Fall detection system, fall detection method, and fall detection program

Abstract

To provide a system, a method, and a program for detecting a falling person, capable of automatically detecting a person falling into a storage facility where waste is stored within a waste treatment facility.SOLUTION: A falling person detection system 10 includes: a first image data acquisition unit that acquires first image data from a first camera 6 imaging inside a storage facility (pit 3) where processed items are stored; a second image data acquisition unit that acquires second image data from a second camera 23 imaging inside a platform 21 adjacent to the storage facility; a first image analysis unit that analyzes the first image data to detect a person within the storage facility; a second image analysis unit that analyzes the second image data to detect a person within the platform and tracks a flow line of the detected person; and a falling person detection unit that determines the presence or absence of a person falling from the platform into the storage facility based on a combination of analysis results of the first image data and analysis results of the second image data.SELECTED DRAWING: Figure 1

Description

【Technical Field】 【0001】 The present invention relates to a fallen person detection system, a fallen person detection method, and a fallen person detection program that automatically detect a person who has fallen into equipment for storing a processed object. 【Background Art】 【0002】 In waste treatment-related facilities, especially in waste incineration facilities, although the number of cases is small, several accidents per year in which people fall into the garbage pit have been reported. 【0003】 When a person falls into the garbage pit, it is necessary to notify a crane operator or the like and temporarily stop the crane operation to ensure the safety of the fallen person. However, in recent years, the automatic driving technology of cranes has been developing. For example, when a fallen person occurs during automatic crane driving (that is, when the crane operator is absent), it takes time to notify a crane operator or the like, and there is a possibility of causing a serious accident. 【0004】 Moreover, especially in waste incineration facilities, there is no accident prevention system that automatically detects a fallen person in the garbage pit and controls the crane. 【0005】 In other fields (especially the railway industry), a system that detects a person who has fallen from a station platform onto the track by image processing or the like is used (see, for example, Patent Documents 1 and 2). 【Prior Art Documents】 【Patent Documents】 【0006】 【Patent Document 1】 Japanese Patent No. 5386744 【Patent Document 2】 Japanese Patent No. 4041678 【Summary of the Invention】 ​​​​​In railway industry fall detection systems, images from surveillance cameras that simultaneously capture images of the platform (the source of the fall) and the tracks (the result of the fall) are processed to detect the person who fell. In contrast, in waste incineration facilities, the platform (the source of the fall) and the waste pit (the result of the fall) are separated by a waste input door, making it difficult to install cameras or other imaging devices in a location that can simultaneously capture images of both the source and the result of the fall. 【0008】 Furthermore, in fall detection systems used in the railway industry, a method is sometimes employed that processes images from surveillance cameras to determine the size of the falling object and then decide whether or not it is a person based on that size. In contrast, in waste incineration facilities, various sizes of waste are dumped into the waste pit from transport vehicles (such as packer trucks and light trucks), making it difficult to determine whether or not a falling object is a person based solely on its size. 【0009】 Furthermore, as mentioned above, it is difficult to install cameras in waste incineration facilities in a location that can simultaneously capture images of both the fall-inducing side (platform) and the fall-resulting side (waste pit). For example, one might consider installing a camera on the waste pit side and processing the images inside the waste pit to detect the person who fell. However, within the waste pit, there is a possibility that the pile of waste may collapse, causing waste to fall on top of the person who fell. In that case, it would be difficult to detect the person who fell by processing the images inside the waste pit. 【0010】 Furthermore, this applies not only to facilities that employ the pit-and-crane system, such as incineration plants, but also to pit-and-crane facilities. Similar challenges exist in waste treatment facilities other than the net-and-crane system, such as those using direct loading into storage facilities or compactor container systems (for example, bulky waste crushing facilities and recycling facilities). 【0011】 The present invention has been made in consideration of the above points. The object of the present invention is to provide a fall detection system, a fall detection method, and a fall detection program that can automatically detect a person who has fallen into a storage facility for storing waste at a waste treatment facility such as a waste incineration facility. [Means for solving the problem] 【0012】 A fall detection system according to a first aspect of the present invention is: A first image data acquisition unit acquires first image data from a first camera that images the inside of a storage facility where the material to be processed is stored, A second image data acquisition unit acquires second image data from a second camera that images the inside of the platform adjacent to the storage facility, A first image analysis unit analyzes the first image data to detect a person inside the storage facility, A second image analysis unit analyzes the second image data to detect people on the platform and tracks the movement of the detected people. A fall detection unit determines whether or not a person has fallen from the platform into the storage facility based on a combination of the analysis results of the first image data and the analysis results of the second image data, It is equipped with. 【0013】 In this configuration, by combining the analysis results of first image data captured inside the storage facility with the analysis results of second image data captured inside the platform, it is possible to automatically detect people who have fallen from the platform into the storage facility. Generally, when detecting people who have fallen from image data using machine learning models, the detection accuracy of the machine learning models varies, and it is impossible to detect people who have fallen with 100% accuracy. However, by combining the analysis results of first image data captured inside the storage facility with the analysis results of second image data captured inside the platform, the accuracy of detecting people who have fallen can be improved. Furthermore, even in existing facilities, the system can be operated simply by installing a first camera that captures images inside the storage facility and a second camera that captures images inside the platform. Therefore, significant modifications to existing facilities are not required to introduce this system, and it is possible to improve the safety of existing facilities at a low cost. 【0014】 A fall detection system according to a second aspect of the present invention is a fall detection system according to a first aspect, The fall detection unit checks the analysis results of the second image data and, if it detects that a person has entered a predetermined area near the loading door separating the storage facility and the platform at the first time point, it checks the analysis results of the first image data at the first time point and, if it detects a person inside the storage facility, it determines that a person has fallen. 【0015】 In this configuration, the analysis results of the second image data, which was captured inside the platform, are checked. If the entry of a person into a predetermined area near the loading door is detected (i.e., a person who has fallen is tentatively detected in the analysis results of the second image data), the analysis results of the first image data are checked. If a person is detected inside the storage facility (i.e., a person who has fallen is tentatively detected in the analysis results of the first image data), a final determination is made that a person who has fallen has occurred, thereby enabling accurate detection of a person who has fallen. 【0016】 A third aspect of the present invention is a fall detection system, which is a fall detection system according to the second aspect, If the fall detection unit detects that a person has entered a predetermined area near the input door at the first time step, it checks the analysis results of the image captured from the first image data at the first time step, which is the area within the storage facility corresponding to the location of the input door. 【0017】 In this configuration, even in large storage facilities equipped with multiple loading doors, if the analysis results of the second image data captured inside the platform are checked and the entry of a person into a predetermined area near the loading door is detected (i.e., if a person who has fallen is tentatively detected in the analysis results of the second image data), instead of checking the analysis results of the entire first image data captured inside the storage facility, the analysis results of the image of the area inside the storage facility corresponding to the loading door where the person who fell is most likely to have occurred are checked, enabling more accurate and faster detection of the person who fell. 【0018】 A fall detection system according to a fourth aspect of the present invention is a fall detection system according to a second or third aspect, If the fall detection unit does not detect a person inside the storage facility after checking the analysis results of the first image data at the first time point, it extracts the difference between the first image data at the first time point and the first image data at the second time point, which is a predetermined time after the first time point has elapsed. If the extracted difference exceeds a predetermined threshold, it determines that a person has fallen; if there is no difference, it determines that no person has fallen. 【0019】 According to such an aspect, for example, when the object to be processed covers the upper part of the fallen person in the storage facility, in the analysis result of the first image data obtained by imaging the inside of the storage facility, (since the object to be processed covers the upper part of the fallen person) no person is detected in the storage facility. However, since the fallen person has already been temporarily detected in the analysis result of the second image data, further, the first image data at the first time is compared with the first image data at the second time after a predetermined time has elapsed from the first time, and the difference therebetween is extracted. When the difference between the first image data at the first time and the first image data at the second time exceeds a predetermined threshold value, it is considered that, for example, the pile of the object to be processed has collapsed in the storage facility, and it is considered that there is a possibility that the fallen person is not visible because the collapsed object to be processed covers the upper part of the fallen person. Therefore, it is determined that there is a fallen person. Thereby, even when the object to be processed covers the upper part of the fallen person in the storage facility, it is possible to automatically detect the fallen person. 【0020】 The fallen person detection system according to the fifth aspect of the present invention is the fallen person detection system according to the first aspect, The fallen person determination unit checks the analysis result of the first image data, and when a person is detected in the storage facility at the first time, checks the analysis result of the second image data up to a predetermined time before the first time, and determines that there is a fallen person when a person has dropped out of the video within a predetermined area near the loading door that partitions between the storage facility and the platform. 【0021】 According to such an aspect, by checking the analysis result of the first image data obtained by imaging the inside of the storage facility, and when a person is detected in the storage facility (that is, when a fallen person is temporarily detected in the analysis result of the first image data), checking the analysis result of the second image data, and when a person has dropped out of the video near the loading door (that is, when a fallen person is also temporarily detected in the analysis result of the second image data), by finally determining that there is a fallen person, the fallen person can be accurately detected. 【0022】 The faller detection system according to the sixth aspect of the present invention is the faller detection system according to the fifth aspect, wherein When a person is detected in the storage facility at the first time, the faller determination unit checks the analysis result of the image obtained by imaging the access door at the position corresponding to the area where the person was detected in the storage facility among the second image data up to a predetermined time before the first time. Check the analysis result of the image of the access door at the position corresponding to the area where the person was detected in the storage facility among the second image data up to a predetermined time before the first time. 【0023】 According to such an aspect, even in a large platform provided with a plurality of access doors, instead of checking the analysis result of the first image data obtained by imaging the inside of the storage facility and checking the analysis result of the entire second image data obtained by imaging the inside of the platform when a person is detected in the storage facility (that is, when a faller is temporarily detected in the analysis result of the first image data), by checking the analysis result of the image of the access door at a position where a faller is likely to occur, it is possible to detect a faller more accurately and in a shorter time. 【0024】 The faller detection system according to the seventh aspect of the present invention is the faller detection system according to the fifth or sixth aspect, wherein When the faller determination unit checks the analysis result of the second image data up to a predetermined time before the first time and a person has dropped out of the frame in the video outside a predetermined area near the access door, it determines that there is no faller. 【0025】 According to such an aspect, when a person has dropped out of the frame in the second image data obtained by imaging the inside of the platform for reasons unrelated to falling into the storage facility (for example, temporarily hidden behind a transport vehicle), it is possible to prevent misjudgment as having a faller, and the accuracy of faller detection can be improved. 【0026】 The faller detection system according to the eighth aspect of the present invention is the faller detection system according to any one of the fifth to seventh aspects, wherein The fall detection unit determines that there was no fall if, after reviewing the analysis results of the second image data up to a predetermined time before the first time, no person is detected entering a predetermined area near the input door. 【0027】 In this configuration, even if, for example, an object of similar size to a person is mistakenly detected as a person in the first image data captured inside the storage facility (i.e., a person who has fallen is mistakenly detected in the analysis results of the first image data), if no person is detected entering the vicinity of the loading door when the analysis results of the second image data captured inside the platform are checked, it is possible to determine that there is no person who has fallen, thereby preventing a false determination that there is a person who has fallen and improving the accuracy of person detection. 【0028】 A fall detection system according to the ninth aspect of the present invention is a fall detection system according to any of the first to eighth aspects, If the fall detection unit determines that there has been a fall, (1) Issue an alarm, (2) Send a control signal to the crane control device to stop the crane that is stirring or transporting the material to be processed stored in the storage facility. (3) A control signal is sent to the loading door control device to close the loading door that separates the storage facility from the platform. (4) Send a control signal to the crane control device to operate the crane and rescue the person who has fallen. (5) A control signal is transmitted to the rescue equipment control device to activate the rescue equipment installed in the storage facility and rescue the person who has fallen in. It further includes an instruction unit that performs at least one of the following processes. 【0029】 This configuration allows for the rapid rescue of those who have fallen, thereby enhancing the safety of the facility. 【0030】 A fall detection system according to the tenth aspect of the present invention is a fall detection system according to any of the first to ninth aspects, The first image analysis unit uses a first detection algorithm, which is constructed by machine learning on first training data generated by assigning artificial labels as information to areas where people or dummy figures resembling people exist in past image data of the storage facility, to detect people in the storage facility using new image data of the storage facility as input. 【0031】 A fall detection system according to the 11th aspect of the present invention is a fall detection system according to any of the 1st to 10th aspects, The second image analysis unit uses a second detection algorithm, constructed by machine learning on second training data generated by assigning artificial labels to areas where people or dummy figures resembling people exist in past image data within the platform, to detect people within the platform using new image data within the platform as input. 【0032】 A fall detection system according to the twelfth aspect of the present invention is a fall detection system according to the eleventh aspect, The second training data was generated by assigning artificial labels to areas where people or dummy figures resembling people existed in past image data within the platform, and by assigning separate artificial labels to areas where loading vehicles existed. 【0033】 In this configuration, people (e.g., workers) frequently work near loading vehicles on the platform, and there is a relationship between the location of the people and the location of the loading vehicles. Therefore, by using a second detection algorithm constructed by machine learning on training data generated by adding artificial labels to areas where people or dummy figures resembling people exist in past image data of the platform, and by adding another artificial label to areas where loading vehicles exist, it is possible to improve the accuracy of detecting people on the platform, as well as to detect loading vehicles falling. 【0034】 A fall detection system according to the 13th aspect of the present invention is a fall detection system according to the 10th aspect, The first detection algorithm includes one or more of the following: maximum likelihood classification, Boltzmann machines, neural networks, support vector machines, Bayesian networks, sparse regression, decision trees, statistical estimation using random forests, reinforcement learning, and deep learning. 【0035】 A fall detection system according to the 14th aspect of the present invention is a fall detection system according to the 11th or 12th aspect, The second detection algorithm includes one or more of the following: maximum likelihood classification, Boltzmann machines, neural networks, support vector machines, Bayesian networks, sparse regression, decision trees, statistical estimation using random forests, reinforcement learning, and deep learning. 【0036】 A fall detection system according to the 15th aspect of the present invention is a fall detection system according to any of the 1 to 14 aspects, The algorithm used by the second image analysis unit to track the movement of a person includes one or more of the following: optical flow, background subtraction method, Kalman filter, particle filter, and deep learning. 【0037】 A fall detection system according to the 16th aspect of the present invention is a fall detection system according to any of the 1 to 15 aspects, The first camera includes one or more of the following: an RGB camera, a near-infrared camera, a 3D camera, or an RGB-D camera. 【0038】 A fall detection system according to the 17th aspect of the present invention is a fall detection system according to any of the 1 to 16 aspects, The second camera includes one or more of the following: an RGB camera, a near-infrared camera, a 3D camera, or an RGB-D camera. 【0039】 A waste treatment facility according to the 18th aspect of the present invention is equipped with a fall detection system according to any of the 1 to 17 aspects. 【0040】 A method for detecting a person who has fallen according to 19 aspects of the present invention is: A step of acquiring first image data from a first camera that images the inside of a storage facility where the material to be processed is stored, The steps include acquiring a second image data from a second camera that images the inside of the platform adjacent to the storage facility, The first step involves analyzing the image data to detect a person inside the storage facility, The second image data is analyzed to detect people on the platform, and the movement of the detected people is tracked. Based on the combination of the analysis results of the first image data and the analysis results of the second image data, a step is made to determine whether or not there is a person who has fallen from the platform into the storage facility. Includes. 【0041】 A fall detection program according to a 20th aspect of the present invention is: On the computer, A step of acquiring first image data from a first camera that images the inside of a storage facility where the material to be processed is stored, The steps include acquiring a second image data from a second camera that images the inside of the platform adjacent to the storage facility, The first step involves analyzing the image data to detect a person inside the storage facility, The second image data is analyzed to detect people on the platform, and the movement of the detected people is tracked. Based on the combination of the analysis results of the first image data and the analysis results of the second image data, a step is made to determine whether or not there is a person who has fallen from the platform into the storage facility. Make it run. [Effects of the Invention] 【0042】 According to the present invention, it is possible to automatically detect people who have fallen into storage facilities for storing waste at waste treatment facilities such as waste incineration plants. [Brief explanation of the drawing] 【0043】 [Figure 1] Figure 1 is a schematic diagram showing the configuration of a waste treatment facility according to one embodiment. [Figure 2] Figure 2 is a block diagram showing the configuration of a fall detection system according to one embodiment. [Figure 3] Figure 3 is a flowchart showing a first example of a method for detecting a person who has fallen using a fall detection system according to one embodiment. [Figure 4A] Figure 4A shows an example of the analysis results of the second image data captured inside the platform. [Figure 4B] Figure 4B shows an example of the analysis results of the first image data captured inside the waste pit. [Figure 5] Figure 5 is a flowchart showing a second example of a fall detection method using a fall detection system according to one embodiment. [Figure 6A] Figure 6A shows an example of the analysis results of the first image data captured inside the waste pit. [Figure 6B] Figure 6B shows an example of the analysis results of the second image data captured inside the platform. [Modes for carrying out the invention] 【0044】 Embodiments of the present invention will be described in detail below with reference to the attached drawings. In the following description and the drawings used therein, the same reference numerals will be used for parts that can be identically configured, and redundant explanations will be omitted. 【0045】 (Configuration of waste treatment facilities) Figure 1 is a schematic diagram showing the configuration of a waste treatment facility 100 according to one embodiment. 【0046】 As shown in Figure 1, the waste treatment facility 100 includes a platform 21 where transport vehicles (such as packer trucks and light trucks) 22 loaded with waste are parked, a waste pit (storage facility) 3 where waste introduced from the platform 21 is stored, a crane 5 for agitating and transporting the waste stored in the waste pit 3, a hopper 4 into which the waste transported by the crane 5 is introduced, an incinerator 1 for incinerating the waste introduced from the hopper 4, and a waste heat boiler 2 for recovering waste heat from the exhaust gas generated in the incinerator 1. The type of incinerator 1 is not limited to a stoker furnace as shown in Figure 1, but also includes a fluidized bed furnace. Furthermore, the structure of the waste pit 3 is not limited to a single-stage pit as shown in Figure 1, but also includes a two-stage pit in which the waste pit is divided into an input section and a storage section. The waste pit 3 and the platform 21 are separated by an input door 24. Furthermore, the waste treatment facility 100 is equipped with an input door control device 20 that controls the operation of the input door 24, and a crane control device 30 that controls the operation of the crane 5. 【0047】 The waste, loaded onto the transport vehicle 22, is fed into the waste pit 3 through the input door 24 from the platform 21 and stored in the waste pit 3. The waste stored in the waste pit 3 is agitated by the crane 5 and then transported by the crane 5 to the hopper 4, which is then fed into the incinerator 1, where it is incinerated and processed. 【0048】 As shown in Figure 1, the waste treatment facility 100 is equipped with a first camera 6 for imaging the inside of the waste pit 3 and a waste identification system 40 for identifying the type of waste in the waste pit 3. 【0049】 The first camera 6 is positioned above the waste pit 3 and, in the illustrated example, is fixed to the rails of the crane 5, enabling it to image the waste stored in the waste pit 3 from above. The first camera 6 may be installed as a single unit or in multiple units. 【0050】 The first camera 6 may be an RGB camera that outputs shape and color image data of waste as an imaging result, a near-infrared camera that outputs near-infrared image data of waste as an imaging result, a 3D camera or RGB-D camera that captures three-dimensional image data of waste as an imaging result, or a combination of two or more of these. It's okay to have it. 【0051】 The waste identification system 40 acquires image data (also called first image data) from the first camera 6 that images the inside of the waste pit 3, and performs image analysis on the first image data to identify the type of waste stored in the waste pit 3. For example, the waste identification system 40 may use an identification algorithm (trained model) constructed by machine learning on training data in which the type of waste is labeled on past image data of the inside of the waste pit 3, and use new image data of the inside of the waste pit 3 as input to identify the type of waste stored in the waste pit 3. 【0052】 The waste identification system 40 generates a map showing the ratio of waste types in each area as a result of identifying the types of waste stored in the waste pit 3, and transmits it to the crane control device 30. Based on the map received from the waste identification system 40, the crane control device 30 operates the crane 5 to agitate the waste in the waste pit 3 so that the ratio of waste types becomes equal in all areas. This enables the automatic operation of the crane 5. 【0053】 Specifically, the waste identification system 40 can utilize, for example, the information processing device described in Japanese Patent Publication No. 6731680. 【0054】 As shown in Figure 1, the waste treatment facility 100 is further equipped with a second camera 23 for imaging the inside of the platform 21 and a fall detection system 10 for detecting people who fall from the platform 21 into the waste pit 3. 【0055】 The second camera 23 is positioned above the platform 21 and, in the illustrated example, is fixed to the wall of the platform 21 near the front of the input door 24, allowing it to image the inside of the platform 21 from near the front of the input door 24. The second camera 23 may be installed as a single unit or in multiple units. 【0056】 The second camera 23 may be an RGB camera that outputs shape and color image data of an object (such as a worker or a transport vehicle 22) as an imaging result, a near-infrared camera that outputs near-infrared image data of an object as an imaging result, a 3D camera or RGB-D camera that captures three-dimensional image data of an object as an imaging result, or a combination of two or more of these. 【0057】 (Configuration of the fall detection system) Next, the configuration of the fall detection system 10, which detects people falling from the platform 21 into the garbage pit 3, will be described. Figure 2 is a block diagram showing the configuration of the fall detection system 10. The fall detection system 10 may consist of a single computer, or it may consist of multiple computers that are connected to each other in a manner that allows them to communicate with one another. 【0058】 As shown in Figure 2, the fall detection system 10 includes a control unit 11, a storage unit 12, and a communication unit 13. Each unit is connected to the others so as to be able to communicate with them via a bus or network. 【0059】 Of these, the communication unit 13 is a communication interface for the first camera 6, the second camera 23, the crane control device 30, and the loading door control device 20. The communication unit 13 transmits and receives information between the first camera 6, the second camera 23, the crane control device 30, and the loading door control device 20 and the fall detection system 10. 【0060】 The storage unit 12 is a non-volatile data storage device such as a hard disk or flash memory. This is the image. The memory unit 12 stores various data handled by the control unit 11. The memory unit 12 also stores the first detection algorithm 12a1 constructed by the first model construction unit 11c1 (described later), the second detection algorithm 12a2 constructed by the second model construction unit 11c2, the first image data 12b1 acquired by the first image data acquisition unit 11a1, the second image data 12b2 acquired by the second image data acquisition unit 11a2, the first training data 12c1 generated by the first training data generation unit 11b1, and the second training data 12c2 generated by the second training data generation unit 11b2. 【0061】 The control unit 11 is a control means that performs various processes of the fall detection system 10. As shown in Figure 2, the control unit 11 includes a first image data acquisition unit 11a1, a second image data acquisition unit 11a2, a first training data generation unit 11b1, a second training data generation unit 11b2, a first model construction unit 11c1, a second model construction unit 11c2, a first image analysis unit 11d1, a first image analysis unit 11d2, a fall detection unit 11e, and an instruction unit 11f. Each of these units may be realized by a processor within the fall detection system 10 executing a predetermined program, or it may be implemented in hardware. 【0062】 Of these, the first image data acquisition unit 11a1 acquires first image data from the first camera 6 that images the inside of the garbage pit 3. The first image data may be a moving image or a series of still images. The frame rate of the first image data can be a general frame rate (around 30fps), and does not need to be particularly high, and may also be low frame rate (around 5-10fps). The metadata of the first image data includes information on the date and time of capture. The first image data 12b1 acquired by the first image data acquisition unit 11a1 is stored in the storage unit 12. 【0063】 The second image data acquisition unit 11a2 acquires second image data from the second camera 23, which images the inside of the platform 21. The second image data may be a moving image or a series of still images. The frame rate of the second image data can be a general frame rate (around 30fps), and does not need to be particularly high; it may also be low frame rate (around 5-10fps). The metadata of the second image data includes information on the date and time of capture. The second image data 12b2 acquired by the second image data acquisition unit 11a2 is stored in the storage unit 12. 【0064】 The first training data generation unit 11b1 generates first training data by labeling information about areas where a person (i.e., a person who has fallen in) or a dummy figure resembling a person, as visually identified by a skilled operator operating the waste incineration facility 100, is located, with past image data captured inside the waste pit 3 (i.e., by assigning an artificial label as information to areas where a person or a dummy figure resembling a person is located). As an example, the first training data generation unit 11b1 may generate first training data for image data captured inside the waste pit 3 after a dummy figure resembling a person has been intentionally dropped into the waste pit 3, or it may generate first training data for image data (composite image data) obtained by combining an image of a person with past image data captured inside the waste pit 3. The first training data generation unit 11b1 may generate first training data by labeling past image data captured inside the waste pit 3 with information about areas where a person (i.e., a person who has fallen in) or a dummy figure resembling a person, as visually identified by a skilled operator operating the waste incineration facility 100, is located, and information about areas where a transport vehicle 22 is located (i.e., artificial labels are assigned as information to areas where a person or a dummy figure resembling a person is located and areas where a transport vehicle 22 is located). The information about areas where a person or a dummy figure resembling a person is located and the information about areas where a transport vehicle 22 is located are labeled (i.e., artificial labels are assigned as information) by overlaying them on the image data as layers, for example. The first training data 12c1 generated by the first training data generation unit 11b1 is stored in the storage unit 12. 【0065】 The second training data generation unit 11b2 generates second training data by labeling past image data captured inside the platform 21 with information about areas where a person or a dummy figure resembling a person, identified visually by a skilled operator operating the waste incineration facility 100, is present (i.e., by assigning an artificial label as information to areas where a person or a dummy figure resembling a person is present). As an example, the second training data generation unit 11b2 may generate second training data for image data captured inside the platform 21 after a dummy figure resembling a person has been intentionally placed inside the platform 21, or it may generate second training data for image data (composite image data) obtained by combining an image of a person with past image data captured inside the platform 21. The second training data generation unit 11b2 may generate second training data by labeling past image data captured inside the waste pit 3 with information about areas where a person or a dummy figure resembling a person, identified visually by a skilled operator operating the waste incineration facility 100, is located, and information about areas where a transport vehicle 22 is located (i.e., artificial labels are assigned as information to areas where a person or a dummy figure resembling a person is located and areas where a transport vehicle is located). Within the platform 21, people (for example, workers) frequently work near the transport vehicle 22, and there is a relationship between the location of the people and the location of the transport vehicle 22. Therefore, the second model construction 11c2, described later, builds a second detection algorithm 12a2 by machine learning training data in which artificial labels are added as information to areas where people or dummy figures resembling people exist, and another artificial label is added as information to areas where the transport vehicle 22 exists, thereby improving the accuracy of detecting people within the platform 21 by the second detection algorithm 12a2, and also enabling the detection of the transport vehicle 22 falling. Information about areas where people or dummy figures resembling people exist and information about areas where the transport vehicle 22 exists are labeled (i.e., artificial labels are added as information) by overlaying them on the image data as layers.The second training data 12c2 generated by the second training data generation unit 11b2 is stored in the storage unit 12. 【0066】 The first model building unit 11c1 builds a first detection algorithm 12a1 (trained model) that detects people (i.e., people who have fallen) in the garbage pit 3 using new image data of the garbage pit 3 as input, by machine learning the first training data 12c1 stored in the memory unit 12. The first detection algorithm 12a1 may include one or more of the following: maximum likelihood classification, Boltzmann machine, neural network, support vector machine, Bayesian network, sparse regression, decision tree, statistical estimation using random forest, reinforcement learning, and deep learning. The first detection algorithm 12a1 built by the first model building unit 11c1 is stored in the memory unit 12. 【0067】 The second model building unit 11c2 builds a second detection algorithm 12a2 (trained model) that detects people on the platform 21 using new image data on the platform 21 as input, by machine learning the second training data 12c2 stored in the memory unit 12. The second detection algorithm 12a2 may include one or more of the following: maximum likelihood classification, Boltzmann machine, neural network, support vector machine, Bayesian network, sparse regression, decision tree, statistical estimation using random forest, reinforcement learning, and deep learning. The second detection algorithm 12a2 built by the second model building unit 11c2 is stored in the memory unit 12. 【0068】 The first image analysis unit 11d1 analyzes the first image data acquired by the first image data acquisition unit 11a1 to detect people inside the garbage pit 3. Specifically, for example, the first image analysis unit 11d uses the first detection algorithm 12a1 (trained model) constructed by the first model construction unit 11c1 to take new image data from inside the garbage pit 3 as input. The system detects people inside the garbage pit 3. As one variation, the first image analysis unit 11d1 may perform image analysis on the first image data acquired by the first image data acquisition unit 11a1 to detect people inside the garbage pit 3 and transport vehicles 22. Specifically, for example, the first image analysis unit 11d may use the first detection algorithm 12a1 (trained model) constructed by the first model construction unit 11c1 to take new image data from the garbage pit 3 as input and detect people inside the garbage pit 3 and transport vehicles 22, respectively. 【0069】 The first image analysis unit 11d1 may divide the surface of the garbage pit 3 into multiple blocks, input the new image data from within the garbage pit 3 into the first detection algorithm 12a1 (trained model) on a block-by-block basis, and obtain detection results for people (or people and transport vehicles 22) on a block-by-block basis. This makes it possible to accurately determine where within the garbage pit 3 a person (or a person and transport vehicle 22) fell. 【0070】 The second image analysis unit 11d2 analyzes the second image data acquired by the second image data acquisition unit 11a2 to detect people on the platform 21 and tracks the movement of the detected people. Specifically, for example, the second image analysis unit 11d uses the second detection algorithm 12a2 (trained model) constructed by the second model construction unit 11c2 to detect people on the platform 21 using new image data of the platform 21 as input. Next, the second image analysis unit 11d tracks the detected people and detects their entry into a predetermined area near the input door 24. As an example, the second image analysis unit 11d2 analyzes the second image data acquired by the second image data acquisition unit 11a2 to detect people and transport vehicles 22 on the platform 21, and tracks the movement of the detected people and transport vehicles 22, respectively. Specifically, for example, the second image analysis unit 11d uses the second detection algorithm 12a2 (trained model) constructed by the second model construction unit 11c2 to take new image data from within the platform 21 as input and detects people and transport vehicles 22 within the platform 21, respectively. Next, the second image analysis unit 11d tracks the detected people and transport vehicles 22, respectively, and detects their entry into a predetermined area near the input door 24. The algorithm used for tracking may include one or more of the following: optical flow, background subtraction method, Kalman filter, particle filter, and deep learning. 【0071】 As one variation, the second image analysis unit 11d2 may perform image analysis on the second image data acquired by the second image data acquisition unit 11a2 to detect people on the platform 21, track the movement of the detected people, and detect whether the detected people are wearing safety equipment (safety belts or helmets) through image processing, and determine whether or not there are people working near the loading door 24 without wearing safety equipment. If it is determined that there are people working near the loading door 24 without wearing safety equipment, the instruction unit 11f, which will be described later, may issue an alarm, or it may send a control signal to the loading door control device 20 to prevent the loading door 24 from opening (if it is closed) or to close it (if it is open). 【0072】 The fall detection unit 11e determines whether or not a person has fallen from the platform 21 into the garbage pit 3 based on a combination of the analysis results of the first image data by the first image analysis unit 11d1 and the analysis results of the second image data by the second image analysis unit 11d2. Generally, when detecting a person who has fallen from image data using a machine learning model, the detection accuracy of the machine learning model varies, and it is impossible to detect a person who has fallen with 100% accuracy. However, in this embodiment, the accuracy of detecting a person who has fallen can be improved by combining the analysis results of the first image data captured inside the garbage pit 3 and the analysis results of the second image data captured inside the platform 21. 【0073】 As an example, the fall detection unit 11e first checks the analysis results of the second image data, as shown in Figure 4. As shown in A, if entry of a person into a predetermined area near the input door 24 is detected at the first time point (i.e., if a person who has fallen is tentatively detected in the analysis results of the second image data), the analysis results of the first image data at the first time point may be checked, and as shown in Figure 4B, if a person is detected inside the garbage pit 3 (i.e., if a person who has fallen is tentatively detected in the analysis results of the first image data), a final determination may be made that a person has fallen. This makes it possible to accurately detect a person who has fallen. 【0074】 As shown in Figure 4A, if the fall detection unit 11e detects that a person has entered a predetermined area near the input door 25 (input door B in the illustrated example) at the first time step, it may, as shown in Figure 4B, check the analysis results of the image of the area within the garbage pit 3 corresponding to the position of the input door (i.e., input door B) (the area enclosed by the dashed line labeled B1 in Figure 4B) from the first image data at the first time step. This allows for more accurate and faster detection of a fall even in a large garbage pit 3 with multiple input doors 24. This is achieved by checking the analysis results of the image of the area B1 within the garbage pit 3 corresponding to the position of the input door B where the fall is most likely to have occurred, rather than checking the analysis results of the second image data captured inside the platform 21 and then checking the analysis results of the entire first image data captured inside the garbage pit 3 when a person has entered a predetermined area near input door B. 【0075】 Furthermore, if the fall detection unit 11e confirms the analysis result of the first image data at the first time point and no person is detected in the garbage pit 3, it may compare the first image data at the first time point with the first image data at the second time point, which is a predetermined time after the first time point (for example, 5 minutes later), and extract the difference between them. If the difference between the first image data at the first time point and the first image data at the second time point exceeds a predetermined threshold, it may determine that a person has fallen in. If there is no difference, it may determine that no person has fallen in. The reason for this is as follows: For example, if waste covers the person who has fallen in the garbage pit 3, the analysis result of the first image data will indicate that no person was detected in the garbage pit 3 (because waste covers the person who has fallen in). However, since a person has already been tentatively detected in the analysis result of the second image data, the first image data at the first time point and the first image data at the second time point are further compared and the difference between them is extracted. If the difference between the first image data at the first time point and the first image data at the second time point exceeds a predetermined threshold, it is considered that the pile of waste in the waste pit 3 has collapsed, and it is possible that the collapsed pile of waste has covered the person who fell, making the person invisible. Therefore, it is determined that a person has fallen. This makes it possible to automatically detect a person who has fallen even if waste has covered the person who has fallen in the waste pit 3. 【0076】 As one variation, the fall detection unit 11e may first check the analysis results of the first image data, and if a person is detected in the garbage pit 3 at the first time (i.e., if a fall is tentatively detected in the analysis results of the first image data), as shown in Figure 6A, it may then check the analysis results of the second image data up to a predetermined time before the first time (for example, 5 minutes before), and if, as shown in Figure 6B, the person has moved out of frame in the video within a predetermined area near the input door 24 (i.e., if a fall is tentatively detected in the analysis results of the second image data), it may make a final determination that a fall has occurred. This makes it possible to accurately detect a fall. 【0077】 As shown in Figure 6A, if a person is detected in the garbage pit 3 at the first time, the person detection unit 11e may, as shown in Figure 6B, check the analysis results of the image of the input door (input door B in the illustrated example) at a position corresponding to the area in the garbage pit 3 where a person was detected (the area enclosed by the dashed line labeled B2 in Figure 6A) from the second image data up to a predetermined time before the first time (for example, 5 minutes before). Even with a large platform 21 equipped with multiple doors 24, if a person is detected inside the garbage pit 3 after checking the analysis results of the first image data captured inside the garbage pit 3, instead of checking the analysis results of the entire second image data captured inside the platform 21, the analysis results of the image of the input door B, which is located at a position where it is highly likely that a person fell, can be checked to enable more accurate and faster detection of the person who fell. 【0078】 Furthermore, the fall detection unit 11e may determine that there is no fall if, after checking the analysis results of the second image data up to a predetermined time before the first time (for example, 5 minutes before), the person is out of frame in the video outside a predetermined area near the input door 24. This prevents a false determination that there is a fall in the second image data captured inside the platform 21 when the person is out of frame for reasons unrelated to falling into the garbage pit 3 (for example, being temporarily hidden behind the transport vehicle 22), thereby improving the accuracy of fall detection. 【0079】 Furthermore, the fall detection unit 11e may determine that there is no fall if, upon reviewing the analysis results of the second image data up to a predetermined time before the first time (for example, 5 minutes before), no entry of a person into a predetermined area near the input door 24 is detected. This prevents a false determination of a fall, even if, for example, a piece of waste roughly the same size as a person is mistakenly detected as a person in the first image data captured inside the garbage pit 3 (i.e., a fall is erroneously tentatively detected in the analysis results of the first image data), by determining that there is no fall if, upon reviewing the analysis results of the second image data captured inside the platform 21, no entry of a person into the vicinity of the input door 24 is detected, thereby improving the accuracy of fall detection. 【0080】 The instruction unit 11f checks the result of the fall detection unit 11e, and if the fall detection unit 11e determines that there is a fall, (1) Issue an alarm, (2) A control signal is sent to the crane control device 30 to stop the crane 5 which is used to agitate or transport the waste stored in the waste pit 3. (3) A control signal is sent to the input door control device 20 to close the input door 24 that separates the waste pit 3 and the platform 21. (4) A control signal is sent to the crane control device 30 to operate the crane 5 and rescue the person who has fallen. (5) A control signal is sent to the rescue equipment control device (not shown) to activate the rescue equipment (not shown) installed in the garbage pit 3 and rescue the person who has fallen in. Perform at least one of the following processes. This will enable the rapid rescue of a person who falls while crane 5 is operating automatically (i.e., when the crane operator is absent), thereby enhancing the safety of the facility. 【0081】 (Example 1 of a method for detecting a person who has fallen) Next, we will describe a first example of a method for detecting a person who has fallen using the fall detection system 10 with the above configuration. Figure 3 is a flowchart showing the first example of the fall detection method. 【0082】 As shown in Figure 3, first, the first image data acquisition unit 11a1 acquires first image data from the first camera 6 that images the inside of the waste pit 3 (step S10). The acquired first image data 12b1 is stored in the storage unit 12. 【0083】 Next, the second image data acquisition unit 11a2 acquires second image data from the second camera 23 that images the inside of the platform 21 (step S11). The acquired first image data 12b2 is stored in the storage unit 12. Note that the order of steps S10 and S11 is... Either one can come first, or they can happen simultaneously. 【0084】 Next, the second image analysis unit 11d2 analyzes the second image data acquired by the second image data acquisition unit 11a2 to detect a person on the platform 21 and tracks the movement of the detected person (step S12). 【0085】 The fall detection unit 11e confirms the analysis results of the second image data by the second image analysis unit 11d2 (step S13). 【0086】 If no person is detected entering a predetermined area near the input door 24 at the first time step (i.e., no person is provisionally detected in the analysis results of the second image data) (step S13: NO), the person fall detection unit 11e determines that there is no person who has fallen (step S19). 【0087】 On the other hand, as shown in Figure 4A, if entry of a person into a predetermined area near the input door 24 is detected at the first time (i.e., if a person who has fallen is tentatively detected in the analysis results of the second image data) (Step S13: YES), the first image analysis unit 11d1 analyzes the first image data at the first time acquired by the first image data acquisition unit 11a1 to detect a person inside the garbage pit 3 (Step S14). 【0088】 Then, the fall detection unit 11e confirms the analysis result of the first image data at the first time point by the first image analysis unit 11d1 (step S15). Here, as shown in Figure 4A, if the fall detection unit 11e detects that a person has entered a predetermined area near the input door 25 (input door B in the illustrated example) at the first time point, it may also confirm the analysis result of the image captured from the first image data at the first time point, which is the area inside the garbage pit 3 corresponding to the position of the input door (i.e., input door B) (the area enclosed by the dashed line labeled B1 in Figure 4B). 【0089】 As shown in Figure 4B, if a person is detected in the garbage pit 3 (i.e., if a person who has fallen is tentatively detected in the analysis results of the first image data) (Step 15: YES), the person who has fallen determination unit 11e determines that there is a person who has fallen (Step S17). 【0090】 On the other hand, if no person is detected in the garbage pit 3 (i.e., if no person is tentatively detected in the analysis results of the first image data) (step 15: NO), the person who fell in 11e extracts the difference between the first image data at the first time and the first image data at the second time, which is a predetermined time after the first time (for example, 5 minutes later), and compares the extracted difference with a predetermined threshold (step S16). 【0091】 If the difference between the first image data at the first time point and the first image data at the second time point exceeds a predetermined threshold (Step S16: YES), it is considered that the pile of waste has collapsed in the waste pit 3, and that the collapsed pile of waste may have covered the person who fell, making the person invisible. Therefore, the person fall detection unit 11e determines that there is a person who has fallen (Step S17). 【0092】 On the other hand, if the difference between the first image data at the first time point and the first image data at the second time point does not exceed a predetermined threshold (step S16: NO), the fall detection unit 11e determines that there is no fall (step S19). 【0093】 Then, if the fall detection unit 11e determines that there is a fall (after step S17), the instruction unit 11f will issue an alarm to notify other workers, A control signal is sent to the crane control device 30 to stop the crane 5 (step S18). 【0094】 In step S18, the instruction unit 11f may send a control signal to the input door control device 20 to close the input door 24 that separates the waste pit 3 and the platform 21, in order to prevent waste from being dumped on the fallen person and making rescue difficult. Alternatively, instead of sending a control signal to the crane control device 30 to stop the crane 5, the instruction unit 11f may send a control signal to the crane control device 30 to operate the crane 5 and rescue the fallen person. Alternatively, the instruction unit 11f may send a control signal to the rescue equipment control device (not shown) to operate the rescue equipment (not shown) installed in the waste pit 3 and rescue the fallen person. This makes it possible to quickly rescue the fallen person even if the fall occurs while the crane 5 is operating automatically (i.e., when the crane operator is absent). 【0095】 (Second example of a method for detecting a person who has fallen) Next, a second example of a method for detecting a person who has fallen using the fall detection system 10 will be described. Figure 5 is a flowchart showing the second example of the fall detection method. 【0096】 As shown in Figure 5, first, the first image data acquisition unit 11a1 acquires first image data from the first camera 6 that images the inside of the waste pit 3 (step S20). The acquired first image data 12b1 is stored in the storage unit 12. 【0097】 Next, the second image data acquisition unit 11a2 acquires second image data from the second camera 23 that images the inside of the platform 21 (step S21). The acquired first image data 12b2 is stored in the storage unit 12. Note that the order of step S20 and step S21 does not matter, and they may occur simultaneously. 【0098】 Next, the first image analysis unit 11d1 analyzes the first image data acquired by the first image data acquisition unit 11a1 to detect a person inside the garbage pit 3 (step S22). 【0099】 The fall detection unit 11e confirms the analysis results of the first image data by the first image analysis unit 11d1 (step S23). 【0100】 If no person is detected in the garbage pit 3 at the first time (i.e., no person is tentatively detected in the analysis results of the first image data) (step S23: NO), the person who fell in the garbage pit 11e determines that there is no person who fell in (step S29). 【0101】 On the other hand, as shown in Figure 6A, if a person is detected in the garbage pit 3 at the first time (i.e., if a person who has fallen is tentatively detected in the analysis results of the first image data) (step S23: YES), the second image analysis unit 11d2 analyzes the second image data acquired by the second image data acquisition unit 11a2 up to a predetermined time before the first time (for example, 5 minutes before) to detect a person on the platform 21 and tracks the movement of the detected person (step S24). 【0102】 Then, the fall detection unit 11e checks whether or not the entry of a person into a predetermined area near the input door 24 has been detected in the analysis results of the second image data by the second image analysis unit 11d2 (step S25). Here, as shown in Figure 6A, if a person is detected in the garbage pit 3 at the first time, the fall detection unit 11e analyzes the image of the input door (input door B in the illustrated example) at a position corresponding to the area in the garbage pit 3 where a person was detected (the area enclosed by the dashed line labeled B2 in Figure 6A) from the second image data up to a predetermined time before the first time (for example, 5 minutes before), as shown in Figure 6B. You may review the analysis results. 【0103】 If no person is detected entering a predetermined area near the input door 24 (i.e., if no person is tentatively detected in the analysis results of the second image data) (step S25: NO), the person falling determination unit 11e determines that there is no person who has fallen (step S29). 【0104】 On the other hand, as shown in Figure 6A, if the analysis of the second image data detects that a person has entered a predetermined area near the input door 24 (step S25: YES), the fall detection unit 11e checks whether the person had moved out of frame in the video within the predetermined area near the input door 24 (step S26). 【0105】 If a person is out of frame in the video within a predetermined area near the input door 24 (i.e., if a person who has fallen is tentatively detected in the analysis results of the second image data) (Step S26: YES), the person who has fallen determination unit 11e determines that a person has fallen (Step S27). 【0106】 On the other hand, if no person has moved out of frame within a predetermined area near the input door 24 (step S26: NO), the fall detection unit 11e determines that no one has fallen (step S29). 【0107】 Then, if the fall detection unit 11e determines that there is a fall (after step S27), the instruction unit 11f issues an alarm to notify other workers and sends a control signal to the crane control device 30 to stop the crane 5 (step S28). 【0108】 In step S28, the instruction unit 11f may send a control signal to the input door control device 20 to close the input door 24 that separates the waste pit 3 and the platform 21, in order to prevent waste from being dumped on the fallen person and making rescue difficult. Alternatively, instead of sending a control signal to the crane control device 30 to stop the crane 5, the instruction unit 11f may send a control signal to the crane control device 30 to operate the crane 5 and rescue the fallen person. Alternatively, the instruction unit 11f may send a control signal to the rescue equipment control device (not shown) to operate the rescue equipment (not shown) installed in the waste pit 3 and rescue the fallen person. This makes it possible to quickly rescue the fallen person even if the fall occurs while the crane 5 is operating automatically (i.e., when the crane operator is absent). 【0109】 Incidentally, in waste incineration facilities, the platform 21, which is the cause of the fall, and the waste pit 3, which is the result of the fall, are separated by a waste input door 24. Therefore, it is difficult to install imaging devices such as cameras in a location that can simultaneously photograph both the cause of the fall and the result of the fall, as is done in fall detection systems used in the railway industry. 【0110】 In contrast, according to this embodiment, even without installing a camera in a location where both the fall-causing side and the fall-result side can be photographed simultaneously, first image data is acquired from a first camera 6 that images the inside of the garbage pit 3, which is the fall-result side, and second image data is acquired from a second camera 23 that images the inside of the platform 21, which is the fall-causing side. By combining the analysis results of the first image data captured inside the garbage pit 3 and the analysis results of the second image data captured inside the platform 21, it is possible to automatically detect a person who has fallen from the platform 21 into the garbage pit 3. 【0111】 Generally, when using machine learning models to detect people who have fallen from image data, the detection accuracy of the machine learning models varies, and it is impossible to detect people who have fallen with 100% accuracy. However, according to this embodiment, the accuracy of detecting people who have fallen can be improved by combining the analysis results of the first image data captured inside the garbage pit 3 with the analysis results of the second image data captured inside the platform 21. 【0112】 Furthermore, according to this embodiment, even in existing facilities, the system can be operated simply by installing a first camera 6 that images the inside of the waste pit 3 and a second camera 23 that images the inside of the platform 21. Therefore, significant modifications to existing facilities are not required to introduce this system, and it is possible to improve the safety of existing facilities at a low cost. 【0113】 In the above-described embodiment, the waste treatment facility 100 equipped with the fall detection system 10 is also equipped with a waste identification system 40, and the crane 5 is configured to be automatically operated based on the identification results of the waste identification system 40. However, the embodiment is not limited to this, and the waste treatment facility 100 equipped with the fall detection system 10 may also be a facility that is not equipped with a waste identification system 40. 【0114】 Furthermore, in the embodiment described above, the fall detection system 10 was configured to determine the presence or absence of a fall based on a combination of the analysis results of first image data captured inside the garbage pit 3 and the analysis results of second image data captured inside the platform 21. However, if sufficient detection accuracy can be obtained from the analysis results of the first image data alone, the presence or absence of a fall may be determined based solely on the analysis results of the first image data. Similarly, if sufficient detection accuracy can be obtained from the analysis results of the second image data alone, the presence or absence of a fall may be determined based solely on the analysis results of the second image data. 【0115】 Although embodiments and modifications of the present invention have been described above by example, the scope of the present invention is not limited thereto, and it is possible to modify and transform it according to the purpose within the scope described in the claims. Furthermore, each embodiment and modification can be appropriately combined as long as the processing content is not contradictory. 【0116】 Furthermore, although the fall detection system 10 according to this embodiment may be composed of one or more computers, the program for implementing the fall detection system 10 on one or more computers and the recording medium on which the program is non-temporarily recorded are also subject to protection in this case. [Explanation of symbols] 【0117】 1 Incinerator 2 Combustion device 3. Garbage pit 4 Hoppers 5 Cranes 6. Camera 1 10. Fall detection system 11 Control Unit 11a1 First image data acquisition unit 11a2 Second image data acquisition unit 11b1 First Training Data Generation Unit 11b2 Second Training Data Generation Unit 11c1 First Model Construction Section 11c2 Second Model Construction Section 11d1 First Image Analysis Unit 11d2 2nd image analysis section 11e Falling Person Determination Section 11f Instruction section 12 Storage section 12a1 First detection algorithm 12a2 Second detection algorithm 12b1 Image Data 1 12b2 Second image data 12c1 First training data 12c2 Second Training Data 13 Communications Department 20. Input Door Control Device 21 Platforms 22 Transport Vehicles 23. Second camera 24 Input Door 30 Crane control device 40. Waste Identification System 100 Waste Treatment Facilities

Claims

[Claim 1] An image data acquisition unit acquires image data from a camera that captures images of the platform adjacent to the storage facility where the materials to be processed are stored, The system includes an image analysis unit that analyzes the aforementioned image data to detect people on the platform and tracks the movement of the detected people. The image analysis unit uses a detection algorithm constructed by machine learning on training data generated by artificially labeling areas where people or dummy figures resembling people exist in past image data of the platform. Using this algorithm, it takes new image data of the platform as input to detect people on the platform and also detects whether the detected people are wearing safety equipment through image processing. A fall detection system characterized by the following features. [Claim 2] The image analysis unit tracks the movement of the detected person and detects the person entering a predetermined area near the loading door that separates the storage facility and the platform. The fall detection system according to claim 1. [Claim 3] The image analysis unit determines whether or not there is a person working near the loading door without wearing safety equipment. The fall detection system according to claim 1 or 2. [Claim 4] The fall detection system according to any one of claims 1 to 3, wherein the safety equipment includes a safety harness and / or helmet. [Claim 5] If the image analysis unit determines that there is a person working near the loading door without wearing safety equipment, (1) Issue an alarm, (2) If the loading door separating the storage facility and the platform is closed, a control signal is sent to the loading door control device to prevent it from opening. (3) If the loading door separating the storage facility and the platform is open, send a control signal to the loading door control device to close it. The fall detection system according to claim 3, further comprising an instruction unit that performs at least one of the following processes. [Claim 6] The fall detection system according to any one of claims 1 to 5, wherein the detection algorithm includes one or more of the following: maximum likelihood classification, Boltzmann machine, neural network, support vector machine, Bayesian network, sparse regression, decision tree, statistical estimation using random forest, reinforcement learning, and deep learning. [Claim 7] The fall detection system according to any one of claims 1 to 6, wherein the algorithm used by the image analysis unit to track the movement of a person includes one or more of the following: optical flow, background subtraction method, Kalman filter, particle filter, and deep learning. [Claim 8] The camera includes one or more of the following: an RGB camera, a near-infrared camera, a 3D camera, or an RGB-D camera. A fall detection system according to any one of claims 1 to 7. [Claim 9] A waste treatment facility equipped with a fall detection system according to any one of claims 1 to 8. [Claim 10] The steps include acquiring image data from a camera that images the inside of a platform adjacent to the storage facility where the material to be processed is stored, The steps include: analyzing the aforementioned image data to detect people on the platform and tracking the movements of the detected people; The steps include detecting the status of the safety equipment worn by the detected person by image processing, A method for detecting a person who has fallen, including the method described above. [Claim 11] On the computer, The steps include acquiring image data from a camera that images the inside of a platform adjacent to the storage facility where the material to be processed is stored, The steps include: analyzing the aforementioned image data to detect people on the platform and tracking the movements of the detected people; The steps include detecting the status of the safety equipment worn by the detected person by image processing, A fall detection program that triggers the execution of the program.

Images