The set mask area 80 is stored in a mask area storage unit 7, and a mask is applied to the learning image by a mask area application unit 8.

[0082] <<Configuration of the Learning Information Generation Device 100>> In order to perform the above functions, the learning information generation device 100 further includes a mask area setting unit 6, a mask area storage unit 7, and a mask area application unit 8. The mask area setting unit 6 is set by the user while viewing the input image (monitoring image 50) or automatically detects the area of ​​the platform 41a using Area Detection or the like, and sets the area other than the platform 41a as a mask area 80. The mask area storage unit 7 stores the set mask area 80. The mask area application unit 8 applies the set mask area 80 to the monitoring image 50.

[0083] <<<<During Analysis>>> <<Functions of Detection Device 102>> The following is a functional overview of the detection device 102 of this embodiment. A mask area application unit 8 applies a mask area 80 stored in a mask area storage unit 7 to an image acquired from a monitoring device 20 such as a camera.

[0084] <<Configuration of Detection Device 102>> In order to perform the above functions, the detection device 102 further includes a mask region storage unit 7 and a mask region application unit 8. Note that the mask region storage unit 7 and the mask region application unit 8 correspond to those included in the learning information generation device 100.

[0085] ***Description of Configuration*** <<<Functional Configuration of the Learning Information Generation Device 100>>> The learning information generation device 100 will be described with reference to Fig. 11. The learning information generation device 100 differs from the first embodiment in that it further includes a mask area setting unit 6, a mask area storage unit 7, and a mask area application unit 8.

[0086] The image receiving unit 1 receives an image including a station platform 41 a as a monitoring image 50. The image acquiring unit 3 acquires an image including the station platform 41 a from the monitoring image 50 received by the image receiving unit 1 as a learning image 51.

[0087] <<Mask Area Setting Unit 6 and Mask Area Storage Unit 7>> The mask area setting unit 6 acquires a monitoring image 50 including the platform 41a from the image receiving unit 1. The mask area setting unit 6 sets a mask area 80 using area detection AI for automatically detecting areas, called Area Detection.

[0088] The mask area 80 is an area excluding the ground area, which is at least one of the ground and the floor surface, in the monitoring area 40. In this embodiment, the mask area 80 is an area excluding the platform 41a.

[0089] The area detection AI is a learning model that has been trained to detect the platform 41a included in the surveillance image 50. The mask area setting unit 6 inputs the surveillance image 50 to the area detection AI and detects the platform 41a. The mask area setting unit 6 sets the area excluding the platform 41a detected by the area detection AI as a mask area 80. The mask area storage unit 7 stores the mask area 80 set by the mask area setting unit 6.

[0090] <<Mask Area Application Unit 8>> The mask area application unit 8 masks the learning image 51 with a mask area 80. The masking process by the mask area application unit 8 will be specifically described with reference to FIG. 12 . The mask area application unit 8 acquires the learning image 51 from the image acquisition unit 3. The mask area application unit 8 reads out the mask area 80 from the mask area storage unit 7. The mask area application unit 8 masks the learning image 51 with the mask area 80, which is the area excluding the home 41 a, to create a masked learning image 52.

[0091] <<<Functional Configuration of Learning Device 101>>> The learning device 101 will be described with reference to Fig. 11. The learning device 101 differs from the first embodiment in that the learning unit 4 of the learning device 101 causes the learning model 5 to learn the masked learning images 52. The learning unit 4 acquires the masked learning images 52 from the mask region application unit 8. The learning unit 4 causes the learning model 5 to learn the acquired masked learning images 52.

[0092] <<<Functional Configuration of Detection Device 102 >>> The detection device 102 will be described with reference to Fig. 13. The detection device 102 differs from the first embodiment in that it further includes a mask region storage unit 7 and a mask region application unit 8.

[0093] <<Mask Region Storage Unit 7 >> The mask region storage unit 7 corresponds to the mask region storage unit 7 of the device for generating information for learning 100 , and operates in the same manner as the mask region storage unit 7 of the device for generating information for learning 100 .

[0094] <<Mask Area Application Unit 8>> The mask area application unit 8 corresponds to the mask area application unit 8 of the learning information generation device 100. The mask area application unit 8 reads out a mask area 80 from the mask area storage unit 7. The mask area application unit 8 masks the mask area 80, which is the area excluding the home 41a, from the surveillance image 50 acquired from the image acquisition unit 3, to create a masked surveillance image 53.

[0095] <<Inference Unit 10 >> The inference unit 10 inputs the masked surveillance image 53 into the learning model 5 that has been trained by the learning device 101 , and outputs the output result 70 .

[0096] <<Analysis Unit 11>> The analysis unit 11 operates in the same manner as the analysis unit 11 of the detection device 102 in the first embodiment.

[0097] <<<<Hardware Configuration of the Training Information Generation Device 100, the Learning Device 101, and the Detection Device 102>>> The hardware configuration of the training information generation device 100, the learning device 101, and the detection device 102 will be described with reference to Figure 14. Note that descriptions of parts that are the same as those in embodiment 1 will be omitted. The storage device 91 stores programs for realizing the functions of the image receiving unit 1, the signal acquiring unit 2, the image acquiring unit 3, the learning unit 4, the inference unit 10, the analysis unit 11, the mask area setting unit 6, and the mask area application unit 8, as well as the learning model 5 and the set mask area 80.

[0098] ***Description of Operation*** <<<Learning Information Generation Device 100, Learning Device 101>>> An example of processing performed by the learning information generation device 100 and the learning device 101 will be described with reference to the flowchart shown in FIG.

[0099] <<Step S101>> The image receiving unit 1 performs the same operation as step S101 in the learning information generation device 100 shown in Fig. 6. After the process of step S101, the process proceeds to step S201.

[0100] <<Step S201>> The mask area setting unit 6 determines whether a mask area 80 has been set (whether a mask area 80 has been stored in the mask area storage unit 7). If the mask area setting unit 6 has already set a mask area, the process proceeds to step S102. If the mask area setting unit 6 has not set a mask area, the process proceeds to step S202.

[0101] <<Step S202>> The mask area setting unit 6 acquires the monitoring image 50 including the home 41a from the image receiving unit 1, and sets the mask area 80. The mask area 80 is not set constantly by the mask area setting unit 6, but only needs to be set once. The mask area 80 set by the mask area setting unit 6 is stored in the mask area storage unit 7. After the processing of step S202, the process proceeds to step S102.

[0102] <<Step S102>> The state signal acquiring unit 2a of the signal acquiring unit 2 performs the same operation as step S102 in the learning information generating device 100 shown in FIG.

[0103] <<Step S103>> The signal determination unit 2b performs the same operation as step S103 in the learning information generation device 100 shown in FIG.

[0104] <<Step S104>> The image acquisition unit 3 performs the same operation as step S104 in the learning information generation device 100 shown in Fig. 6. After the process of step S104, the process proceeds to step S203.

[0105] <<Step S203>> The mask area application unit 8 masks the learning image 51. Specifically, the mask area application unit 8 reads the mask area 80 from the mask area storage unit 7. The mask area application unit 8 masks the learning image 51 with the mask area 80 to create a masked learning image 52. After processing in step S203, the process proceeds to step S105.

[0106] <<Step S105>> The learning unit 4 uses the masked learning images 52 to train the learning model 5. After that, the process returns to step S101.

[0107] <<<Detection Device 102 >>> An example of processing performed by the detection device 102 will be described with reference to the flowchart shown in FIG.

[0108] <<Steps S101 to S103>> In step S101, the image receiving unit 1 performs the same operation as in step S101 in the detection device 102 shown in Fig. 7. In step S102, the signal acquiring unit 2 performs the same operation as in step S102 in the detection device 102 shown in Fig. 7. In step S103, the signal determining unit 2b performs the same operation as in step S103 in the detection device 102 shown in Fig. 7.

[0109] <<Step S104>> In step S104, the image acquisition unit 3 receives the acquisition signal 61 from the signal acquisition unit 2 and acquires, from the monitoring images 50 received by the image receiving unit 1, the monitoring image 50 when the signal acquisition unit 2 acquired the reference signal. After the processing in step S104, the process proceeds to step S203.

[0110] <<Step S203>> The mask area application unit 8 masks the monitoring image 50. Specifically, the mask area application unit 8 reads out the mask area 80 from the mask area storage unit 7. The mask area application unit 8 masks the monitoring image 50 with the mask area 80 to create a masked monitoring image 53. After the processing in step S203, the process proceeds to step S111.

[0111] <<Step S111>> The inference unit 10 inputs the masked surveillance image 53 created by the mask region application unit 8 to the learning model 5 that the learning device 101 has trained based on the masked training image 52. The learning model 5 outputs the output result 70.

[0112] <<Step S112>> The analysis unit 11 performs an analysis based on the output result 70. The analysis unit 11 calculates the difference between the masked surveillance image 53 input to the learning model 5 by the inference unit 10 and the output result 70 output by the learning model 5, and analyzes whether the surveillance area 40 is in a reference state.

[0113] <<Step S113>> In step S113, the analysis unit 11 performs the same operation as in step S113 in the detection device 102 shown in FIG.

[0114] ***Effects of Embodiment 2*** As described above, the training information generation device 100 of Embodiment 2 can ignore movement of the background 42 of the training image 51 by masking the area of ​​the training image 51 other than the platform 41a. Furthermore, the detection device 102 of Embodiment 2 can ignore movement of the background 42 of the monitoring image 50 by masking the area of ​​the monitoring image 50 other than the platform 41a. Therefore, the detection device 102 of Embodiment 2 can improve the obstacle detection accuracy of an object (obstacle 43) placed on the platform 41a.

[0115] *** Supplementary Note on Second Embodiment *** A modification of the second embodiment will now be described.

[0116] <<<<Modifications Related to the Mask Area Setting Unit 6 of the Learning Information Generation Device 100>>> <<When the Monitoring Area 40 and the Opening and Closing Device 30 are Other Than the Platform Doors 30a>> In the present embodiment, a case has been described in which the image receiving unit 1 receives a monitoring image 50 in which the monitoring area 40 is an area near the platform doors 30a. Therefore, the mask area setting unit 6 sets an area excluding the platform 41a as the mask area 80, but this is not limited thereto. The mask area setting unit 6 may set an area excluding a ground area different from the platform 41a as the mask area 80.

[0117] As shown in FIG. 17 , the mask area 80 may be changed as desired to suit the opening / closing device 30. If the opening / closing device 30 is a barrier and the monitoring area 40 is within a railroad crossing, the mask area 80 is the area excluding the ground surface within the railroad crossing (the ground surface includes the tracks). If the opening / closing device 30 is an elevator and the monitoring area 40 is within a car, the mask area 80 is the area excluding the floor surface within the car. If the monitoring area 40 is an elevator hall, the mask area 80 is the area excluding the floor surface of the elevator hall. If the opening / closing device 30 is a ticket gate and the monitoring area 40 is an entrance to the ticket gate, the mask area 80 is the area excluding the floor surface of the entrance to the ticket gate. If the ticket gate is installed outdoors, the mask area 80 is the area excluding the ground surface of the entrance to the ticket gate. If the opening / closing device 30 is a security gate and the monitoring area 40 is an entrance to the security gate, the mask area 80 is the area excluding the floor surface of the entrance to the security gate. If the security gate is installed outdoors, the mask area 80 is the area excluding the ground at the security gate entrance. If the opening / closing device 30 is an automatic door and the monitoring area 40 is the area around the automatic door installed inside a building, the mask area 80 is the area excluding the floor around the automatic door. If the monitoring area 40 is the area around an automatic door installed at the entrance of a building, the mask area 80 is the area excluding the floor and ground around the automatic door. If the opening / closing device 30 is an automatic revolving door and the monitoring area 40 is inside an automatic revolving door installed inside a building, the mask area 80 is the area excluding the floor inside the automatic revolving door. If the monitoring area 40 is inside an automatic revolving door installed at the entrance of a building, the mask area 80 is the area excluding the ground inside the automatic revolving door. If the opening / closing device 30 is a vehicle and the monitoring area 40 is inside the vehicle, the mask area 80 is the area excluding the seats inside the vehicle. Alternatively, the mask area 80 may be the area excluding the floor inside the vehicle.

[0118] The image receiving unit 1 receives an image including a ground area, which is at least one of the ground and the floor, as the monitoring image 50. The image acquiring unit 3 acquires an image including the ground area as the learning image 51 from the monitoring image 50 received by the image receiving unit 1.

[0119] The area detection AI used in the mask area setting unit 6 is a model trained to detect a ground area, which is at least one of the ground and floor surface, included in the monitoring image 50 according to the mask area 80 corresponding to the opening and closing device 30.

[0120] <<When the mask area setting unit 6 sets the mask area 80 without using area detection AI>> In the present embodiment, a case has been described in which the mask area setting unit 6 sets the mask area 80 using area detection AI, but the present invention is not limited to this. The mask area setting unit 6 may set, as the mask area 80, an area set by the user while viewing the monitoring image 50.

[0121] <<<<Modifications Related to the Mask Area Application Unit 8 of the Detection Device 102>>> <<When the Mask Area Application Unit 8 Acquires the Surveillance Image 50 from the Image Receiving Unit 1>> In the present embodiment, the mask area application unit 8 of the detection device 102 acquires the surveillance image 50 from the image acquiring unit 3 and creates the masked surveillance image 53, but this is not limiting. The mask area application unit 8 may acquire the surveillance image 50 from the image receiving unit 1 and create the masked surveillance image 53. In particular, when the detection device 102 does not include the image acquiring unit 3, the mask area application unit 8 acquires the surveillance image 50 from the image receiving unit 1.

[0122] Embodiment 3. The learning information generation device 100, learning device 101, and detection device 102 will be described with reference to FIGS. 18 to 21. In the embodiments and drawings, the same or corresponding elements are designated by the same reference numerals. Descriptions of elements designated by the same reference numerals as previously described elements will be omitted or simplified as appropriate. Arrows in the figures primarily indicate the flow of data or the flow of processing.

[0123] ***Outline of Third Embodiment*** First, an outline of this embodiment will be described. An area where a train 44 exists is detected from a video (monitoring image 50) captured by a 2D optical video camera (monitoring device 20), and masking processing is performed. The method of detecting the area where the train 44 exists is to utilize the fact that the train 44 travels in a constant direction, and mask an area where optical flows 45 occur collectively in the same direction. Alternatively, the train 44 is detected by object detection, and the detected area is masked.

[0124] <<<During Learning>>> <<Functions of the Learning Information Generating Device 100>> The following is a functional overview of the learning information generating device 100 of this embodiment. The motion vector calculating unit 9 calculates an optical flow 45 for a video (monitoring image 50) acquired from a 2D optical video camera (monitoring device 20). The motion vector calculating unit 9 extracts, as a mask area 80, an area where optical flows 45 occur collectively in the same direction as the area where the train 44 is running.

[0125] <<Configuration of the Learning Information Generation Device 100>> In order to perform the above functions, the learning information generation device 100 further includes a motion vector calculation unit 9. The motion vector calculation unit 9 calculates motion vectors (optical flows 45) from the input image (monitoring image 50) and extracts areas of the optical flows 45 that are generated collectively in the same direction.

[0126] ***Description of Configuration*** <<<Functional Configuration of the Training Information Generation Device 100>>> The training information generation device 100 will be described with reference to Fig. 18. The training information generation device 100 differs from the first embodiment in that it further includes a mask region storage unit 7, a mask region application unit 8, and a motion vector calculation unit 9.

[0127] <<Image Receiving Unit 1 and Image Acquiring Unit 3>> The image receiving unit 1 receives an image including a train 44 as a monitoring image 50. The image acquiring unit 3 acquires an image including the train 44 as a learning image 51 from the monitoring image 50 received by the image receiving unit 1.

[0128] The mask area 80 in this embodiment is an area in the monitoring image 50 where the train 44 is present. The train 44 runs on tracks 41b outside the platform 41a. The train 44 may be in any operating state, such as running or stopped. The train 44 may be in any position when stopped. The train 44 may be of any type.

[0129] <<Motion Vector Calculation Unit 9>> The operation of the motion vector calculation unit 9 to detect a mask area 80 will be described. The motion vector calculation unit 9 calculates an optical flow 45 indicating the movement of a certain area from a plurality of consecutive monitoring images 50 received by the image receiving unit 1. The motion vector calculation unit 9 detects a mask area 80 based on the calculated optical flow 45. Specifically, the motion vector calculation unit 9 determines whether the optical flow 45 generated collectively in the same direction is larger than a preset threshold. If the optical flow 45 is larger than the threshold, the motion vector calculation unit 9 considers the area where the optical flow 45 is generated as an area where a train 44 exists and detects it as a mask area 80. The motion vector calculation unit 9 stores the detected mask area 80 in the mask area storage unit 7. If the optical flow 45 is smaller than the threshold, the motion vector calculation unit 9 does not detect the mask area 80.

[0130] <<Mask Region Storage Unit 7 >> The mask region storage unit 7 stores the mask region 80 detected by the motion vector calculation unit 9 .

[0131] <<Mask Area Application Unit 8>> The mask area application unit 8 masks the learning image 51 with a mask area 80. The masking process by the mask area application unit 8 will be specifically described with reference to FIG. 19 . The mask area application unit 8 acquires the learning image 51 from the image acquisition unit 3. The mask area application unit 8 reads out the mask area 80 from the mask area storage unit 7. The mask area application unit 8 masks the learning image 51 with the mask area calculated by the motion vector calculation unit 9, to create a masked learning image 52.

[0132] <<<Functional Configuration of Learning Device 101>>> The learning device 101 will be described with reference to Fig. 18. The learning device 101 in embodiment 3 operates in the same manner as the learning device 101 in embodiment 2. However, the masked learning image 52 input by the learning unit 4 to the learning model 5 is an image obtained by masking the mask area 80 detected by the motion vector calculation unit 9 from the learning image 51.

[0133] <<<Functional Configuration of Detection Device 102>>> The detection device 102 has the same configuration as the detection device 102 in embodiment 2. The detection device 102 operates in the same manner as the detection device 102 in embodiment 2. However, the mask area 80 stored in the mask area storage unit 7 is the mask area 80 detected by the motion vector calculation unit 9 of the learning information generation device 100. Furthermore, the mask area application unit 8 masks the mask area 80, which is the area where the train 44 is present and which was detected by the motion vector calculation unit 9, in the surveillance image 50 acquired from the image acquisition unit 3, to create a masked surveillance image 53.

[0134] <<<Hardware Configuration of the Training Information Generation Device 100, the Learning Device 101, and the Detection Device 102>>> The hardware configuration of the training information generation device 100, the learning device 101, and the detection device 102 will be described with reference to Figure 20. Note that descriptions of parts that are the same as those in the first and second embodiments will be omitted. The storage device 91 stores programs for realizing the functions of the image receiving unit 1, the signal acquiring unit 2, the image acquiring unit 3, the learning unit 4, the inference unit 10, the analysis unit 11, the motion vector calculation unit 9, and the mask area application unit 8, as well as a learning model 5 and a set mask area 80.

[0135] ***Description of Operation*** <<<Learning Information Generation Device 100, Learning Device 101>>> An example of processing performed by the learning information generation device 100 and the learning device 101 will be described with reference to the flowchart shown in FIG.

[0136] <<Step S101>> The image receiving unit 1 performs the same operation as step S101 in the learning information generation device 100 shown in Fig. 15. After the process of step S101, the process proceeds to step S201.

[0137] <<Step S201>> The motion vector calculation unit 9 determines whether the mask area 80 has been set (whether the mask area 80 has been stored in the mask area storage unit 7). If the mask area 80 has already been set in the motion vector calculation unit 9, the process proceeds to step S102. If the mask area 80 has not been set in the motion vector calculation unit 9, the process proceeds to step S301.

[0138] <<Step S301>> The motion vector calculation unit 9 acquires the monitoring image 50 including the train 44 from the image receiving unit 1, and sets a mask area 80. The mask area 80 set by the motion vector calculation unit 9 is stored in the mask area storage unit 7. After the processing of step S301, the process proceeds to step S102.

[0139] <<Step S102>> The state signal acquiring unit 2a of the signal acquiring unit 2 performs the same operation as step S102 in the learning information generating device 100 shown in FIG.

[0140] <<Step S103>> The signal determination unit 2b performs the same operation as step S103 in the learning information generation device 100 shown in FIG.

[0141] <<Step S104>> The image acquisition unit 3 performs the same operation as step S104 in the learning information generation device 100 shown in Fig. 15. After the process of step S104, the process proceeds to step S203.

[0142] <<Step S203>> The mask area application unit 8 performs the same operation as step S203 in the training information generation device 100 shown in Fig. 15. However, the mask area application unit 8 masks the mask area 80 detected by the motion vector calculation unit 9 on the training image 51 to create a masked training image 52. After processing in step S203, the process proceeds to step S105.

[0143] <<Step S105>> The learning unit 4 performs the same operation as step S105 in the learning information generation device 100 shown in Fig. 15. Then, the process returns to step S101.

[0144] <<<Detection Device 102>>> The processing performed by the detection device 102 in the third embodiment is the same as the processing performed by the detection device 102 in the second embodiment shown in Fig. 16. However, in step S203, the mask area application unit 8 uses the mask area 80 detected by the motion vector calculation unit 9 as the mask for the monitoring image 50.

[0145] ***Effects of Embodiment 3*** The train 44 outside the platform 41a included in the monitoring image 50 is likely to vary greatly depending on its operating state (stopped or running), its position when stopped, and the types of trains. As described above, the learning information generation device 100 of Embodiment 3 can mask the area in which the train 44 is located in the training image 51 to ignore movement in the background 42 of the training image 51. Furthermore, because the motion vector calculation unit 9 detects the area in which the train 44 is located based on calculation of the optical flow 45, the burden on detecting the masked area 80 can be reduced. The detection device 102 of Embodiment 3 can improve the accuracy of obstacle detection by excluding the train 44 included in the monitoring image 50 from obstacle detection targets.

[0146] *** Supplementary Note on Third Embodiment *** A modification of the third embodiment will now be described.

[0147] <<<<Modifications Related to the Motion Vector Calculation Unit 9>>> <<Detection Operation of the Mask Region 80 by the Motion Vector Calculation Unit 9>> In the present embodiment, the motion vector calculation unit 9 detects the mask region 80, which is the region where the train 44 exists, by calculating the optical flow 45 from a plurality of consecutive monitoring images 50. However, this is not limited to this. The motion vector calculation unit 9 may detect the train 44 by object detection and set the detected region as the mask region 80. Specifically, the motion vector calculation unit 9 includes an object detection AI. The object detection AI is a model that has previously learned to detect the train 44 included in the monitoring image 50. The motion vector calculation unit 9 inputs the monitoring image 50 to the object detection AI. The motion vector calculation unit 9 stores the region where the train 44 exists in the monitoring image 50, detected by the object detection AI, in the mask region storage unit 7 as the mask region 80.

[0148] The motion vector calculation unit 9 may combine detection of the mask area 80 by calculating the optical flow 45 and detection of the mask area 80 by the object detection AI. Specifically, the motion vector calculation unit 9 detects the mask area 80 by calculating the optical flow 45, and then detects the mask area 80 by the object detection AI. The motion vector calculation unit 9 calculates the difference between the mask area 80 by calculating the optical flow 45 and the mask area 80 by the object detection AI.

[0149] The motion vector calculation unit 9 determines that the two mask areas 80 represent the same area when the difference between the two mask areas 80 is equal to or less than a preset threshold. Thereafter, the motion vector calculation unit 9 stores the mask area 80 calculated using the optical flow 45 in the mask area storage unit 7. Note that the motion vector calculation unit 9 may also store the mask area 80 calculated using the area detection AI in the mask area storage unit 7.

[0150] The motion vector calculation unit 9 determines that the two mask areas 80 represent different areas when the difference between the two mask areas 80 is equal to or greater than a preset threshold. In other words, the motion vector calculation unit 9 determines that at least one of the mask area 80 calculated using the optical flow 45 and the mask area 80 calculated using the object detection AI is an incorrect mask area 80. In this case, the motion vector calculation unit 9 acquires a new monitoring image 50 and re-detects the mask area 80.

[0151] Aspects of the present disclosure are described below as appendices. (Appendix 1) A learning information generation device comprising: an image receiving unit that receives monitoring images of a monitoring area; a signal acquiring unit that acquires a reference signal indicating that the monitoring area is in a reference state; and an image acquiring unit that, in response to the signal acquiring unit acquiring the reference signal, acquires, as a learning image, from the monitoring images received by the image receiving unit, a monitoring image that was acquired when the signal acquiring unit acquired the reference signal. (Appendix 2) The signal acquiring unit has a status signal acquiring unit that acquires a status signal indicating the state of the monitoring area, and a signal determining unit that determines whether the status signal acquired by the status signal acquiring unit is the reference signal, and the image acquiring unit acquires the learning image when the signal determining unit determines that the signal acquiring unit has acquired the reference signal. (Supplementary Note 3) The learning information generation device according to Supplementary Note 1 or 2, wherein the signal acquisition unit has an image capture instruction unit that instructs a monitoring device that captures the monitoring image to capture the monitoring image in response to the signal acquisition unit acquiring the reference signal, and the image acquisition unit acquires the monitoring image captured in accordance with the instruction of the image capture instruction unit as the learning image. (Supplementary Note 4) The learning information generation device according to any one of Supplements 1 to 3, wherein the image receiving unit receives, as the monitoring image, an image captured around a switching device that blocks traffic, and the signal acquisition unit acquires, as the reference signal, a signal indicating a closed state of the switching device. (Supplementary Note 5) The learning information generation device according to any one of Supplements 1 to 4, wherein the image receiving unit receives the monitoring image of the monitoring area that is an area including a station platform, and the signal acquisition unit acquires, as the reference signal, a signal indicating a closed state of a platform door installed on the platform. (Appendix 6) The learning information generation device described in any one of Appendices 1 to 3, wherein the image receiving unit receives the monitoring image of the monitoring area, which is an area including a station platform, and the signal acquiring unit acquires as the reference signal a signal indicating the state of a train that has entered the platform and stopped until the train doors open.(Supplementary Note 7) The learning information generation device according to any one of Supplementary Notes 1 to 6, wherein the image receiving unit receives, as the monitoring image, an image including a ground region which is at least one of the ground and a floor surface, the image acquiring unit acquires, as the learning image, an image including the ground region from the monitoring image received by the image receiving unit, the image including the ground region, and is provided with a mask region application unit that sets a region excluding the ground region of the learning image as a mask region and masks the mask region from the learning image. (Supplementary Note 8) The learning information generation device according to any one of Supplementary Notes 1 to 6, wherein the image receiving unit receives, as the monitoring image, an image including a station platform, the image acquiring unit acquires, as the learning image, an image including the platform from the monitoring image received by the image receiving unit, the image including the platform as a mask region and is provided with a mask region application unit that sets a region excluding the platform of the learning image as a mask region and masks the mask region from the learning image. (Supplementary Note 9) The learning information generation device according to any one of Supplementary Notes 1 to 6, further comprising: the image receiving unit receives an image including a train as the surveillance image; the image acquisition unit acquires an image including the train from the surveillance image received by the image receiving unit as the learning image; and a mask area application unit that sets an area in the learning image where the train is present as a mask area and masks the mask area from the learning image. (Supplementary Note 10) The learning information generation device according to Supplementary Note 9, further comprising: a motion vector calculation unit that calculates an optical flow indicating movement of a certain area from a plurality of consecutive surveillance images received by the image receiving unit and detects the mask area based on the optical flow; and the mask area application unit masks the mask area calculated by the motion vector calculation unit from the learning image. (Supplementary Note 11) A learning device comprising: a learning information generation device that receives a monitoring image of a monitoring area and acquires, as a learning image, the monitoring image when a reference signal indicating that the monitoring area is in a reference state is acquired; and a learning unit that trains a learning model for inferring whether the monitoring area is in the reference state based on the learning image acquired by the learning information generation device.(Supplementary Note 12) A learning device comprising: a learning information generation device that receives monitoring images of a monitored area and acquires, as training images, monitoring images obtained when a reference signal indicating that the monitored area is in a reference state has been acquired; and a learning unit that trains, based on the training images acquired by the training information generation device, a learning model including an autoencoder that outputs an output result for inferring whether or not the monitored area is in a reference state from the monitoring images received by the training information generation device. (Supplementary Note 13) A learning device comprising: a learning information generation device that receives monitoring images of a monitored area and acquires, as training images, monitoring images obtained when a reference signal indicating that the monitored area is in a reference state has been acquired; and a learning unit that trains, based on the training images acquired by the training information generation device, a learning model including an autoencoder that, when an input image is input, extracts feature amounts of the input image and outputs an output result obtained by reconstructing the input image based on the feature amounts, the learning model that trains so that the output result approaches the input image. (Supplementary Note 14) A detection device comprising: an image receiving unit that receives monitoring images of a monitored area, an inference unit that inputs the monitoring images received by the image receiving unit into a learning model that has been trained using training images that are monitoring images when a reference signal indicating that the monitoring area is in a reference state has been acquired, the learning model outputting an output result for inferring whether or not the monitoring area is in the reference state when a monitoring image is input, and outputs the output result from the learning model, and an analysis unit that analyzes whether or not the monitoring area is in the reference state based on the output result output by the inference unit. (Supplementary Note 15) A detection device comprising: an image receiving unit that receives monitoring images of a monitored area, an inference unit that inputs the monitoring images received by the image receiving unit into a learning model that includes an autoencoder that has been trained using training images that are monitoring images when a reference signal indicating that the monitoring area is in the reference state has been acquired, and outputs an output result for inferring from the monitoring image whether or not the monitoring area is in the reference state, and outputs the output result from the learning model, and an analysis unit that analyzes whether or not the monitoring area is in the reference state based on the output result output by the inference unit.(Supplementary Note 16) A detection device comprising: an image receiving unit that receives monitoring images of a monitoring area; an inference unit that extracts feature amounts of a training image that is a monitoring image when a reference signal indicating that the monitoring area is in a reference state is acquired, inputs the monitoring images received by the image receiving unit into a learning model including an autoencoder that has been trained to extract feature amounts of a training image that is a monitoring image when a reference signal indicating that the monitoring area is in a reference state is acquired, and outputs the output result from the learning model; and an analysis unit that analyzes whether the monitoring area is in the reference state based on a difference between the output result output by the inference unit and the monitoring image input to the inference unit. (Supplementary Note 17) A method for generating information for learning, comprising: receiving monitoring images of a monitoring area; acquiring a reference signal that indicates that the monitoring area is in a reference state; and, in response to acquiring the reference signal, acquiring, from the monitoring images, a monitoring image when the reference signal was acquired, as a learning image.

[0152] REFERENCE SIGNS LIST 1 Image receiving unit 2 Signal acquisition unit 2a Status signal acquisition unit 2b Signal determination unit 2c Photography instruction unit 3 Image acquisition unit 4 Learning unit 5 Learning model 5a Autoencoder 6 Mask area setting unit 7 Mask area storage unit 8 Mask area application unit 9 Motion vector calculation unit 10 Inference unit 11 Analysis unit 20 Monitoring device 30 Opening and closing device 30a Platform door 40 Monitoring area 41a Platform 41b Track 42 Background 43 Obstacle 44 Train 45 Optical flow 50 Monitoring image 51 Learning image 52 Masked learning image 53 Masked monitoring image 60 Status signal 61 Acquisition signal 62 Photography instruction signal 70 Output result 71 Detection result 80 Mask area 90 Processor 91 Storage device 92 Communication device 93 Input / output interface 100 Learning information generating device 101 Learning device 102 Detection device

Claims

1. A learning information generation device comprising: an image receiving unit that receives monitoring images of a monitoring area; a signal acquiring unit that acquires a reference signal indicating that the monitoring area is in a reference state; and an image acquiring unit that, in response to the signal acquiring unit acquiring the reference signal, acquires, as a learning image, from the monitoring images received by the image receiving unit, the monitoring image at the time when the signal acquiring unit acquired the reference signal.

2. The learning information generation device of claim 1, wherein the signal acquisition unit has a status signal acquisition unit that acquires a status signal indicating the status of the monitoring area, and a signal determination unit that determines whether the status signal acquired by the status signal acquisition unit is the reference signal, and the image acquisition unit acquires the learning image when the signal determination unit determines that the signal acquisition unit has acquired the reference signal.

3. The learning information generation device described in claim 1, wherein the signal acquisition unit has a shooting instruction unit that instructs a monitoring device that captures the monitoring image to capture the monitoring image upon acquisition of the reference signal by the signal acquisition unit, and the image acquisition unit acquires the monitoring image captured in accordance with the instruction of the shooting instruction unit as the learning image.

4. The learning information generation device described in claim 1, wherein the image receiving unit receives, as the surveillance image, an image of the area around a switching device that blocks traffic, and the signal acquiring unit acquires, as the reference signal, a signal indicating the closed state of the switching device.

5. The learning information generation device described in claim 1, wherein the image receiving unit receives the monitoring image of the monitoring area, which is an area including the station platform, and the signal acquiring unit acquires a signal indicating the closed state of platform doors installed on the platform as the reference signal.

6. The learning information generation device described in claim 1, wherein the image receiving unit receives the monitoring image of the monitoring area, which is an area including a station platform, and the signal acquiring unit acquires as the reference signal a signal indicating the state of a train that has entered the platform and stopped until the train doors open.

7. A learning information generation device as described in any one of claims 1 to 6, wherein the image receiving unit receives an image including a ground area, which is at least one of the ground and a floor, as the monitoring image; the image acquisition unit acquires an image including the ground area from the monitoring image received by the image receiving unit as the learning image; and the device is provided with a mask area application unit that sets an area excluding the ground area of ​​the learning image as a mask area and masks the mask area from the learning image.

8. A learning information generation device as described in any one of claims 1 to 6, wherein the image receiving unit receives an image including a station platform as the surveillance image, the image acquisition unit acquires an image including the platform from the surveillance image received by the image receiving unit as the learning image, and the device is provided with a mask area application unit that sets an area of ​​the learning image excluding the platform as a mask area and masks the mask area from the learning image.

9. The learning information generation device according to any one of claims 1 to 6, wherein the image receiving unit receives an image including a train as the surveillance image, the image acquisition unit acquires an image including the train from the surveillance image received by the image receiving unit as the learning image, and the device is further provided with a mask area application unit that sets an area in the learning image where the train is present as a mask area and masks the mask area from the learning image.

10. A learning information generation device as described in claim 9, further comprising a motion vector calculation unit that calculates an optical flow indicating the movement of a certain area from a plurality of consecutive surveillance images received by the image receiving unit and detects the mask area based on the optical flow, and the mask area application unit masks the mask area calculated by the motion vector calculation unit on the learning image.

11. A learning device comprising: a learning information generation device that receives monitoring images of a monitoring area and acquires the monitoring images when a reference signal indicating that the monitoring area is in a reference state is acquired as learning images; and a learning unit that trains a learning model for inferring whether the monitoring area is in the reference state based on the learning images acquired by the learning information generation device.

12. A learning device comprising: a learning information generation device that receives monitoring images of a monitoring area and acquires the monitoring images when a reference signal indicating that the monitoring area is in a reference state is acquired as learning images; and a learning unit that trains a learning model including an autoencoder that outputs an output result for inferring whether the monitoring area is in a reference state from the monitoring images received by the learning information generation device, based on the learning images acquired by the learning information generation device.

13. A learning device comprising: a learning information generation device that receives monitoring images of a monitoring area and acquires, as a training image, the monitoring image when a reference signal indicating that the monitoring area is in a reference state is acquired; and a learning unit that trains, based on the training images acquired by the learning information generation device, a learning model including an autoencoder that, when an input image is input, extracts features of the input image and outputs an output result that reconstructs the input image based on the features, such that the output result approaches the input image.

14. A detection device comprising: an image receiving unit that receives monitoring images of a monitored area; an inference unit that inputs the monitoring images received by the image receiving unit into a learning model that has been trained using training images that are monitoring images when a reference signal indicating that the monitored area is in a reference state is acquired, the learning model outputting an output result for inferring whether or not the monitored area is in a reference state when a monitoring image is input, and outputs the output result from the learning model; and an analysis unit that analyzes whether or not the monitored area is in the reference state based on the output result output by the inference unit.

15. A detection device comprising: an image receiving unit that receives monitoring images of a monitored area; an inference unit that inputs the monitoring images received by the image receiving unit into a learning model that includes an autoencoder trained using training images that are monitoring images when a reference signal indicating that the monitored area is in a reference state is acquired, and that outputs an output result for inferring from the monitoring images whether or not the monitored area is in a reference state, and outputs the output result from the learning model; and an analysis unit that analyzes whether or not the monitored area is in the reference state based on the output result output by the inference unit.

16. A detection device comprising: an image receiving unit that receives surveillance images of a surveillance area; an inference unit that inputs the surveillance images received by the image receiving unit into a learning model including an autoencoder that has been trained to extract features of a training image, which is a surveillance image when a reference signal indicating that the surveillance area is in a reference state is acquired, and output an output result from the feature values ​​that approaches the training image, and outputs the output result from the learning model; and an analysis unit that analyzes whether the surveillance area is in the reference state based on the difference between the output result output by the inference unit and the surveillance image input to the inference unit.

17. A method for generating learning information, comprising receiving a monitoring image of a monitoring area, acquiring a reference signal indicating that the monitoring area is in a reference state, and, upon acquiring the reference signal, acquiring, as a learning image, one of the monitoring images at the time the reference signal was acquired.

Images