JP2005106748A - Heat generation monitoring method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide heat generation monitoring method and device for surely detecting a high-temperature strip mixed in normal-temperature strips in a conveyer which transports strips, especially, and to provide heat generation monitoring method and device for surely detecting a heating RDF in an RDF manufacturing plant for alarming. <P>SOLUTION: A position where a strip 13 falls, at an exit of a strip carrying conveyer 11, is photographed with an infrared TV camera to obtain a temperature distribution image. The temperature distribution image obtained is image-processed to extract a high-temperature portion, and an alarm is issued when such a high-temperature portion is present. It is preferred to confirm, in subsequent images, that the high-temperature portion is detected in a position predicted based on the free fall equation, determine that it is surely a fall of a high-temperature RDF 14 or the like, and then issue an alarm. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method and apparatus for detecting heated debris and crushed material buried or hidden between those at normal temperature, and more particularly, a method and apparatus for monitoring the heat generation of RDF particles in a RDF (refuse derived fuel) manufacturing plant. About.

Since the amount of plastic waste is increasing year by year, with the aim of recycling waste, solid waste from municipal waste and construction waste plastic is solidified and recovered as solid RDF (refuse derived fuel). However, technologies for use as alternative energy for oil in power plants and the like have been actively developed.
RDF is in the form of pellets that are dried, crushed, added with chemicals, and compressed and molded from combustible waste as a raw material. It has no foul odor, is resistant to spoilage, and can be stored for a long period of time. It is a high-calorie solid fuel that is easy to handle.

By the way, when manufacturing RDF, RDF particles that are self-fired in the compression drying process and become high temperature may be mixed in the product RDF, and if this is mixed with RDF accumulated in a silo or storage tank, a fire may occur. There is.
Therefore, it is desired to detect and warn of high-temperature RDF particles during transportation of the conveyor before storing the finished RDF product in a silo or the like.
It is preferable to take measures to detect and remove the waste pieces and crushed abnormal heat generation materials as raw materials for RDF in advance during the RDF manufacturing process because they also pose a danger.

  As a method for detecting a high temperature body in a conveyor, for example, there is a method disclosed in Patent Document 1. In the method disclosed in Patent Document 1, waste plastics or the like used as an auxiliary material in a steel smelting furnace are dropped or scattered from the conveyor while being transported on the belt conveyor and accumulated to generate heat. A large visual field is monitored at once using a contact temperature sensor.

When trying to apply the method disclosed in Patent Document 1 to the conveyor transport process of RDF raw material or product RDF,
1. When transporting a large amount of RDF, garbage, etc., hot strips may be buried in normal strips on the conveyor and may not be detected by an infrared camera installed above.
2. Infrared cameras must be prevented from falsely detecting infrared rays from sunlight, illumination light, etc.
There are difficulties such as.

JP 2002-205813 A

  The problem to be solved by the present invention is to provide a heat generation monitoring method and apparatus for reliably detecting high-temperature strips mixed in strips of normal temperature in a conveyor for transporting strips, and in particular, an RDF manufacturing plant. The present invention provides a heat generation monitoring method and apparatus for reliably detecting and alarming a heat generation RDF.

In order to solve the above problem, the heat generation monitoring method of the present invention obtains a temperature distribution image by photographing the position where the strip at the exit of the strip conveyor conveys with an infrared TV camera, and obtains the obtained temperature distribution image. A high temperature part is extracted by image processing, and an alarm is issued when the high temperature part exists.
Moreover, it is preferable to acquire a temperature distribution image at a predetermined time interval, and to generate an alarm when a high temperature portion is detected at a position predicted by a free fall equation in subsequent images.
Further, it is possible to confirm the coincidence with the free fall type using three or more images.
The heat generation monitoring method of the present invention can be applied to RDF strips.

Further, the heat generation monitoring device of the present invention includes an infrared camera installed at a position where the position where the strip falls at the exit of the strip conveyor can be seen, and an image processing apparatus that inputs the output of the infrared camera and extracts a high-temperature portion. An arithmetic unit that calculates the position that should appear in the next image as the hot part falls freely using the output of the image processing device, and compares it with the position of the hot part in the actual next image; And a display device for displaying the determination result.
Note that a plurality of element image processing devices may be provided in the image processing device, and the temperature distribution image input by each element processing device may be subjected to image processing one frame at a time to extract the high temperature portion.

When transporting RDF while it is thickly deposited on a conveyor, high-temperature RDF particles that are generated during the manufacturing process without generating heat or being cooled cannot be observed from the surface because they are buried under the deposited RDF. . The particles buried in the interior can be exposed and observed by the camera when falling from the end of the conveyor.
Therefore, an infrared camera is installed at the fall position, the fall state is photographed at predetermined time intervals, and image processing is performed to create a temperature distribution diagram. The obtained temperature distribution diagram is color-coded corresponding to the temperature level, for example, and when displayed on the display device of the infrared camera, the position and size of the high temperature portion can be known at a glance.

In the present invention, this temperature distribution map is further taken into the image processing apparatus, and the RDF in an overheated state is extracted by cutting out a portion having a predetermined temperature or higher, for example. When there is a high-temperature part in the temperature distribution diagram, there is a high possibility that an overheated RDF exists, so an alarm is issued to the operator and necessary measures are taken to ensure safety. Can do.
Further, it is possible to detect the overheat RDF more reliably by using a plurality of temperature distribution diagrams acquired at predetermined time intervals by the infrared camera.

Extract the overheated part in the first image, and assume that this overheated part will fall freely. Estimate the position where the overheated part appears in the subsequent image and check it against the actual temperature distribution map. By doing so, erroneous detection caused by stray light or the like is prevented, and the detection accuracy of the overheated RDF is greatly improved.
The free fall equation can be easily solved and calculated in a relatively short time.
In addition, in the heat generation monitoring device that includes a plurality of elemental image processing devices that perform image processing of each temperature distribution image one by one, the image processing device that operates at a relatively low speed can sufficiently function, As a whole, the device can be manufactured at low cost.

Hereinafter, the heat generation monitoring method and apparatus of the present invention will be described in detail with reference to examples.
FIG. 1 is an explanatory diagram showing a state in which the heat generation monitoring device of the embodiment according to the present invention is installed in the conveyor section, FIG. 2 is a configuration diagram of the heat generation monitoring device of this embodiment, and FIG. 3 is an output image example of an infrared camera FIG. 4 is a diagram for explaining the relationship between the position and distance between the real space and the image space, and FIG. 5 is a diagram for explaining a method for detecting a high temperature portion.

The heat generation monitoring device of this embodiment is applied to waste pieces / crushed materials or RDF particles (hereinafter, represented by RDF) conveyed by a conveyor.
As shown in FIG. 2, the heat generation monitoring device includes an infrared camera 1, a video signal distributor 2, an image processing device 3, a display device 4, and an alarm buzzer 5. The contact signal of the image processing device 3 is controlled. Is transmitted to the device 6. The control device 6 can take measures such as an emergency stop of the plant in response to the contact signal. The image processing apparatus 3 includes one or more image processing arithmetic units 7, a determination unit 8, and an output unit 9 inside. The image processing arithmetic unit 7 and the determination unit 8 are a computer or a processing board having a central processing unit (CPU) and an arithmetic function.

In some manufacturing processes, RDF or the like may generate heat and become a high-temperature state for some reason, and high-temperature RDF may be mixed at the RDF storage location, and may spread and cause a serious accident.
For this reason, the product RDF 13 is cooled upstream of the transport conveyor 11 to prevent a heat generation state. However, when the cooling is not sufficient, the RDF particles 14 having a high temperature are mixed with the normal temperature RDF 13 and transported on the conveyor 11. Sometimes.

An infrared camera can be used to find hot particles. The infrared camera 1 outputs an image having higher brightness as a higher temperature portion as a video signal. A normal, relatively low-temperature object is photographed as a dark image, and an object having a high temperature such as a heated RDF is a bright image. Taken as. Therefore, it is possible to extract high temperature particles by binarizing the brightness of the image using an appropriate threshold and cutting out the high temperature portion.
However, when the hot particles are buried in a large amount of normal temperature particles, the hot particles cannot be found even when observed from the upper surface of the belt of the conveyor 11.

Therefore, in the heat generation monitoring device of the present embodiment, the infrared camera 1 is installed at the discharge side end of the transport conveyor 11 at a position where the transported RDF particles and the like are dropped onto the next transport conveyor 12, and is falling. The RDF particles are observed.
In such a fall state, high-temperature particles that could not be observed from the upper surface due to normal temperature particles on the conveyor are scattered and fall apart and can be captured with high probability in the captured image. .

However, in order to reliably capture the high temperature particles in the successively falling object in the camera image, the imaging interval and the image processing time are problems.
Sunlight or illumination light exists in the shooting environment, and the light may be reflected by RDF particles or the like, or may enter the background of the screen, and these rays may enter the image as stray light. Infrared light contained in stray light is imprinted in the output image of the infrared camera and may be mistaken for a heating element. However, such stray light can be distinguished because the motion state in the imaging screen is different from the falling object.

The heat generation monitoring device according to the present embodiment includes the image processing device 3 and reliably detects the hot particles 14 by image processing and determination logic operation.
The RDF particles 13 spilled from the discharge side end of the conveyor 11 fall on the upper surface of the next conveyor 12 while receiving the acceleration of gravity according to the principle of free fall and increasing the fall speed. Therefore, it can be determined separately from noise components such as stray light by tracking the movement of the hot particles and confirming that it follows the fall law.

In order to determine whether or not the fall principle is followed, it is necessary to track the movement of hot particles so that several images can be acquired while one particle falls in the field of view of the infrared camera 1. To. If necessary, multiple infrared cameras may be used to track hot particles, but the calculation logic is complicated, such as handling different camera fields depending on the camera, so one infrared camera is placed at a fixed point. It is preferable to use the video.
When only one infrared camera 1 is used, the distance between the camera and the subject is adjusted to determine the size of the real space that can be accommodated in the camera field of view, and the camera field of view can be viewed vertically by looking at the free fall distance of particles. Is set to be several times the imaging interval determined by the camera, for example, 1/30 second.

Each time the infrared camera 1 captures an image, the infrared camera 1 converts it into a temperature distribution image to generate a video signal and transmits it to the image processing device 3. Based on the video signal, the image processing arithmetic unit 7 of the image processing apparatus 3 performs image processing using a temperature level that can distinguish between the normal temperature RDF and the high temperature RDF, and generates an image in which only the high temperature portion remains. Most simply, a binary image may be used.
Further, the center of gravity position of the high temperature portion is obtained based on the image processed in this way.
The obtained center-of-gravity position coordinates are sent to the determination unit 8, and the position in the next image is estimated using a mathematical model based on the principle of free fall.

Since the image processing operation increases in time in proportion to the number of pixels, when using an infrared camera that outputs a large image, an arithmetic processing unit such as that used industrially has a low speed instead of a high reliability. There are many cases where arithmetic processing is performed, and there may occur a situation where the arithmetic processing for the previous image is not completed before the video signal of the next image is input.
Such a case can be dealt with by preparing a plurality of image processing operation units 7 and performing parallel operations. The number of units is determined by the relationship between the calculation time and the update interval of the video signal. When performing parallel operation, the video signal distributor 2 is used to sequentially distribute the video signals for each frame supplied from the infrared camera 1 to the units.

  Note that even if the image interval is too short, the amount of change in position when the hot particles fall is too small to make a highly accurate determination. Therefore, the imaging screen does not perform image processing for all the continuous frames, but distributes the video signal by appropriately thinning out so as to bring about an appropriate position change according to free fall. Note that the mechanism for thinning out a frame and capturing a video signal can also be used when there is only one arithmetic unit.

FIG. 3 is a diagram schematically showing an image photographed by the infrared camera 1. The image of the infrared camera 1 is displayed as a two-dimensional screen divided into zones for each temperature. Three screens in FIG. 3 represent images acquired at appropriate times tm, tm + 1, and tm + 2 in order from the top.
Further, the image processing apparatus 3 takes in the information of the time when the image is generated from the infrared camera 1 to be used for calculation together with the video signal.
In addition, since the frame image update cycle of the infrared camera 1 and the image processing arithmetic unit are not synchronized, sometimes the frame image may be missed, but the generation time of the captured frame image is known. There is no problem in calculation.

The upper drawing (m) is a frame image taken at a certain time point tm, and an image Am of the heating element is displayed at a position slightly above and to the right of the center. It can be seen that the heating element has a high-temperature part in the center and a slightly high-temperature part around it. In addition, a high temperature portion Bm is observed at a position slightly above the center of the left region.
For the video signal, an appropriate temperature level for distinguishing between the normal temperature RDF and the abnormally high temperature RDF is determined, and only the high temperature portion is cut out, and the barycentric position of the cut out high temperature region is calculated.

The middle drawing (m + 1) in FIG. 3 is a frame image at a time point tm + 1 when a certain time has elapsed from the time point tm. The image of the heating element A is displayed as a high-temperature object image Am + 1 whose shape is slightly changed at a position vertically below the screen. Further, the high temperature portion Bm + 1 that appears on the left side is in a position that does not change compared to the position that appears in the drawing (m).
The same processing is performed on the high-temperature object image in FIG. 3 (m + 1) to calculate the center of gravity position of the high-temperature region.

  If the high temperature object images Am and Bm in FIG. 3 (m) are the same as the high temperature object images Am + 1 and Bm + 1 in FIG. 3 (m + 1), at least the high temperature object images Am + 1 and Bm + 1 in FIG. 3 must exist on the vertical line of the high-temperature object images Am and Bm in FIG. However, due to the direction of motion, the rotation of the object, the collision of particles, etc., there is always a vertical line standing at the center position of the high temperature object Am on the previous screen and a normal line passing through the center of gravity of the high temperature object Am + 1 on the screen (m + 1). It does not overlap, but shows some fluctuation. Therefore, a margin width Δ is set in the horizontal direction for recognizing a fall phenomenon, and if the horizontal coordinate of the center of gravity is within the margin Δ, the images of the same object are recognized.

Further, if the high-temperature object is a free-falling object such as RDF, the vertical coordinate in FIG. 3 (m) and the vertical coordinate in FIG. There must be a difference corresponding to the distance.
For example, the high temperature object image Am + 1 in FIG. 3 (m + 1) represents the state after the free fall of the high temperature object image Am in FIG. 3 (m), and can be estimated as the high temperature RDF, but the high temperature object images Bm and Bm + 1 are Since the height position hardly changes, it is not free-falling, but is likely to be ambient light, for example.
In FIG. 3 (m + 2), a new high-temperature object image C appears at the horizontal position of the high-temperature object image A. However, since there is a contradiction in the vertical position, it can be regarded as noise.

In order to make such a determination, a determination unit 8 is prepared.
The determination unit 8 estimates the movement in the real space using the data in the image space in consideration of the relationship between the position and distance between the real space and the image space, and the high temperature region in the image is a free fall object. If it can be estimated, it is determined that the object is a high-temperature object, and a contact signal is generated to alert or take emergency measures.

FIG. 4 is a diagram illustrating the relationship between the position and distance between the real space and the image space.
The image 21 of the infrared camera is a projection of a certain region 22 in the real space including the locus where the object 23 falls. The real space region 22 is parallel to the camera image 21, and the distance l in the image and the corresponding distance L in the real space always have a constant ratio lc: La.
That is, the ratio γ of the actual distance to the distance on the screen is
γ = La / lc
It becomes.

The high temperature RDF dropped from the upper end of the real space area 22 to the position L0 when the time t passed after starting to fall free from the position of the height a above the area corresponding to the screen in the real space with zero initial velocity. Suppose that a point was taken. The speed of the object O in the real space at this time is set to V0. Also, let lco be the distance on the screen between the top edge of the screen and the object image i.
Then, from the principle of free fall,
a + L0 = 1/2 gt ²
V0 = gt
Because there is a relationship
t = (2 (a + L0) / g) ^1/2
= (2 (a + γlco) / g) ^1/2
And therefore
V0 = (2g (a + L0)) ^1/2
= (2g (a + γlco)) ^1/2
It becomes.

Since the RDF drop start position is the end of the conveyor, the height a is known in advance. Therefore, V0 in the above relational expression can be derived using the observation value lco on the screen.
The fall distance dx when the subsequent time Δtx has passed is determined using a parameter obtained from the observed value on the screen.
dx = 1 / 2gΔtx ² + V0Δtx (1)
Can be calculated.

Therefore, as shown in FIG. 5, a procedure for determining whether or not the object is a high-temperature fall object using three images acquired at appropriate time intervals will be described. If the number of images used is large, the certainty of the determination is improved. However, there is a disadvantage that the calculation capacity increases and the determination time and the calculation device are excessive, and therefore it is necessary to balance at an appropriate place.
For example, assuming that the position from the drop start point to a position 30 cm below is stored on one screen, the RDF particles pass through this screen in 0.24 seconds. If this is photographed with a camera that acquires one image in 1/30 second, RDF is imprinted on seven consecutive images. Accordingly, when every other analysis target image is adopted, the same RDF is found in the three images.

  The three image frames are referred to as m-th image, m + 1-th image, and m + 2-th image in order of collection time. The determination unit 8 stores the barycentric coordinates of the high temperature area detected in each image frame by the image processing arithmetic unit 7 and the acquisition time information for each image frame. Based on the acquisition time of the mth image, a time interval Δt1 until the acquisition time of the m + 1st image and a time interval Δt2 until the acquisition time of the m + 2th image are obtained.

The extent to which the object detected in the m-th image falls in the m + 1-th image and the m + 2-th image is estimated using the above equation (1). However, Expression (1) is obtained by converting the moving distance due to dropping into the number of pixels on the image data.
If the center of gravity of the high-temperature region observed in the m-th image acquired at time tm falls freely, the estimated value d12 of the amount of further falling in the m + 1-th image acquired at time tm + 1 is the frame acquisition Using the time interval Δt1 = tm + 1−tm,
d12 = 1 / 2gΔt1 ² + V0Δt1
It becomes.
A difference Δd12 between the estimated fall distance d12 and the actual fall distance (d2−d1) in the image is calculated.

In this embodiment, in order to further improve the reliability of determination, the determination result for the m + 2th image is also determined.
The estimated value d13 of the amount of further falling in the m + 2 image acquired at time tm + 2 is calculated by using the frame acquisition time interval Δt2 = tm + 2−tm.
d13 = 1 / 2gΔt2 ² + V0Δt2
It becomes. A difference Δd13 between the estimated fall distance d13 and the actual fall distance (d3−d1) in the image is calculated.

If the high-temperature region to be tracked is an object that has fallen freely, the difference Δd12 and Δd13 between the estimated value of the fall distance and the actual value should be within the range of error. When this difference Δd12 or Δd13 is large, there is a very high possibility that it is a noise component other than the free-falling light such as ambient light.
Therefore, a certain threshold θ is determined,
| Δd12 | + | Δd13 |> θ
Is ignored as a false positive,
| Δd12 | + | Δd13 | ≦ θ
In this case, since the object is a fallen object, it is determined that the object is a high-temperature RDF or the like, and an alarm is generated via the alarm unit 5 to prompt the operator to take some measures.

In addition, when an alarm signal is transmitted to the control device 6 via the output unit 9 that outputs a contact signal, an emergency measure such as an automatic stop can be taken.
In these determinations, as described above, the identity of the target is determined by taking the horizontal margin width Δ into consideration for the barycentric coordinates of the high-temperature region.
When there are a plurality of objects in one image, the moving distance is estimated for each object, and the determination is made by comparing with the actual moving distance.

As described above, the number of image frames to be tracked is not limited to three.
In the above description, analysis determination is performed using the frame as a reference, but analysis determination may be performed using the discovered high temperature region as a reference. In this case, when a new high-temperature area is found in the horizontal position, the frame is used as the first frame, and it is checked whether there is a high-temperature area at the same horizontal coordinate position for several subsequent frames. Check to see if they match.

  The heat generation monitoring device of the present embodiment can be reliably judged and excluded when the high temperature region is not a fallen object, so it becomes a rugged device extremely resistant to environmental noise, but the high temperature region caused by external light etc. If a large number of occurrences occur, the calculation load increases, which is not preferable. Accordingly, it goes without saying that it is practically effective to provide a shield in the shooting target area of the infrared camera or to install a light absorption pair in the background so that disturbance stray light does not affect the shooting screen.

It is explanatory drawing which shows the state which installed the heat_generation | fever monitoring apparatus of the Example which concerns on this invention in the conveyance conveyor part. It is a block diagram of the heat_generation | fever monitoring apparatus of a present Example. It is a conceptual diagram which shows the example of an output image of the infrared camera in a present Example. It is drawing explaining the relationship between the position and distance of real space and image space. It is drawing explaining the method to detect the high temperature area | region in a screen.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Infrared camera 2 Video signal distributor 3 Image processing apparatus 4 Display apparatus
5 Alarm buzzer 6 Control device 7 Image processing operation unit 8 Judgment unit 9 Output unit 11, 12 Conveyor 13 RDF particle 14 High temperature RDF particle 21 Infrared camera image 22 Real space region 23 Object

Claims (7)

  A temperature distribution image is obtained by photographing the position where the strip at the exit of the conveyor carrying the strip falls with an infrared TV camera, and the temperature distribution image is image-processed to extract a high-temperature portion. An exothermic monitoring method characterized by issuing an alarm when present.   2. The temperature distribution image is acquired at predetermined time intervals, and an alarm is generated when a high temperature portion is detected at a position predicted by a free fall equation in subsequent images. Fever monitoring method.   3. The heat generation monitoring method according to claim 2, wherein the coincidence with the free fall type is confirmed using three or more temperature distribution images.   4. The heat generation monitoring method according to claim 1, wherein the strip is an RDF (Refuse Derived Fuel) piece.   An infrared camera installed at a position where the position where the strip falls at the outlet of the strip conveyor can be seen, an image processing apparatus for extracting the high temperature portion by inputting the output of the infrared camera, and an output of the image processing apparatus are input. An arithmetic unit that calculates the position that should appear in the next image as the hot part naturally falls and determines the presence of the hot part in comparison with the position of the hot part in the next image actually taken, and determination A heat generation monitoring device comprising: a display device for displaying a result.   The image processing apparatus includes two or more element image processing apparatuses that process the temperature distribution images acquired by the infrared camera one by one, and performs an independent calculation process for each frame of the temperature distribution image to extract a high temperature portion. 6. The heat generation monitoring device according to claim 5, wherein the heat generation monitoring device is performed. 7. The heat generation monitoring apparatus according to claim 5, wherein the strip is an RDF piece.

Images