Fire detection device, fire detection system, and computer program

Abstract

This technology allows for fire suppression and a reduction in the frequency of firefighting operations at waste disposal facilities. [Solution] The fire detection device 6 comprises a camera (first camera 71) that photographs waste in the waste processing process at a waste treatment facility (crushing and sorting equipment 1), and a detection unit (first detection unit 632) that detects the ignition of specific waste based on the color information of the image captured by the camera.

Description

Technical Field

[0001] The technology disclosed herein relates to an apparatus, a system, and a computer program for detecting the ignition of specific waste in a waste treatment facility.

Background Art

[0002] Measures for preventing fires in waste treatment facilities have been conventionally taken. Currently, the causative object of fires in waste treatment facilities is often a lithium-ion battery contained in the waste processed in the waste treatment facility. For example, as described in Patent Document 1, a lithium-ion battery generates heat due to a short circuit caused by an external impact such as crushing or compression, leading to ignition. Since the flame of a lithium-ion battery continues to appear, there is a risk that the flame will spread to the surrounding waste. Compared with the case where waste other than a lithium-ion battery catches fire, the ignition of a lithium-ion battery is highly likely to lead to a fire.

[0003] Patent Document 2 describes a fire detection system for a waste treatment facility. This system detects ignition during the conveyance of crushed waste. Specifically, this fire detection system includes a first conveyor that conveys the crushed waste discharged from a crusher and a second conveyor that conveys the crushed waste dropped from the first conveyor. In the fire detection system, an infrared / visible composite camera images the crushed waste dropped from the first conveyor to the second conveyor, and it is determined whether an ignition phenomenon has occurred based on the captured image.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0005] In the fire detection system described in Patent Document 2, if a fire is detected, waste disposal operations are stopped, and the crushed waste being transported is dumped into a hazardous materials storage tank containing cooling water to extinguish the fire. However, for these fire extinguishing operations and post-fire treatment, the fire detection system in Patent Document 2 requires the waste disposal facility to be shut down every time a fire is detected, which disrupts the operation of the facility. While there is a demand to minimize the frequency of unnecessary fire extinguishing operations by not performing fire extinguishing operations if a fire does not result, there is also a need to effectively suppress fires in waste disposal facilities.

[0006] The technology disclosed herein suppresses fires in waste treatment facilities while maintaining the operational efficiency of the facilities. [Means for solving the problem]

[0007] The inventors of this application focused on the fact that fires in waste treatment facilities include fires caused by specific types of waste that are prone to ignition, and fires caused by waste other than these specific types that are less likely to cause fires. The fire detection device disclosed herein distinguishes and detects fires caused by specific types of waste that are prone to ignition, and fires caused by other types of waste. The fire detection device can selectively detect fires that require firefighting, thereby suppressing fires in waste treatment facilities while reducing the frequency of unnecessary firefighting operations and maintaining the operational efficiency of the facility.

[0008] The technology disclosed herein relates to a device for detecting ignition in a waste treatment facility. The device comprises a photographing device for photographing waste in the waste treatment process at the waste treatment facility, and a detection unit for detecting ignition of specific waste based on the color information of the image captured by the photographing device.

[0009] The imaging device photographs waste during the waste treatment process at a waste treatment facility. This treatment includes the crushing of waste. It also includes the transportation of waste before or after crushing. However, the treatment is not limited to crushing or transportation.

[0010] The detection unit detects the ignition of specific waste based on the color information of the image captured by the imaging device. Color includes elements of hue, lightness, and saturation. For example, the color of the flame when a specific type of waste that is prone to fire burns may be different from the color of the flame when other types of waste burn, and may have a unique color. The detection unit can detect the ignition of specific waste based on the color information of the image.

[0011] Fire detection devices, for example, detect the ignition of specific waste materials that are prone to causing fires, thus reducing the frequency of detecting fires that do not require firefighting and maintaining the operational efficiency of the facility. Furthermore, because fire detection devices detect fires that do require firefighting, necessary firefighting operations can be carried out quickly at waste treatment facilities. This helps to suppress fires at waste treatment facilities.

[0012] Another fire detection device disclosed herein comprises, in the waste processing process at a waste treatment facility, a photographing device for photographing the waste, and a detection unit for distinguishing and detecting the ignition of specific waste based on the color information of the image captured by the photographing device.

[0013] The detection unit can detect the ignition of a specific type of waste based on the color information of the image, distinguishing it from the ignition of other types of waste.

[0014] The fire detection device can distinguish between fires caused by specific types of waste that are prone to ignition and fires caused by other types of waste, thereby reducing the frequency of detecting fires that do not require firefighting and maintaining the operational efficiency of the facility. Furthermore, because the fire detection device detects fires that do require firefighting, necessary firefighting operations can be carried out quickly at waste treatment facilities. This helps to suppress fires at waste treatment facilities.

[0015] The detection unit may detect that the specific waste has ignited if the image contains a predetermined number or more cells of the color corresponding to the flame reaction of the specific waste.

[0016] Here, a "cell" is one of the regions into which the image of the imaging device has been pre-divided. An image pixel is a type of "cell," and the term "cell" here may also refer to a pixel. Furthermore, a specific type of waste is an object that exhibits a flame test. The detection unit can detect the ignition of a specific type of waste by detecting cells of the color corresponding to the flame test. The detection unit also detects the ignition of a specific type of waste based on the fact that the image contains a predetermined number or more cells of the color corresponding to the flame test, that is, that the area of ​​the color corresponding to the flame test is large in the image. The detection unit can detect the ignition of a specific type of waste with high accuracy.

[0017] The detection unit may encode the colors of the image using a predetermined color space and then determine the color corresponding to the flame reaction of the specific waste.

[0018] The specified color space can be any of the various known color spaces, such as RGB (Red, Green, Blue), HSV (Hue, Saturation, Value), HSL (Hue, Saturation, Lightness), or L*a*b* (CIELAB).

[0019] By encoding the colors in the image, the detection unit can accurately determine the color corresponding to the flame test of a specific type of waste. The detection unit can accurately detect ignition of a specific type of waste.

[0020] The aforementioned specific waste is a lithium-ion battery, The detection unit may determine the color corresponding to the flame test of lithium ions.

[0021] As described above, a lithium-ion battery is an example of waste that is likely to lead to a fire. The color of the flame reaction of lithium ions is deep red, and there is no waste among the waste processed in the waste treatment facility that shows a flame reaction of a similar color. The detection unit can accurately detect the ignition of the lithium-ion battery based on the color information of the image.

[0022] The imaging device may be configured to image the waste being crushed in the crusher.

[0023] In the crusher, an external force is applied to the waste. As the external force is applied, various wastes, such as spray cans, etc., may catch fire in the crusher. As described above, the detection unit can accurately detect the ignition of a specific waste based on the color information of the image. The detection unit can quickly detect the ignition of a specific waste in the crusher where ignition is likely to occur. The quick detection of ignition enables the quick extinguishing of the waste. If the waste is quickly extinguished, the stoppage of the operation of the waste treatment facility is suppressed to a minimum.

[0024] The fire detection device may further include an output unit that outputs a signal for executing water spraying to the fire extinguishing equipment of the crusher when the detection unit detects the ignition of the specific waste in the crusher.

[0025] When the detection unit detects the ignition of a specific waste in the crusher, the fire extinguishing equipment receives the signal for executing water spraying and sprays water in the crusher where the ignition has been detected. The ignited waste is quickly extinguished.

[0026] The fire detection device may further include a second detection unit that has a learned model and detects the ignition of the waste based on the image captured by the imaging device and the learned model, and a determination unit that determines the ignition of the specific waste when each of the detection unit and the second detection unit detects ignition.

[0027] The second detection unit detects ignition of waste, regardless of whether it is a specific type of waste or not. The determination unit determines that the ignition is of a specific type of waste when both the detection unit and the second detection unit detect ignition. The ignition detection device can detect ignition of specific waste with greater accuracy.

[0028] The system disclosed herein is a system for detecting ignition in the waste processing process at a waste treatment facility, comprising: an input unit that receives images from a camera that photographs the waste in the processing process; and a detection unit that detects ignition of specific waste based on the color information of the images.

[0029] This system can accurately detect when specific waste materials have ignited at a waste disposal facility.

[0030] The computer program disclosed herein is a computer program that causes a computer to perform a process for detecting ignition in the waste processing process at a waste treatment facility, and causes the computer to perform an input process that receives an image from a camera that photographs the waste in the processing process, and a detection process that detects ignition of a specific waste based on the color information of the image.

[0031] This computer program can accurately detect when specific waste materials have ignited in a waste disposal facility. [Effects of the Invention]

[0032] As described above, the aforementioned fire detection device, fire detection system, and computer program can suppress fires in waste treatment facilities while reducing the frequency of unnecessary firefighting operations. [Brief explanation of the drawing]

[0033] [Figure 1] Figure 1 is a schematic diagram of the waste treatment facility. [Figure 2] Figure 2 is a block diagram illustrating an example of a fire detection device. [Figure 3] Figure 3 illustrates the difference between an image captured by a camera and an image identifying cells corresponding to the color of the lithium-ion flame test. [Figure 4] Figure 4 is a flowchart illustrating the control procedure for a fire detection device. [Figure 5] Figure 5 is a flowchart illustrating the control procedure of a modified fire detection device. [Figure 6] Figure 6 is a flowchart illustrating the control procedure of a modified fire detection device. [Figure 7] Figure 7 is a flowchart illustrating the control procedure of a modified fire detection device. [Figure 8] Figure 8 is a block diagram illustrating a modified fire detection device. [Modes for carrying out the invention]

[0034] The following describes embodiments of the fire detection device, fire detection system, and computer program with reference to the drawings. The fire detection device and other components described herein are illustrative examples.

[0035] (Configuration of waste disposal facilities) Figure 1 illustrates the configuration of a waste treatment facility. The waste treatment facility in Figure 1 is a crushing and sorting facility 1 for non-combustible waste and non-combustible bulky waste. The ignition detection device 6 (see Figure 2) described here is installed in the crushing and sorting facility 1. However, the ignition detection device 6 is not limited to being installed in the crushing and sorting facility 1.

[0036] The crushing and sorting processing facility 1 is broadly divided into a waste pit 11, a sorting room 12, a crushing machine room 13, and a sorting device 14. The locations of the waste pit 11, sorting room 12, crushing machine room 13, and sorting device 14 shown in Figure 1 do not accurately represent the relative positions of these waste pits 11, sorting rooms 12, crushing machine rooms 13, and sorting devices 14.

[0037] The waste pit 11 stores non-combustible waste and non-combustible bulky waste, i.e., waste, collected from various locations. The dumping box 101 installed in the yard 10 rotates to deposit the waste into the waste pit 11. Although not shown in the diagram, collection trucks that collect waste from various locations may also directly deposit the waste into the waste pit 11.

[0038] The waste crane 15 transports the waste from the waste pit 11 to the conveyor 121 in the sorting room 12. The conveyor 121 transports the waste to the bag-breaking machine 122. The bag-breaking machine 122 breaks open the bags containing the waste. The waste that has passed through the bag-breaking machine 122 is placed on the sorting conveyor 123.

[0039] The sorting conveyor 123 is equipped with a belt wrapped around a pulley. When the pulley rotates, the belt moves. Waste placed on the belt is transported along the belt. While the sorting conveyor 123 is transporting the waste, the waste is sorted manually by workers. Workers remove bottles, spray cans, unsuitable or difficult-to-process materials from the sorting conveyor 123 (see arrows in Figure 1). Even if lithium-ion batteries are included in the waste, they are basically removed from the sorting conveyor 123 by workers in the sorting room 12. The waste that has passed through the sorting conveyor 123 is sent to the crushing room 13. Lithium-ion batteries may be included in the waste that is sent from the sorting room 12 to the crushing room 13 due to being missed during sorting in the sorting room 12.

[0040] Waste sent from the sorting room 12 to the crushing room 13 is fed into the receiving hopper 211 of the first conveyor 21. The first conveyor 21 transports the waste to the low-speed rotary crusher 3. The first conveyor 21 is the supply conveyor for the low-speed rotary crusher 3. The first conveyor 21 comprises a pair of pulleys and a belt wrapped around the pair of pulleys. When the pulleys rotate, the belt moves. Waste placed on the belt is transported along the belt.

[0041] The low-speed rotary crusher 3 is installed inside the crusher room 13. The receiving hopper 211 is located outside the crusher room 13. The first conveyor 21 transports the waste fed into the receiving hopper 211 upwards into the crusher room 13. The waste transported by the first conveyor 21 is fed into the low-speed rotary crusher 3.

[0042] The low-speed rotary crusher 3 coarsely crushes the waste. The low-speed rotary crusher 3 has a pair of rotating blades that rotate at a low speed. The pair of rotating blades are on a horizontal axis and are arranged side by side. Note that the configuration of the crusher that coarsely crushes the waste is not limited to the illustrated example.

[0043] The waste, roughly crushed by the low-speed rotary crusher 3, is fed into the second conveyor 22. The second conveyor 22 transports the waste to the high-speed rotary crusher 4. The second conveyor 22 is the supply conveyor for the high-speed rotary crusher 4. The second conveyor 22, like the first conveyor 21, is equipped with a pair of pulleys and a belt. The second conveyor 22 transports the waste crushed by the low-speed rotary crusher 3 upwards. The waste transported by the second conveyor 22 is fed into the inlet of the high-speed rotary crusher 4. The inlet of the high-speed rotary crusher 4 opens upwards at the top of the high-speed rotary crusher 4.

[0044] The high-speed rotary crusher 4 further crushes the waste into smaller pieces. The high-speed rotary crusher 4 has a multi-stage grinder that rotates at high speed. The grinder has a vertical axis. Note that the configuration of the crusher that finely crushes the waste is not limited to the example shown.

[0045] The waste, finely crushed by the high-speed rotary crusher 4, is fed from the outlet of the high-speed rotary crusher 4 into the input section 231 of the third conveyor 23. The outlet of the high-speed rotary crusher 4 opens laterally at the bottom of the high-speed rotary crusher 4. The input section 231 is located below the outlet of the high-speed rotary crusher 4. The third conveyor 23 transports the waste to the sorting device 14, which is installed outside the crusher room 13.

[0046] The sorting device 14 sorts the waste into three types: non-combustible residue, combustible residue, aluminum, and iron. The sorting device 14 is equipped with a magnetic separator 51, a crushed material separator 52, an aluminum separator 53, an aluminum air separator 54, and an iron air separator 55.

[0047] The fourth conveyor 24 is continuous with the third conveyor 23. The fourth conveyor 24 transports the crushed material conveyed by the third conveyor 23 to the magnetic separator 51.

[0048] The magnetic separator 51 uses magnetism to separate the crushed waste into iron and other materials. The iron separated by the magnetic separator 51 is sent to the iron air separator 55. The waste other than iron is sent to the crushed material separator 52.

[0049] The crushing and sorting machine 52 separates the waste into non-combustible residue and the rest. The non-combustible residue separated by the crushing and sorting machine 52 is transported to the pit 56 by the residue conveying conveyor 25. The waste other than the non-combustible residue separated by the crushing and sorting machine 52 is sent to the aluminum sorting machine 53.

[0050] The aluminum sorting machine 53 separates aluminum from the waste. The remaining waste after the aluminum has been separated by the aluminum sorting machine 53 becomes combustible residue. The aluminum separated by the aluminum sorting machine 53 is sent to the aluminum air separator 54. The combustible residue is transported to the pit 56 by the residue conveyor 25.

[0051] The aluminum wind separator 54 removes lightweight materials from the aluminum sent from the aluminum separator 53. After passing through the aluminum wind separator 54, the aluminum is transported to the aluminum storage hopper 58. The aluminum storage hopper 58 stores the aluminum.

[0052] The iron wind separator 55 removes lighter materials from the iron sent from the magnetic separator 51. After passing through the iron wind separator 55, the iron is transported to the iron storage hopper 59. The iron storage hopper 59 stores the iron.

[0053] Compared to a processing facility that incinerates combustible waste in an incinerator, the crushing and sorting facility 1 has the potential for ignition at every stage from acceptance to shipment of the materials being processed. In other words, while combustible waste is basically immune to ignition once it is incinerated and reduced to ash, the waste handled by the crushing and sorting facility 1 is susceptible to ignition at every stage. Therefore, early detection of ignition is required.

[0054] Among the waste materials, lithium-ion batteries can generate heat and catch fire due to short circuits caused by external impacts such as crushing or compression. Because lithium-ion batteries generate heat and produce a continuous flame, there is a risk of the flames spreading to surrounding waste. Compared to other waste materials, lithium-ion battery ignition is more likely to lead to a fire in the crushing and sorting facility 1. Therefore, there is a need to quickly detect lithium-ion battery ignition.

[0055] (Configuration of the fire detection device) A fire detection device 6 is installed in the crushing and sorting processing facility 1. Figure 2 shows the fire detection device 6. The fire detection device 6 selectively detects fires in lithium-ion batteries.

[0056] The fire detection device 6 detects the ignition of lithium-ion batteries in the high-speed rotary crusher 4. This is because lithium-ion batteries are highly likely to ignite when crushed by the high-speed rotary crusher 4. Furthermore, if the fire is not detected and extinguished in the high-speed rotary crusher 4, the burning lithium-ion battery may be sent to the next sorting device 14, potentially spreading the fire. To prevent the burning lithium-ion battery from being transported to the sorting device 14, it is effective to promptly detect the ignition of lithium-ion batteries in the high-speed rotary crusher 4.

[0057] The fire detection device 6 includes a first camera 71 as a photographic device. The first camera 71 is, for example, a digital camera. The first camera 71 can capture color moving images or color still images.

[0058] The first camera 71 is installed in the high-speed rotary crusher 4. More specifically, the first camera 71 is installed in the input hood of the high-speed rotary crusher 4, at a position above the main body of the high-speed rotary crusher 4, so as to photograph the entrance to the main body of the high-speed rotary crusher 4 from above. The first camera 71 can photograph the process of crushing waste inside the high-speed rotary crusher 4. The first camera 71 can photograph any fire that occurs inside the high-speed rotary crusher 4.

[0059] The fire detection device 6 is equipped with a second camera 72. The second camera 72 is installed near the outlet of the high-speed rotary crusher 4. The second camera 72 can photograph the waste supplied from the outlet of the high-speed rotary crusher 4 to the third conveyor 23.

[0060] The fire detection device 6 includes a control device 63. The control device 63 includes an input unit 631, a first detection unit 632, a second detection unit 633, a determination unit 634, an output unit 635, and a memory 636. The control device 63 can be configured, for example, by a general-purpose computer. The control device 63 can be configured to include a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The control device 63 constitutes a fire detection system that executes various processes described later by reading a computer program. At least the input unit 631, the first detection unit 632, the second detection unit 633, the determination unit 634, and the output unit 635 are functional blocks realized by the control device 63 after reading the program.

[0061] The first camera 71 is connected to the control device 63. Various known configurations can be used for the connection between the first camera 71 and the control device 63 as appropriate. The second camera 72 may also be connected to the control device 63.

[0062] The color image data captured by the first camera 71 is input to the input unit 631 of the control device 63. The image from the first camera 71 input to the input unit 631 is displayed on the monitor 73. The monitor 73 displays the image from the first camera 71, for example, in real time. Note that the monitor 73 may be omitted. The image from the first camera 71 input to the input unit 631 is stored in the memory 636. The memory 636 is composed of a large-capacity memory or non-volatile memory such as a hard disk or SSD.

[0063] The first detection unit 632 detects ignition in the high-speed rotary crusher 4 based on image data from the first camera 71. Specifically, the first detection unit 632 detects ignition of a specific type of waste based on the color information of the image. In this case, the specific waste is a lithium-ion battery. When a lithium-ion battery ignites, the flame color turns crimson due to the flame reaction of lithium ions. The first detection unit 632 detects the ignition of the lithium-ion battery by detecting a crimson flame from the image data of the first camera 71.

[0064] Specifically, the first detection unit 632 encodes the color of each pixel in the image data from the first camera 71 using a predetermined color space. Various known color spaces can be appropriately adopted. Examples of known color spaces include RGB, HSV, HSL, or L*a*b*.

[0065] The first detection unit 632 also extracts pixels that can be determined to be crimson based on the code of each pixel. The range of code values ​​that can be determined to be crimson is set in advance. The first detection unit 632 calculates the total number of extracted pixels. The first detection unit 632 calculates the area of ​​the crimson pixels, in other words, the area occupied by the crimson flames, in the image from the first camera 71. In the following, crimson pixels may be referred to as "red cells".

[0066] Here, Figure 3 illustrates the image from the first camera 71 (upper part of Figure 3) and the state in which red cells corresponding to the flame reaction of lithium ions have been identified from the image (lower part of Figure 3). In the upper part of Figure 3, reference numeral 31 indicates the edge 31 of the monitoring window of the high-speed rotary crusher 4, and the area inside the edge 31 of the monitoring window corresponds to the inside of the main body of the high-speed rotary crusher 4. In the upper part of Figure 3, the shaded area labeled reference numeral 32 indicates the ignition point inside the high-speed rotary crusher 4. In the lower part of Figure 3, reference numeral 33 indicates the red cell 33 that the first detection unit 632 has determined to be crimson. Note that in the lower part of Figure 3, one red cell 33 is drawn larger for ease of understanding. The first detection unit 632 calculates the total number of red cells 33 extracted in the lower part of Figure 3. The total number of pixels corresponds to the area of ​​the red cells 33 in the image from the first camera 71. The first detection unit 632 outputs the calculated sum of the red cells 33 to the determination unit 634.

[0067] The second detection unit 633 detects ignition of all waste generated inside the high-speed rotary crusher 4, not just lithium-ion batteries. The second detection unit 633 uses AI technology to detect ignition from image data. Specifically, the second detection unit 633 has a trained model 637. The trained model 637 may be pre-generated, for example, by machine learning. The trained model 637 can be generated using various known methods. For example, the trained model 637 may be generated, for example, by deep learning, using a training dataset consisting of training images that include images taken inside the high-speed rotary crusher 4 where ignition has not occurred and images where ignition has occurred, and the judgment result of whether or not ignition has occurred for each of the training images. The second detection unit 633 can also perform reinforcement learning using image data from the first camera 71 while the crushing and sorting processing equipment 1 is in operation.

[0068] The second detection unit 633 also outputs multiple classes corresponding to the degree of ignition. Specifically, as an example, the second detection unit 633 outputs a total of seven classes to the determination unit 634, ranging from class 0 (no ignition) to class 1 through class 6, in order of increasing degree of ignition. The second detection unit 633 can detect not only ignition but also the degree of ignition.

[0069] The determination unit 634 receives the output of the first detection unit 632 and the output of the second detection unit 633, and determines whether or not the lithium-ion battery has ignited inside the high-speed rotary crusher 4. The specific procedure for determination by the determination unit 634 will be described later.

[0070] The crushing and sorting equipment 1 is equipped with a fire extinguishing system 8. The fire extinguishing system 8 includes nozzles 81, 82, and 83, an on-off valve 84, and piping 85. Nozzle 81 is installed above the inlet of the high-speed rotary crusher 4 inside the input hood. Nozzle 81 sprays water into the high-speed rotary crusher 4 through its inlet. Nozzle 82 is installed near the outlet of the high-speed rotary crusher 4. Nozzle 82 sprays water onto the waste discharged from the outlet of the high-speed rotary crusher 4. Nozzle 83 is installed above the input section 231 of the third conveyor 23. Nozzle 83 sprays water onto the waste fed from the outlet of the high-speed rotary crusher 4 into the input section 231 of the third conveyor 23. The fire extinguishing system 8 can quickly extinguish a fire if the waste ignites in the high-speed rotary crusher 4. This prevents ignition materials from being transported to the sorting device 14 by the third conveyor 23.

[0071] Nozzles 81, 82, and 83 are all connected to a shut-off valve 84 via piping 85. When the shut-off valve 84 opens, water is supplied to nozzles 81, 82, and 83 via piping 85. Each of the nozzles 81, 82, and 83 then sprays water.

[0072] The on-off valve 84 is, for example, an electromagnetic on-off valve, and the output unit 635 of the control device 63 outputs an open signal for the on-off valve 84. When the output unit 635 outputs an open signal, the on-off valve 84 opens and water is sprayed through the nozzles 81, 82, and 83. Alternatively, instead of the control device 63 directly outputting an open signal to the on-off valve 84, a control panel that controls the on-off valve 84 may be interposed between the control device 63 and the on-off valve 84. The output unit of the control device 63 may output an open signal to the control panel, and the control panel, upon receiving the open signal, may control the on-off valve 84 to open it.

[0073] (Control procedure for fire detection device) Figure 4 shows an example of the control procedure for the fire detection device 6. In step S41 after the start, the input unit 631 reads the image from the first camera 71 (input processing). In the following step S42, the second detection unit 633 inputs the image to the trained model 637, and in the following step S43, the second detection unit 633 outputs one of classes 0 to 6 to the determination unit 634.

[0074] In step S44, the determination unit 634 determines whether the number of outputs for classes 1 to 6 has exceeded a set number. Classes 1 to 6 correspond to instances of ignition occurring in the high-speed rotary crusher 4, and the number of outputs for classes 1 to 6 is the number of outputs counted individually for each of the classes. In step S44, it is determined whether the number of outputs for any of the classes 1 to 6 has exceeded a predetermined number. The predetermined number (i.e., the threshold) may be the same for each of the classes 1 to 6, or it may be different for each of the classes. If the determination in step S44 is Yes, that is, if ignition is detected by the second detection unit 633, the process in Figure 4 proceeds to step S45; otherwise, the process in Figure 4 proceeds to step S48.

[0075] In step S44, the determination unit 634 may determine whether the number of outputs for Class 1 to 6 has exceeded a set number, regardless of which of Class 1 to 6 it is. In other words, in step S44, the determination unit 634 may determine "Yes" if the number of times ignition has been detected exceeds a predetermined number, regardless of the degree of ignition.

[0076] In step S45, the first detection unit 632 calculates the sum of the red cells 33 based on the image from the first camera 71 and outputs it to the determination unit 634. In the following step S46, the determination unit 634 determines whether the sum of the red cells 33 exceeds a threshold. If it exceeds the threshold, the determination unit 634 determines that the lithium-ion battery has caught fire (detection process). In this way, the computer program causes the computer to execute input processing and detection processing. On the other hand, if the number of outputs for classes 1 to 6 does not exceed a set number, or if the number of outputs for classes 1 to 6 exceeds a set number but the sum of the red cells 33 does not exceed a threshold, the determination unit 634 does not determine that the lithium-ion battery has caught fire. If the determination in step S46 is Yes, the process in Figure 4 proceeds to step S47; otherwise, it proceeds to step S48.

[0077] In step S47, the output unit 635 outputs a signal to open the on-off valve 84, that is, a signal to start water spraying to the fire extinguishing system 8. The on-off valve 84 opens, and water is sprayed through the nozzles 81, 82, and 83. In step S48, the output unit 635 does not output an open valve signal.

[0078] According to the aforementioned fire detection device 6, fire detection system, and computer program, the fire detection device 6, etc., can selectively detect the ignition of lithium-ion batteries among the waste ignitions based on the color information of the image from the first camera 71. The ignition of lithium-ion batteries is highly likely to lead to a fire in the crushing and sorting processing facility 1. The fire detection device 6, etc., can quickly detect the ignition of lithium-ion batteries, which are prone to causing fires.

[0079] The fire detection device 6, etc., outputs a valve open signal to initiate firefighting operations when it detects a fire in a lithium-ion battery. However, even if a fire is detected, if it is caused by waste other than a lithium-ion battery, it does not output a valve open signal. The fire detection device 6, etc., can reduce the frequency of fire detections that do not require firefighting operations. This makes it possible to maintain the operational efficiency of the facility. Furthermore, by detecting a fire in a lithium-ion battery that requires firefighting operations, firefighting operations can be started quickly, thus effectively suppressing the fire.

[0080] Furthermore, the ignition detection device 6, etc., detects ignition of a lithium-ion battery based on the presence of a predetermined number or more of deep red cells, i.e., red cells 33, in the image, which correspond to the flame reaction of lithium ions contained in the lithium-ion battery. The ignition detection device 6, etc., can detect ignition of a lithium-ion battery with high accuracy.

[0081] Furthermore, the ignition detection device 6, etc., encodes the image color using a predetermined color space and then determines the color corresponding to the flame test of lithium ions. The ignition detection device 6, etc., can accurately determine the deep red color corresponding to the flame test of lithium ions. Among the waste processed in the crushing and sorting facility 1, there is no waste that exhibits a flame test color similar to that of lithium ions. Based on the unique color corresponding to the flame test of lithium ions, the ignition detection device 6, etc., can accurately detect the ignition of lithium-ion batteries.

[0082] Furthermore, since the fire detection device 6 uses images from the first camera 71 installed in the high-speed rotary crusher 4, the fire detection device 6 can quickly detect ignition of lithium-ion batteries in the high-speed rotary crusher 4. In addition to quickly detecting ignition of lithium-ion batteries in the high-speed rotary crusher 4, the fire extinguishing equipment 8 sprays water on the high-speed rotary crusher 4, so even if a lithium-ion battery ignites, the fire can be quickly extinguished. The spread of the fire is effectively suppressed. Because the spread of the fire is suppressed, the fire detection device 6 can minimize the shutdown of the crushing and sorting processing equipment 1.

[0083] The ignition detection device 6, etc., detects the ignition of a lithium-ion battery based on both the detection result of the second detection unit 633 and the detection result of the first detection unit 632, thereby enabling more accurate detection of lithium-ion battery ignition.

[0084] In addition, in the flowchart of Figure 4, the order of block 401, which includes the detection step by the second detection unit 633, and block 402, which includes the detection step by the first detection unit 632, can be swapped.

[0085] (modified version) Figure 5 shows a control procedure for a modified example of the ignition detection device 6. The ignition detection device 6 may omit detection by the second detection unit 633.

[0086] Specifically, in step S51 after the start, the input unit 631 reads the image from the first camera 71. In the following step S52, the first detection unit 632 calculates the sum of the red cells 33 based on the image from the first camera 71 and outputs it to the determination unit 634. In the following step S53, the determination unit 634 determines whether the sum of the red cells 33 exceeds a threshold. If it exceeds the threshold, the determination unit 634 determines that the lithium-ion battery has caught fire. In the following step S54, the output unit 635 outputs an open signal for the on-off valve 84. The on-off valve 84 opens, and water is sprayed through the nozzles 81, 82, and 83. On the other hand, if it does not exceed the threshold, the determination unit 634 does not determine that the lithium-ion battery has caught fire. In step S55, the output unit 635 does not output an open signal.

[0087] Even in modified forms, the fire detection device 6 can selectively detect the ignition of lithium-ion batteries among the ignitions of waste based on the color information of the image from the first camera 71, and can quickly detect the ignition of lithium-ion batteries, which are prone to causing fires.

[0088] Figure 6 shows a control procedure for another modified example of the fire detection device 6. The fire detection device 6 may also incorporate the detection result from the second detection unit 633 into the detection decision made by the first detection unit 632.

[0089] In step S61 after the start, the input unit 631 reads the image from the first camera 71. In the following step S62, the second detection unit 633 inputs the image to the trained model 637, and in the following step S63, the second detection unit 633 outputs one of classes 0 to 6 to the determination unit 634.

[0090] In step S64, the determination unit 634 determines whether the number of outputs for classes 1 to 6 has exceeded the set number. Step S64 is the same as step S44 described above. If the determination in step S64 is Yes, the process in Figure 6 proceeds to step S65; otherwise, it proceeds to step S68.

[0091] In step S65, the first detection unit 632 calculates the sum of the red cells 33 based on the image from the first camera 71 and outputs it to the determination unit 634. In the following step S66, the determination unit 634 determines whether the sum of the red cells 33 exceeds a threshold. Here, the threshold is set according to the class output by the second detection unit 633. The lower the output class (class 1, 2), the smaller the threshold may be, and the higher the output class (class 5, 6), the larger the threshold may be.

[0092] If the sum of the red cells 33 exceeds the threshold, the determination unit 634 determines that the lithium-ion battery has caught fire. In the following step S67, the output unit 635 outputs a signal to open the on / off valve 84. If the threshold is not exceeded, the determination unit 634 does not determine that the lithium-ion battery has caught fire. In step S68, the output unit 635 does not output a signal to open the valve.

[0093] A lower output class indicates a lower degree of ignition. On the other hand, even if the area of ​​the red cell 33 is small, fire extinguishing will occur. Regardless of the degree of ignition, the system can accurately detect lithium-ion battery fires and initiate fire extinguishing.

[0094] Since the detection result from the second detection unit 633 is reflected in the detection decision of the first detection unit 632, the fire detection device 6, etc., can more accurately detect the ignition of a lithium-ion battery that could lead to a fire, and can start firefighting.

[0095] In the flow chart of Figure 6, the detection by the second detection unit 633 (steps S62 and S63) and the detection by the first detection unit 632 (step S65) can be reversed in order.

[0096] Figure 7 shows a control procedure for a further modification of the ignition detection device 6. The control procedure in Figure 7 corresponds to a combination of the control procedure in Figure 5 and the control procedure in Figure 6.

[0097] In step S71 after the start, the input unit 631 reads the image from the first camera 71. In the following step S72, the second detection unit 633 inputs the image to the trained model 637, and in the following step S73, the second detection unit 633 outputs one of classes 0 to 6 to the determination unit 634 according to the trained model 637.

[0098] In step S74, the first detection unit 632 calculates the sum of the red cells 33 based on the image from the first camera 71 and outputs it to the determination unit 634. In the following step S75, the determination unit 634 determines whether the sum of the red cells 33 exceeds a threshold set according to the class output by the second detection unit 633. If the sum of the red cells 33 exceeds the threshold, the determination unit 634 determines that the lithium-ion battery has caught fire. In the following step S76, the output unit 635 outputs an open signal for the on / off valve 84. If the threshold is not exceeded, the determination unit 634 does not determine that the lithium-ion battery has caught fire. In step S77, the output unit 635 does not output an open signal.

[0099] Since the detection result from the second detection unit 633 is reflected in the detection decision of the first detection unit 632, the ignition detection device 6, etc., can more accurately detect the ignition of a lithium-ion battery that could lead to a fire.

[0100] In the flow chart of Figure 7, the detection by the second detection unit 633 (steps S72 and S73) and the detection by the first detection unit 632 (step S74) can be performed in any order.

[0101] (Other variations) The fire detection device 6, etc., may change and control the amount of water sprayed during firefighting operations according to the degree of ignition (or fire) detected. For example, the fire detection device 6, etc., may increase the amount of water sprayed if the total number of red cells is large, and decrease the amount of water sprayed if the total number of red cells is small. Since the total number of red cells corresponds to the degree of ignition, increasing the amount of water sprayed when the total number of red cells is large can achieve effective firefighting. By decreasing the amount of water sprayed when the total number of red cells is small, effective firefighting can be achieved with minimal water spraying. A smaller amount of water sprayed shortens the time required for post-treatment after water spraying, and shortens the time required from the suspension of treatment due to ignition to the resumption of treatment. As a result, the operational efficiency of the facility can be maintained. Note that the amount of water sprayed may also be adjusted, for example, by adjusting the opening time of the on-off valve 84.

[0102] The fire detection device 6, etc., may change and control the water spraying locations (or water spraying timing) according to the degree of ignition (or fire) detected. For example, the fire detection device 6, etc. may increase the number of water spraying locations if the total number of red cells is large, and decrease the number of water spraying locations if the total number of red cells is small. If the total number of red cells is large, the fire detection device 6, etc. can achieve effective fire extinguishing through wide-area water spraying by spraying water on both the high-speed rotary crusher 4 and the third conveyor 23, for example. If the total number of red cells is small, the fire detection device 6, etc. will spray water only on the high-speed rotary crusher 4. By narrowing the water spraying range, effective fire extinguishing can be achieved with minimal water spraying. A narrow water spraying range also reduces the time required for post-treatment after water spraying, and shortens the time required from the suspension of processing due to ignition to the resumption of processing. As a result, the operational efficiency of the facility can be maintained.

[0103] Figure 8 shows a modified version of the fire detection device 60. The fire detection device 60 uses AI to detect the ignition of specific waste, namely lithium-ion batteries, based on the color information of the image.

[0104] The control device 630 of the fire detection device 60 has an input unit 631, a determination unit 634, an output unit 635, a memory 636, and a detection unit 638. The detection unit 638 has a trained model 639. The trained model 639 is a model that determines whether or not a lithium-ion battery has caught fire based on the color information of an image, using crimson as a feature. The trained model 639 can be generated using various known methods. The detection unit 638 inputs an image from the first camera 71 to the trained model 639 and detects the ignition of the lithium-ion battery.

[0105] The determination unit 634 receives the output from the detection unit 638 and determines whether or not the lithium-ion battery has caught fire inside the high-speed rotary crusher 4.

[0106] The fire detection device 60 can also detect ignition of lithium-ion batteries in a distinct manner from ignition of other waste materials. The fire detection device 60 can effectively suppress fires and reduce the frequency of detecting ignitions that do not require firefighting, thereby maintaining the operational efficiency of the facility.

[0107] The ignition detection device 6 can also detect the ignition of a lithium-ion battery using images from the second camera 72 instead of using images from the first camera 71 installed on the high-speed rotary crusher 4. Furthermore, the ignition detection device 6 can also detect the ignition of a lithium-ion battery using images from both the first camera 71 and the second camera 72.

[0108] The fire detection device 6 is not limited to detecting the ignition of lithium-ion batteries. The fire detection device 6 can detect the ignition of specific waste based on the color information of an image. The fire detection device 6 may also detect the ignition of specific waste based on the color corresponding to the flame reaction of that specific waste. As described above, the fire detection device 6 may encode the colors of an image using a predetermined color space and then determine the color corresponding to the flame reaction of the specific waste.

[0109] Furthermore, although the ignition detection device 6 is intended for the high-speed rotary crusher 4, the ignition detection device 6 may also be intended for the low-speed rotary crusher 3 and may detect ignition of specific waste. In addition, the ignition detection device 6 may be intended for waste being transported by any of the conveyors 21, 22, 23, or 24 in the crushing and sorting processing facility 1 and may detect ignition.

[0110] The configuration of the crushing and sorting equipment 1 shown in Figure 1 is just one example, and various configurations can be appropriately adopted for the crushing and sorting equipment 1.

[0111] Furthermore, the fire detection device, fire detection system, and computer program disclosed herein are not limited to application to the crushing and sorting equipment 1, but can be broadly applied to waste treatment facilities where waste is processed. [Explanation of Symbols]

[0112] 1. Crushing and sorting equipment (waste treatment facility) 33 Red Cells 4. High-speed rotary crusher (crusher) 6. Fire detection device 60. Fire detection device 63 Control device (fire detection system) 632 First detection unit 633 Second detection unit 634 Judgment section 637 Pre-trained models 638 Detection unit 71. First camera (imaging device) 72. Second camera (imaging device)

Claims

1. In the waste processing process at a waste treatment facility, a photographic device for photographing the waste, A fire detection device comprising: a detection unit that detects the ignition of a specific waste based on the color information of an image captured by the aforementioned imaging device.

2. In the waste processing process at a waste treatment facility, a photographic device for photographing the waste, A fire detection device comprising: a detection unit that distinguishes and detects the ignition of specific waste based on the color information of an image captured by the aforementioned imaging device.

3. In the fire detection device according to claim 1 or 2, The detection unit detects that the specific waste has ignited when the image contains a predetermined number or more cells of the color corresponding to the flame reaction of the specific waste.

4. In the fire detection device according to claim 3, The detection unit is a fire detection device that encodes the colors of the image using a predetermined color space and then determines the color corresponding to the flame reaction of the specific waste.

5. In the fire detection device according to claim 3, The aforementioned specific waste is a lithium-ion battery, The aforementioned detection unit is a fire detection device that determines the color corresponding to the flame reaction of lithium ions.

6. In the fire detection device according to claim 1 or 2, The aforementioned imaging device is a fire detection device that photographs the waste being crushed in the crusher.

7. In the fire detection device according to claim 6, A fire detection device further comprising an output unit that outputs a water spraying execution signal to the fire extinguishing equipment of the crusher when the detection unit detects ignition of the specific waste in the crusher.

8. In the fire detection device according to claim 1 or 2, A second detection unit having a trained model and detecting the ignition of the waste based on the image captured by the imaging device and the trained model, A fire detection device further comprising: a determination unit that determines that the fire is caused by the specific waste when the detection unit and the second detection unit each detect fire.

9. A system for detecting ignition in the waste processing process at a waste treatment facility, An input unit that receives images from a camera that photographs the waste in the processing step, A fire detection system comprising: a detection unit that detects the ignition of a specific type of waste based on the color information of the aforementioned image.

10. A computer program that causes a computer to perform a process to detect ignition in the waste processing process at a waste treatment facility, Input processing that receives images from a camera that photographs the waste in the processing step, A computer program that causes a computer to perform a detection process to detect the ignition of a specific type of waste based on the color information of the aforementioned image.

Images