JP2659080B2 - Noise removal method for surveillance equipment using infrared image by infrared camera - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a monitoring apparatus for monitoring an abnormality of various devices in a facility such as a factory, intrusion of a suspicious person or an animal, and more particularly to a monitoring apparatus using an infrared image by an infrared camera. It relates to a noise removal method.

[0002]

2. Description of the Related Art Conventionally, there has been a method of monitoring an abnormality or the like in a facility such as a factory by using an infrared image obtained by an infrared camera. This monitoring method uses an infrared camera as a component to monitor a monitoring area. It monitors infrared radiation distribution, that is, temperature distribution, and detects abnormalities such as leakage of fluid from the device and intrusion of suspicious persons and animals by detecting abnormal changes. For example, refer to JP-A-62-162939 and JP-A-63-221226 for known examples.

In such a method, an infrared image showing a temperature distribution of an object in a field of view of an infrared camera is subjected to a difference calculation for each pixel at predetermined time intervals, threshold value processing is performed, and a temporal temperature equal to or higher than a predetermined value is obtained. By detecting a change, the above-described abnormality is detected.

[0004] However, a temporal change in temperature in an infrared image is caused not only by the occurrence of an abnormality as an object to be monitored as described above but also by a natural phenomenon. For example, when an object in the field of view moves due to the wind or the camera itself moves, if the outline is clear in the infrared image, a temperature change appears in the pixels at many positions corresponding to the outline, or the object However, in a kind of grass, pixels in which a temperature change appears may be distributed and distributed over a wide range of the infrared image. Also, when the sun rays are reflected on the object in the field of view or the shadows caused by clouds etc. move in the field of view, the pixels where the temperature change appears are relatively dispersed and distributed over a wide area of the image. To be in a state. The occurrence of a temperature change in an infrared image due to such a natural phenomenon becomes noise with respect to a detection signal of an abnormality as a monitoring target.

Conventionally, it has not been practiced to automatically discriminate between a temperature change as noise due to such natural phenomena and a temperature change due to occurrence of an abnormality to remove noise. At the time when an alarm is generated as an abnormality including the temperature change of the noise, the observer checks the image on an image display device such as a CRT and makes a determination.

[0006] In addition, the monitor adjusts the threshold value to a high level when the above-mentioned noise occurs constantly, thereby preventing the abnormal signal from being buried in the noise.
If it is determined from the image that the threshold is too high, the threshold is adjusted lower.

[0007]

Unless the above-mentioned noise is removed, false alarms often occur due to the noise caused by the natural phenomena. Each time a discrimination by the monitor is necessary, the burden on the monitor is large. Also, it is difficult to quickly respond to the occurrence of an abnormality as a monitoring target.

[0008] If the threshold value is not properly adjusted, the above-described noise may cause a false alarm, and may cause a problem of a failure to detect the occurrence of an abnormality as a monitoring target. . The present invention has been made in view of the above circumstances, that is, it is an object of the present invention to automatically remove noise caused by the above-described natural phenomenon and to reliably detect occurrence of an abnormality. It is assumed that.

[0009]

In order to solve the above-mentioned problems, the present invention first performs a difference calculation for each pixel at predetermined time intervals for an infrared image showing a temperature distribution of an object in a visual field of an infrared camera. In a monitoring apparatus using an infrared image that performs threshold processing and detects an abnormality by detecting a temporal temperature change that is equal to or more than a predetermined value, a pixel that detects the temporal temperature change is counted. A noise elimination method is provided in which, when the number is equal to or more than a predetermined number, it is determined as noise, and the processing as an abnormality is not performed.

Further, according to the present invention, a difference calculation is performed for each pixel of an infrared image showing a temperature distribution of an object in a field of view of an infrared camera at predetermined time intervals, threshold processing is performed, and a temporal temperature change exceeding a predetermined value is obtained. In a monitoring device using an infrared image that detects an abnormality by detecting, a pixel that detects the above-mentioned temporal temperature change is counted, and if it is a predetermined number or more, it is determined as noise, and the abnormality is determined. When the state determined as noise continues for a predetermined time or more, the threshold value is adjusted in a higher direction, and the detection of the abnormality is continued. Provide a way.

In addition, the present invention performs a difference calculation for each pixel at predetermined time intervals for an infrared image showing a temperature distribution of an object in a visual field of an infrared camera, performs threshold processing, and detects a temporal temperature change of a predetermined or more. In a monitoring device using an infrared image that detects an abnormality by detecting the number of pixels, the number of pixels that have detected the above-mentioned temporal temperature change is counted. The present invention provides a noise elimination method that does not perform the processing as described above.

Further, the present invention performs a difference calculation for each pixel at predetermined time intervals for an infrared image showing a temperature distribution of an object in a visual field of an infrared camera, performs threshold processing, and detects a temporal temperature change of a predetermined or more. In a monitoring device using an infrared image that detects an abnormality by detecting the number of pixels, the number of pixels that have detected the above-mentioned temporal temperature change is counted. When the state determined as noise continues for a predetermined time or more, the threshold value is adjusted in a smaller direction so that the detection of the abnormality is continued. Provide a way.

[0013]

[Function] Temperature changes caused by leakage of low-temperature gas from pipes and devices, and temperature changes caused by intrusion of suspicious individuals and animals occur locally in the infrared image corresponding to the field of view of the infrared camera. Therefore, the number of pixels extracted as having undergone a temporal temperature change of a predetermined value or more after the difference calculation and the threshold value processing is relatively small.

On the other hand, temperature changes in an infrared image caused by natural phenomena such as the movement of an object in the field of view and the infrared camera due to the wind or the reflection of sun rays and the movement of shadows are dispersed over a wide range. Often in a state of distribution,
Therefore, the number of pixels extracted as described above due to temperature changes caused by these natural phenomena is relatively large.

Therefore, it is possible to determine whether the cause of the temperature change is a natural phenomenon, that is, noise or an abnormality as a monitoring target, based on the number of pixels extracted as described above. it can. That is, the number of extracted pixels is counted, and when the number is equal to or larger than a predetermined number, it is determined that the noise is caused by a natural phenomenon, and the abnormal processing is not performed. Detection accuracy is improved and false alarms are reduced.

If the temperature change in the infrared image determined as noise occurs steadily and continues for a predetermined time or more, a relatively large temperature change due to the abnormality that occurred during that time will be buried. In this case, by adjusting the threshold value in the higher direction so that relatively small temperature changes due to natural phenomena are not detected, it is possible to prevent the temperature changes from being buried due to abnormalities. Detection accuracy is improved.

On the other hand, when the leaves of the tree in the field of view of the infrared camera sway due to the wind, a relatively small number of pixels may be extracted, and the number of extracted pixels is determined when an abnormality occurs. It may be less than the number of pixels to be processed.
Therefore, if the number of pixels extracted corresponding to these is equal to or less than a predetermined number, this is also determined as noise due to natural phenomena, and processing as abnormal is not performed.

However, the reason why the number of pixels to be extracted is small may be that the threshold value is too high. Therefore, if the above-mentioned state continues for a predetermined time or more, the threshold value is adjusted in a lower direction. Thus, it is possible to prevent an abnormality from being missed due to an excessively high threshold value.

(Embodiment) Next, an embodiment of the present invention will be described. FIG. 1 is an explanatory perspective view conceptually showing a monitoring method using an infrared image by an infrared camera. Reference numeral 1 denotes a device in a monitoring area, and reference numeral 2 denotes an infrared camera. The rectangular frame indicated by reference numeral 3 is the field of view of the infrared camera 2, and the inside of the field of view 3 is a monitoring target. Reference numeral 4 denotes an image processing device, 5
Denotes an image display device such as a CRT, and forms a monitoring device together with the infrared camera 2 described above. Reference numeral 7 indicates leakage of a low-temperature fluid as an example of an abnormality. FIG. 2 shows an example of a display on the image display device 5. In the image, the temperature distribution of the leaked low-temperature fluid 7 is displayed together with the device and the like.

The image processing device 4 generates an infrared image showing the temperature distribution of the object within the visual field 3 of the infrared camera 2 for a predetermined time t1.
A predetermined process is performed for each pixel 6 (for example, about 10 seconds). Therefore, the contents of the processing in the image processing apparatus 4 will be described based on the flowchart of FIG.

First, in step S1, the image processing device 4 performs a difference process between the current image and the image before the predetermined time t1 for each pixel 6. Next, in step S2, threshold value processing is performed to extract pixels 6 indicating a temperature change of a predetermined value or more. Then, the number of the extracted pixels 6 is counted in step S3.

Next, the counted number of pixels 6 is determined in step S4.
When the pixel 6 is not extracted at all in step S3 or when the number is very small and is not more than the set number n, an abnormality has occurred as a monitoring target. If it is determined that there is no threshold value, the process branches to a first threshold value adjusting process a.

In the first threshold value adjusting process a, first, a time is measured in step S5, and the accumulated time is compared with a set time in the next step S6, and a predetermined time t2 elapses. If not, the above step S1
As described above, the abnormality detection process from step S1 is repeated every time the predetermined time t1 elapses as described above. Then, when such a processing loop is continued, the time measurement in step S5 proceeds. On the other hand, this timing is reset when the process proceeds from step S4 to step S8.

On the other hand, if n or more pixels are extracted in step S3, it is determined that there is a possibility of occurrence of an abnormality, and in the next step S8, a determination is made by comparing with the set number m. I do. That is, in step S8, the number of pixels 6 counted in step S3 is compared with a set number m. If the number is more than m, the temporal temperature change of the pixels 6 is caused by a natural phenomenon. That is, it is determined to be noise, and the process branches to a second threshold value adjusting process b.

On the other hand, if the counted number of pixels 6 is not more than m, it is determined that the above-described abnormality as a monitoring target has occurred, and the processing corresponding to the occurrence of the abnormality, that is, at step S9. The process branches to the process of generating an abnormality such as generation of an alarm.

In the second threshold value adjusting process b, first, time measurement is performed in step S10, and the next step S1 is performed.
In step 1, the accumulated time is compared with the set time, and a predetermined time t3
If has not elapsed, the flow branches to the step S1 side, and as described above, the step S1 is executed every time the predetermined time t1 elapses.
The abnormality detection processing from step 1 is repeated. If such a processing loop continues, the above-described step S10
The timekeeping in progresses. On the other hand, this timing is performed in step S
If it is determined in step 8 that an abnormality has occurred, the operation is reset.

When the accumulated time measured by the timer in step S10 has passed the predetermined time t3, it is determined that the above-mentioned pixels of a predetermined number m or more are constantly extracted, that is, noise is constantly generated. And the next step S
In step 12, the threshold value in the threshold value process in step S2 is changed. That is, after increasing the threshold value by a predetermined temperature range, the process branches to the step S1 and repeats the abnormality detection processing from the step S1 every time the predetermined time t1 elapses. As described above, the steps for the passage of time in this repetitive processing are not shown.

The first and second threshold value adjusting processes a,
By passing through b, the threshold value can be automatically changed in the threshold value processing in step S2. If the threshold value is too low, stationary noise is generated. It is possible to prevent omission of an abnormality in the detection.
However, such a change of the threshold value can be configured to be manually performed by the observer. In this case, the process proceeds from step S4 or S8 to step S1 via a path indicated by a two-dot chain line in the drawing. Any configuration may be used.

Specific numerical examples of the above operation are as follows. Monitoring range (field of view of infrared camera) 50m x 100m Number of pixels: 256 x 512 = 131072 (pixels: 16 x 16 dots) t1: 10 seconds t2 (= t3): 60 seconds n: 50 m: 10000 Due to natural phenomena Number of pixels to be extracted Number of pixels to be extracted due to movement of an object or camera due to wind, or reflection or shadow of sunlight: 15000 Number of pixels to be extracted due to sway of leaves by wind: 40 Number of pixels to be extracted due to abnormality Number of pixels extracted due to leakage of fluid from the joint:
100

[0030]

As described above, according to the present invention, an infrared image showing the temperature distribution of an object in the field of view of an infrared camera is calculated for each pixel at predetermined time intervals, threshold value processing is performed, and a threshold value processing is performed. In a monitoring device that uses an infrared image to detect anomalies by detecting thermal temperature changes, temperature changes as noise due to natural phenomena and temperature changes due to the occurrence of abnormalities are determined by the number of extracted pixels. Since the noise can be automatically determined and the noise can be removed, the detection accuracy of the abnormality in the monitoring device is improved, and the number of false alarms is reduced.

Also, the level of the threshold value in the threshold value processing is determined based on the number of extracted pixels. That is, when the threshold value is low and noise due to natural phenomena is constantly generated. Automatically adjusts the threshold to a higher value, and automatically adjusts the threshold value to a lower value if the threshold value is too high and generates too little noise. As a result, it is possible to prevent the temperature change from being buried due to the abnormality, and in this respect, the detection accuracy of the abnormality is also improved.

As described above, the detection accuracy of the abnormality is improved and the number of false alarms is reduced, so that the burden on the supervisor is reduced and a quick response to the occurrence of the abnormality becomes possible.

[Brief description of the drawings]

FIG. 1 is an explanatory perspective view conceptually showing a monitoring method to which the method of the present invention is applied.

FIG. 2 is an explanatory front view showing an example of a display of the image display device in the method of FIG.

FIG. 3 is a flowchart showing an example of a processing flow in the present invention.

[Explanation of symbols]

 Reference Signs List 1 device in monitored area 2 infrared camera 3 field of view 4 processing device 5 image display device 6 pixel 7 fluid leakage

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification number Agency reference number FI Technical indication // H04N 5/33 G06F 15/68 350

Claims (4)

(57) [Claims] 1. An infrared camera showing a temperature distribution of an object in a field of view of an infrared camera, performing a difference operation for each pixel at predetermined time intervals, performing a threshold value process, and detecting a temporal temperature change exceeding a predetermined value. In a monitoring apparatus using an infrared image that detects an abnormality by counting, the number of pixels for which the temporal temperature change is detected is counted, and when the number is equal to or more than a predetermined number, the pixel is determined as noise, and processing as an abnormality is performed. Noise elimination method in a monitoring device using an infrared image by an infrared camera, wherein the noise is not performed 2. An infrared image showing a temperature distribution of an object in a field of view of an infrared camera is subjected to a difference calculation for each pixel at predetermined time intervals, threshold value processing is performed, and a temporal temperature change exceeding a predetermined value is detected. In a monitoring apparatus using an infrared image that detects an abnormality by counting, the number of pixels for which the temporal temperature change is detected is counted, and when the number is equal to or more than a predetermined number, the pixel is determined as noise, and processing as an abnormality is performed. And when the state determined as noise continues for a predetermined time or more, the threshold value is adjusted in a higher direction to continue the detection of the abnormality. Noise removal method for surveillance equipment using infrared image by infrared camera 3. An infrared image showing a temperature distribution of an object in a field of view of an infrared camera is subjected to a difference calculation for each pixel at predetermined time intervals, threshold value processing is performed, and a temporal temperature change exceeding a predetermined value is detected. In a monitoring apparatus using an infrared image that detects an abnormality by counting, the number of pixels for which the above-mentioned temporal temperature change is detected is counted, and if the number is less than a predetermined number, it is determined as noise, and processing as an abnormality is performed. Noise elimination method in a monitoring device using an infrared image by an infrared camera, wherein the noise is not performed 4. An infrared camera showing a temperature distribution of an object in a field of view of an infrared camera, performing a difference calculation for each pixel at predetermined time intervals, performing threshold processing, and detecting a temporal temperature change of a predetermined time or more. In a monitoring apparatus using an infrared image that detects an abnormality by counting, the number of pixels for which the above-mentioned temporal temperature change is detected is counted, and if the number is less than a predetermined number, it is determined as noise, and processing as an abnormality is performed. And when the state determined as noise continues for a predetermined time or more, the threshold value is adjusted in a smaller direction to continue the detection of the abnormality. Noise removal method for surveillance equipment using infrared image by infrared camera