JP3715705B2 - Monitoring image processing device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a monitoring image processing apparatus installed at a center for remotely monitoring an unmanned electric station, for example.
[0002]
[Prior art]
The configuration of a conventional monitoring image processing apparatus will be described with reference to FIG. FIG. 14 is a diagram showing a schematic configuration of a conventional monitoring image processing apparatus.
[0003]
In FIG. 14, 1 is a video input unit that receives a video signal output from a TV camera or the like installed at an electric station or the like, 2 is an A / D conversion unit that quantizes a video signal input from the video input unit 1, Current image data calculation storage unit 4 that receives the output from the A / D conversion unit 2 and prepares the current image data. 4 receives the output from the A / D conversion unit 2 and performs corresponding calculations to create comparison image data. The comparison image data calculation storage unit 5 performs a comparison operation on the current image data from the current image data calculation storage unit 3 and the comparison image data from the comparison image data calculation storage unit 4 to extract a change area. It is a change area extraction part. Generally, a binarization processing unit is built in here, and the output is converted into binary data. Reference numeral 6 denotes a recognition processing unit that receives an output from the change area extraction unit 5 and determines whether the object is an object, and 7 is a notification processing unit that receives an instruction from the recognition processing unit 6 and issues a notification. .
[0004]
Next, the operation of the conventional monitoring image processing apparatus will be described with reference to FIG. FIG. 15 is a diagram for explaining the operation of the recognition processing unit of the conventional monitoring image processing apparatus.
[0005]
First, a schematic overall operation is as follows. A video signal (video signal) given from the outside to the apparatus is received by the video input unit 1 and sent to the A / D conversion unit 2. The A / D converter 2 converts the received video signal into digital data for later processing. In general, conversion into multi-value data of 8 bits per pixel is performed.
[0006]
Next, the current image data calculation storage unit 3 prepares “current image data” based on the input data, and prepares for subsequent processing. In many cases, the input data itself is output. On the other hand, the comparison image data calculation storage unit 4 prepares “comparison image data” based on the input data and prepares for subsequent processing. This is often the current image data periodically sampled or the video of the previous frame.
[0007]
Next, the change area extraction unit 5 compares the current image data output from the current image data calculation storage unit 3 with the comparison image data output from the comparison image data calculation storage unit 4, and finds a different portion. Find it and output it as a “change area”. In general, the change area is mainly a moving object in an image, but in rare cases, the change area may be a luminance change such as light reflection.
[0008]
Next, the recognition processing unit 6 receives the change area data from the change area extraction unit 5 and determines whether or not the change area is the target (“report target”). Details of this determination will be described later. The notification processing unit 7 receives the instruction from the recognition processing unit 6 and performs a notification process. This notification process is usually a process of sending a specific signal to a higher-level PC (personal computer) or the like or sounding a device buzzer. In general, the recognition processing unit 6 and the notification processing unit 7 are processed by S / W.
[0009]
Next, a representative example (general theory) of the recognition processing performed by the recognition processing unit 6 will be described. The recognition processing unit 6 performs a determination process for each change region delivered from the change region extraction unit 5 according to a flow as shown in FIG.
[0010]
Conditional branches indicated by reference numerals 10 to 13 are portions for collating conditions for determining whether or not the change area is a notification target. The following four conditions are shown in this example.
[0011]
At reference numeral 10, it is determined whether or not the change area is within the monitoring range.
In reference numeral 11, it is determined whether or not the change area has a predetermined area.
In reference numeral 12, it is determined whether or not the change area has a predetermined vertical and horizontal dimension.
In reference numeral 13, it is determined whether or not the change area has a predetermined speed.
[0012]
Each condition is set separately, and only condition matching is described here. A change area that meets the above conditions is regarded as a report target in reference numeral 14. When a notification target is found, it is transmitted to the notification processing unit 7 which is the next process. On the other hand, if the condition is not met, it is considered that there is no issue target in reference numeral 15. In this case, nothing is reported to the next process.
[0013]
[Problems to be solved by the invention]
In the conventional monitoring image processing apparatus as described above, even if the change region extracted by the change region extraction unit 5 is a change region that is not a report target, if it satisfies the conditions of reference numerals 10 to 13, There was a problem of judging.
[0014]
In particular, when performing surveillance work outdoors, unnecessary detection may occur due to the effect of changing areas due to shadows caused by sunlight and the influence of changing areas caused by the movement of car headlights at night. Many.
[0015]
A specific description will be given taking the example shown in FIG. FIGS. 16A and 16B are temporally continuous images. In the figure (a), the puddle under the feet is calm, but in the figure (b) at the next moment, a water pattern is shown there.
[0016]
In this case, the change area seen from the apparatus is “water pattern”. The frame line written in the image is a circumscribed rectangle (ferret diameter) derived from the change area of “water pattern”. If the conditions of the change area due to the water pattern match the conditions indicated by reference numerals 10 to 13, it is determined that the condition is to be issued. In response to this, the issue processing unit 7 performs the issue processing.
[0017]
However, in this case, in a human sense, that is, for example, a person or a car does not appear on the screen. From the perspective of the device, "water pattern" appears on the screen, but it is not a thing from the human sense. Because of this gap, this notification is considered “unnecessary detection”.
[0018]
The present invention has been made to solve the above-mentioned problems. The unnecessary detection has the same scale as the conventional apparatus, is inexpensive, and does not significantly impair the processing speed. It is an object of the present invention to obtain a monitoring image processing apparatus capable of dealing with information so as not to be reported.
[0021]
[Means for Solving the Problems]
Monitoring the image processing apparatus according to the invention of this, if a change region extraction means for extracting difference data by comparing operation on the comparison image data and current image data, said differential data matches the predetermined condition originating A recognition processing means for recognizing as a reporting target, a disturbance determining means for determining that a disturbance has occurred when the difference data appears at a predetermined position at a predetermined time interval, and a notification target when the disturbance has occurred The difference data is a change area, and the disturbance determination means is configured such that the circumscribing rectangle of the change area appears at a predetermined position and at a predetermined time interval. It is determined that a disturbance has occurred.
[0024]
In the monitoring image processing apparatus according to the present invention, the contents of the restriction are omitted from the notification.
[0025]
In the monitoring image processing apparatus according to the present invention, the predetermined position is a predetermined area having a higher probability of occurrence of disturbance in the current image data than other areas.
[0026]
In the monitoring image processing apparatus according to the present invention, the predetermined time interval is regular.
[0027]
Further, the monitoring image processing apparatus according to the present invention includes an A / D conversion unit that quantizes an input video, a current image data calculation storage unit that prepares the current image data from an output of the A / D conversion unit, The image processing apparatus further includes comparison image data calculation storage means for creating the comparison image data from the output of the A / D conversion means.
[0028]
DETAILED DESCRIPTION OF THE INVENTION
The monitoring image processing apparatus according to the present invention has a function of determining that a disturbance has occurred in a video when a change region appears at a specific time interval at a specific location, and restricting the notification to that. As a result, unnecessary detection can be reduced.
[0029]
In addition, the monitoring image processing apparatus according to the present invention determines that a disturbance has occurred in the video when a change area appears at a specific time interval in a specific location, and restricts the notification to that. It has a function and its “specific location” is updated as needed based on the location information obtained as a result of the report. Can be reduced.
[0030]
Furthermore, the monitoring image processing apparatus according to the present invention determines that a disturbance has occurred in the video when a change area appears at a specific time interval in a specific location, and restricts the notification to that. When the function has a function and the "specific time interval" is updated as needed based on the time information obtained as a result of the report, and as a result, the frequency of disturbances occurring in the video changes slowly In this way, it is possible to reduce unnecessary detection.
[0031]
Embodiment 1 FIG.
The configuration of the first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a diagram showing an overall configuration of Embodiment 1 of the present invention. In addition, in each figure, the same code | symbol shows the same or equivalent part.
[0032]
In FIG. 1, 1 is a video input unit that receives a video signal output from a TV camera or the like installed at an electric station or the like, 2 is an A / D conversion unit that quantizes a video signal input from the video input unit 1, and 3 is Current image data calculation storage unit 4 that receives the output from the A / D conversion unit 2 and prepares the current image data. 4 receives the output from the A / D conversion unit 2 and performs corresponding calculations to create comparison image data. The comparison image data calculation storage unit 5 performs a comparison calculation between the current image data output from the current image data calculation storage unit 3 and the comparison image data output from the comparison image data calculation storage unit 4. It is a change area extraction part which extracts a change area. Generally, a binarization processing unit is built in here, and the output is converted into binary data.
[0033]
In the figure, 6 is a recognition processing unit that receives the output of the change region extraction unit 5 and determines whether or not it is an object, and 8 receives the output of the change region extraction unit 5 and generates disturbance. A disturbance determination unit 7A for determining whether or not a notification is generated, and a notification processing unit 7A that performs a notification process in response to instructions from the recognition processing unit 6 and the disturbance determination unit 8. In general, the recognition processing unit 6, the disturbance determination unit 8, and the alert processing unit 7A are processed by S / W.
[0034]
Next, the operation of the first embodiment will be described with reference to FIGS. FIG. 2 is a diagram for explaining the operation of the disturbance determination unit according to the first embodiment of the present invention. FIG. 3 is a diagram for explaining the operation of the alarm processing unit according to the first embodiment of the present invention.
[0035]
First, the overall schematic operation will be described. A video signal (video signal) given from the outside to the apparatus is received by the video input unit 1 and sent to the A / D conversion unit 2. The A / D converter 2 converts the received video signal into digital data for later processing. In general, conversion into multi-value data of 8 bits per pixel is performed.
[0036]
Next, the current image data calculation storage unit 3 prepares “current image data” based on the input data, and prepares for subsequent processing. In many cases, the input data itself is output. The comparison image data calculation storage unit 4 prepares “comparison image data” based on the input data, and prepares for subsequent processing. This is often the current image data periodically sampled or the video of the previous frame.
[0037]
Next, the change area extraction unit 5 compares the outputs of the current image data calculation accumulation unit 3 and the comparison image data calculation accumulation unit 4, finds a different part, and outputs the part as a change area. In general, the change area is mainly a moving object in an image, but there is a rare case where the brightness changes such as reflection of light.
[0038]
Next, the recognition processing unit 6 receives the change area data from the change area extraction unit 5 and determines whether or not it is a notification target. The details of this determination are the same as in the conventional example. On the other hand, the disturbance determination unit 8 receives the change area data from the change area extraction unit 5 and determines whether there is a disturbance in the change area data. Details of this determination will be described later.
[0039]
Next, the notification processing unit 7A receives the instructions from the recognition processing unit 6 and the disturbance determination unit 8 and performs a notification process. This notification process normally sends a specific signal to a PC or the like at the upper level or sounds a buzzer of the apparatus, but when a disturbance occurs, these processes are restricted.
[0040]
Next, the flow of determination processing performed by the disturbance determination unit 8 will be described. The disturbance determination unit 8 performs a determination process for each change region passed from the change region extraction unit 5 according to the flow shown in FIG. Conditional branches indicated by reference numerals 20 and 21 are portions for collating conditions for determining whether or not a disturbance has occurred. Here, the determination is performed under the following conditions.
[0041]
In reference numeral 20, it is determined whether or not a change area appears in a specific place.
At reference numeral 21, it is determined whether or not a change area appears at a specific time interval.
Since the setting of the “certain location, time interval” value is performed separately, only the determination is described here.
[0042]
The change region that matches the above conditions is regarded as a disturbance at 22. Here, the effects particularly obtained by this process are as follows. “Unnecessary detection can be reduced for areas of change that continue to occur at regular intervals, such as water drops caused by rain.” In general, the subject of the report is a person or a car. , Regularity in time. In the first embodiment, disturbances are determined for a notification target that repeats regularly in time, using the regular points as clues, the notification is restricted, and unnecessary detection is reduced. Further, when the occurrence of a disturbance is found, it is transmitted to the notification processing unit 7A, which is the next process. In addition, it may be transmitted to the recognition processing unit 6 and processed there. On the other hand, if the condition is not met, it is considered that no disturbance has occurred in reference numeral 23. In this case, nothing is reported to the next process.
[0043]
Next, the flow of the notification process when a disturbance occurs performed in the notification processing unit 7A will be described. Here, for convenience, it is described that the processing is performed by the notification processing unit 7A. However, as described above, the processing may be performed by the recognition processing unit 6.
[0044]
The notification processing unit 7A receives the notification of the occurrence of the disturbance passed from the disturbance determination unit 8, and performs the notification processing according to the flow as shown in FIG. The conditional branch indicated by reference numeral 30 is a part for checking whether or not a disturbance has occurred. When the notification of the occurrence of disturbance is received, at 32, the notification process is restricted. On the other hand, if no disturbance is generated, a normal notification process is performed at reference numeral 31.
[0045]
The details of the regulations are as follows.
In Example 1, no alarm is issued during the occurrence of a disturbance.
In Example 2, a part of the report is omitted during the occurrence of the disturbance.
In example 3, the condition of the recognition process is corrected by the apparatus itself during the occurrence of disturbance.
In Example 4, information indicating that a disturbance has occurred is notified to a host PC or the like.
[0046]
Further, a specific example will be given to explain how the process of the present invention solves the problems of the conventional apparatus with reference to FIG. 4A and 4B are temporally continuous images.
[0047]
In the figure (a), the puddle under the feet is calm, but in the figure (b) at the next moment, there is a water ripple due to rain. In this case, the change area seen from the apparatus is “water pattern”. A solid frame line written in the image is a circumscribed rectangle derived from the change area of “water pattern”. In the conventional apparatus, in this case, if the recognition processing unit 6 determines that the water pattern is a notification target, the notification processing unit 7 performs the notification process.
[0048]
In the monitoring image processing apparatus according to the first embodiment, the “position” of the change area and the “time interval from the previous report” are viewed. Temporarily, “a specific position” is defined as “an area at the lower right quarter of the screen”. This is because the area has been empirically found to be prone to puddles. Also, “a certain time interval” is defined as “within 60 seconds”. This is based on the assumption that the same event will be repeated within 60 seconds at the latest in the case of water ripples due to rain. In FIG. 5B, a water pattern is shown, and a change area due to the water pattern is in the “lower right quarter area of the screen”. Further, if water patterns are generated at intervals of about 5 seconds, this is always “within 60 seconds”. That is, the condition for disturbance determination is satisfied. Thereby, it is determined that “disturbance occurs” in the video. Accordingly, although the “water pattern” has been determined to be reported, the notification is restricted by the notification processing unit 7A, and the occurrence of unnecessary detection due to the “water pattern” is suppressed.
[0049]
On the other hand, a change area that appears in a place that is not the “lower right quarter of the screen” is not regarded as a disturbance, but this is probably a change area that is not related to a puddle. Therefore, it is correct to judge that it is not a disturbance and it can be left as it is. In addition, change areas that appear at time intervals exceeding “60 seconds” are not regarded as disturbances, but it can be determined that the rain has stopped if there is a gap between them. It is correct to judge that it is not a disturbance from the point that it may be a change area due to a moving object, and it can be left as it is. Only the change region of the condition where both are compatible can be determined as the change region due to the “water pattern” in this case.
[0050]
As described above, according to the first embodiment, when a change area that occurs at a specific time interval appears in a specific place, it is determined that a disturbance has occurred and the notification is restricted. It is possible to efficiently reduce “unnecessary detection for a change area that continuously occurs at a certain interval, such as a water drop caused by rain”.
[0051]
Embodiment 2. FIG.
The configuration of the second embodiment of the present invention will be described with reference to FIG. FIG. 5 is a diagram showing an overall configuration of the second embodiment of the present invention.
[0052]
In FIG. 5, 1 is a video input unit that receives a video signal output from a TV camera or the like installed at an electric station or the like, 2 is an A / D conversion unit that quantizes a video signal input from the video input unit 1, Current image data calculation storage unit 4 that receives the output from the A / D conversion unit 2 and prepares the current image data. 4 receives the output from the A / D conversion unit 2 and performs corresponding calculations to create comparison image data. The comparison image data calculation storage unit 5 performs a comparison calculation between the current image data output from the current image data calculation storage unit 3 and the comparison image data output from the comparison image data calculation storage unit 4. It is a change area extraction part which extracts a change area. Generally, a binarization processing unit is built in here, and the output is converted into binary data.
[0053]
In the figure, 6 is the recognition processing unit that receives the output of the change region extraction unit 5 and determines whether or not it is an object, and 8A receives the output of the change region extraction unit 5 and generates disturbance. A disturbance determination unit 7A that determines whether or not a notification is generated, and a notification processing unit 7A that performs a notification process in response to instructions from the recognition processing unit 6 and the disturbance determination unit 8A. In general, the recognition processing unit 6, the disturbance determination unit 8A, and the alert processing unit 7A are processed by S / W.
[0054]
Next, the operation of the second embodiment will be described with reference to FIGS. FIG. 6 is a diagram for explaining the operation of the disturbance determination unit according to the second embodiment of the present invention. FIG. 7 is a diagram for explaining the operation of the alarm processing unit according to the second embodiment of the present invention.
[0055]
First, the overall schematic operation will be described. A video signal (video signal) given from the outside to the apparatus is received by the video input unit 1 and sent to the A / D conversion unit 2. The A / D converter 2 converts the received video signal into digital data for later processing. In general, conversion into multi-value data of 8 bits per pixel is performed.
[0056]
Next, the current image data calculation storage unit 3 prepares “current image data” based on the input data, and prepares for subsequent processing. In many cases, the input data itself is output. The comparison image data calculation storage unit 4 prepares “comparison image data” based on the input data, and prepares for subsequent processing. This is often the current image data periodically sampled or the video of the previous frame.
[0057]
Next, the change area extraction unit 5 compares the outputs of the current image data calculation accumulation unit 3 and the comparison image data calculation accumulation unit 4, finds a different part, and outputs the part as a change area. In general, the change area is mainly a moving object in an image, but there is a rare case where the brightness changes such as reflection of light.
[0058]
Next, the recognition processing unit 6 receives the change area data from the change area extraction unit 5 and determines whether or not it is a notification target. The details of this determination are the same as in the conventional example. On the other hand, the disturbance determination unit 8A receives the change area data from the change area extraction unit 5, and determines whether or not there is a disturbance. Details of this determination will be described later.
[0059]
Next, the notification processing unit 7A receives the instructions from the recognition processing unit 6 and the disturbance determination unit 8A and performs notification processing. This notification process normally sends a specific signal to a PC or the like at the upper level or sounds a buzzer of the apparatus, but when a disturbance occurs, these processes are restricted.
[0060]
Next, the flow of determination processing performed by the disturbance determination unit 8A will be described. The disturbance determination unit 8A performs a determination process for each change region passed from the change region extraction unit 5 according to a flow as shown in FIG. Conditional branches indicated by reference numerals 40 and 41 are parts for collating conditions for determining whether or not a disturbance has occurred. Here, the determination is performed under the following conditions.
[0061]
In reference numeral 40, it is determined whether or not a change area appears in a specific place.
In reference numeral 41, it is determined whether or not a change area appears at a specific time interval.
The initial value of “a specific place, time interval” is set separately, and only the determination is described here.
[0062]
In the first embodiment, the change region that meets the above conditions is regarded as a disturbance at reference numeral 42. Here, the effect obtained particularly by this process is that “unnecessary detection can be reduced with respect to a change region that continuously occurs at a certain interval, such as water droplets due to rain”, as described in the first embodiment. It is done.
[0063]
In the second embodiment, simple knowledge processing is further added thereto. Reference numeral 44 denotes a knowledge processing unit that updates the “certain location” based on the position information of the change area that occurred in the past. The update rule stores, for example, a change area that has occurred most recently or a circumscribed rectangle of the change area related to the report. In addition, a certain peripheral range is regarded as a “certain location” after the update. Inevitably, as a result of the update, the “certain location” depends on the location where the report is repeatedly generated.
[0064]
Originally, the object of the present invention is to reduce unnecessary detection for a change area that continues to occur at a certain interval, such as water droplets caused by rain. Change the position. In order to follow this, it is necessary to have a function of tracking a place where a change area frequently occurs. The second embodiment provides this means. This is a place where the knowledge processing unit 44 actually processes the above update rule.
[0065]
When the occurrence of a disturbance is found, it is transmitted to the notification processing unit 7A, which is the next process. In addition, it may be transmitted to the recognition processing unit 6 and processed there. On the other hand, if the condition is not met, it is considered that no disturbance has occurred in reference numeral 43. In this case, nothing is reported to the next process.
[0066]
Next, the flow of the notification process when a disturbance occurs performed in the notification processing unit 7A will be described. Here, for convenience, it is described that the processing is performed by the notification processing unit 7A. However, as described above, the processing may be performed by the recognition processing unit 6.
[0067]
The notification processing unit 7A receives the notification of the occurrence of the disturbance passed from the disturbance determination unit 8A, and performs the notification processing according to the flow as shown in FIG. The conditional branch indicated by reference numeral 30 is a part for checking whether or not a disturbance has occurred. When the notification of the occurrence of disturbance is received, at 32, the notification process is restricted. On the other hand, if no disturbance is generated, a normal notification process is performed at reference numeral 31.
[0068]
The details of the regulations are as follows.
In Example 1, no alarm is issued during the occurrence of a disturbance.
In Example 2, a part of the report is omitted during the occurrence of the disturbance.
In example 3, the condition of the recognition process is corrected by the apparatus itself during the occurrence of disturbance.
In Example 4, information indicating that a disturbance has occurred is notified to a host PC or the like.
[0069]
Further, a specific example will be given to explain how the process of the present invention solves the problems of the conventional apparatus with reference to FIG. FIGS. 8A and 8B are temporally continuous images.
[0070]
In the figure (a), the puddle under the feet is calm, but in the figure (b) at the next moment, a water pattern is shown there. In this case, the change area seen from the apparatus is “water pattern”. A solid frame line written in the image is a circumscribed rectangle derived from the change area of “water pattern”. In the conventional apparatus, in this case, if the recognition processing unit 6 determines that the water pattern is a notification target, the notification processing unit 7 performs the notification process.
[0071]
In the monitoring image processing apparatus according to the second embodiment, the “position” of the change area and the “time interval from the previous report” are as described in the first embodiment. In the first embodiment, “a specific position” is defined as “an area at the lower right quarter of the screen”, but here, for example, it is updated as follows.
[0072]
The circumscribed rectangle of the related change area at the time of the latest report is stored. Further, a certain range around the circumscribed rectangle is set as a “certain specific position” for the next time. The reason why the “certain position” is defined as “the area at the lower right quarter of the screen” is based on the reason that it is empirically known that the puddle is easily formed. If the update is performed as described above, the “certain location” automatically comes in the puddle according to the actual puddle. Therefore, the position accuracy is improved.
[0073]
In FIG. 5B, a water pattern is shown, but the position where the change area due to the water pattern immediately before this is present is the position of the dotted line in the figure. Further, “a certain time interval” is defined as “within 60 seconds”. The current waterprint is in contact with the previous change area position. Also, if water ripples occur at intervals of about 5 seconds, this is always “within 60 seconds”. That is, the condition for disturbance determination is satisfied. Thereby, it is determined that “disturbance occurs” in the video. Accordingly, although the “water pattern” has been determined to be reported, the notification is restricted by the notification processing unit 7A, and the occurrence of unnecessary detection due to the “water pattern” is suppressed.
[0074]
Further, the difference from the first embodiment is the position followability. In the first embodiment, even if a water pattern occurs outside the range of the lower right quarter of the screen, it is not subject to disturbance. According to the second embodiment, even if the water puddle spreads and a water pattern is generated outside the range of the lower right quarter of the screen, the “specific position” is gradually updated to the actual puddle position. As in the first embodiment, there is little risk that the disturbance cannot be determined halfway. As described above, the feature is the freedom of tracking the event and determining the disturbance and the positional accuracy.
[0075]
On the contrary, a change area that appears in a place away from the “updated“ specific place ”” is not regarded as a disturbance, but this change area is probably a change area that is not related to a puddle. It is correct to judge that it is not a disturbance from the deaf point, and it can be left as it is.
[0076]
As described above, according to the second embodiment, when a change region that occurs at a specific time interval appears at a specific location, it is determined that a disturbance has occurred and the notification is restricted. It is possible to efficiently reduce unnecessary detection for a change area that continues to occur at a certain interval, such as water droplets due to rain. In addition, a specific location can be determined from the position information of the change area that occurred last time. Since it is updated as appropriate to bring it closer to the place where the event has occurred, it is possible to flexibly follow the position where the water pattern moves due to changes in the puddle, etc., while reducing unnecessary detection. Contrary to the intention, there is an effect of suppressing the harmful effect of mistaking the report target as a disturbance.
[0077]
Embodiment 3 FIG.
The configuration of the third embodiment of the present invention will be described with reference to FIG. FIG. 9 is a diagram showing an overall configuration of the third embodiment of the present invention.
[0078]
In FIG. 9, 1 is an input unit that receives a video signal output from a TV camera or the like installed at an electric station or the like, 2 is an A / D conversion unit that quantizes a video signal input from the video input unit 1, and 3 is A A current image data calculation / storage unit 4 that receives the output from the / D conversion unit 2 and prepares the current image data, and 4 receives the output from the A / D conversion unit 2 and performs corresponding calculations to create comparison image data. The comparison image data calculation storage unit 5 compares the current image data output from the current image data calculation storage unit 3 with the comparison image data output from the comparison image data calculation storage unit 4, and changes It is a change area extraction part which extracts an area. Generally, a binarization processing unit is built in here, and the output is converted into binary data.
[0079]
In the figure, 6 is the recognition processing unit that receives the output of the change region extraction unit 5 and determines whether or not it is a target object, and 8B receives the output of the change region extraction unit 5 and causes disturbance. A disturbance determination unit 7A for determining whether or not a notification is generated, and a notification processing unit 7A that performs a notification process in response to instructions from the recognition processing unit 6 and the disturbance determination unit 8B. In general, the recognition processing unit 6, the disturbance determination unit 8B, and the alert processing unit 7A are processed by S / W.
[0080]
Next, the operation of the third embodiment will be described with reference to FIGS. FIG. 10 is a diagram for explaining the operation of the disturbance determination unit according to the third embodiment of the present invention. FIG. 11 is a diagram for explaining the operation of the alarm processing unit according to the third embodiment of the present invention.
[0081]
First, the overall schematic operation will be described. A video signal (video signal) given from the outside to the apparatus is received by the video input unit 1 and sent to the A / D conversion unit 2. The A / D converter 2 converts the received video signal into digital data for later processing. In general, conversion into multi-value data of 8 bits per pixel is performed.
[0082]
Next, the current image data calculation storage unit 3 prepares “current image data” based on the input data, and prepares for subsequent processing. In many cases, the input data itself is output. The comparison image data calculation storage unit 4 prepares “comparison image data” based on the input data, and prepares for subsequent processing. This is often the current image data periodically sampled or the video of the previous frame.
[0083]
Next, the change area extraction unit 5 compares the outputs of the current image data calculation accumulation unit 3 and the comparison image data calculation accumulation unit 4, finds a different part, and outputs the part as a change area. In general, the change area is mainly a moving object in an image, but there is a rare case where the brightness changes such as reflection of light.
[0084]
Next, the recognition processing unit 6 receives the change area data from the change area extraction unit 5 and determines whether or not it is a notification target. The details of this determination are the same as in the conventional example. On the other hand, the disturbance determination unit 8B receives the change area data from the change area extraction unit 5, and determines whether or not there is a disturbance. Details of this determination will be described later.
[0085]
Next, the notification processing unit 7A receives the instructions from the recognition processing unit 6 and the disturbance determination unit 8B and performs a notification process. This notification process normally sends a specific signal to a PC or the like at the upper level or sounds a buzzer of the apparatus, but when a disturbance occurs, these processes are restricted.
[0086]
Next, the flow of determination processing performed by the disturbance determination unit 8B will be described. The disturbance determination unit 8B performs the determination process for each change region passed from the change region extraction unit 5 in the flow as shown in FIG. Conditional branches indicated by reference numerals 50 and 51 are parts for collating conditions for determining whether or not a disturbance has occurred. Here, the determination is performed under the following conditions.
[0087]
In reference numeral 50, it is determined whether or not a change area appears in a specific place.
In reference numeral 51, it is determined whether or not a change area appears at a specific time interval.
The initial value of “a specific place, time interval” is set separately, and only the determination is described here.
[0088]
In the first embodiment, the change region meeting the above condition is regarded as a disturbance with reference numeral 22. Here, the effect obtained particularly by this process is that “unnecessary detection can be reduced with respect to a change region that continuously occurs at a certain interval, such as water droplets due to rain”, as described in the first embodiment. It is done.
[0089]
In the third embodiment , simple knowledge processing is further added thereto. Reference numeral 54 denotes a knowledge processing unit that updates the “certain time interval” based on time information of a change area that occurred in the past and a related change area at the time of reporting. The update rule stores, for example, the most recent change area or the occurrence time of the change area related to the report. If the time is larger than the current “time interval”, the current “time interval” is updated by a specific time. There is an upper limit to increase. Conversely, if there is no report for a long time, the current “time interval” is updated by reducing the specific time. There is a lower limit to reduce. Inevitably, as a result of the update, the “certain time interval” comes at the rate of occurrence of repeated alerts.
[0090]
Originally, an object of the present invention is to reduce unnecessary detection for a change area that continues to occur at a certain interval, such as a water drop caused by rain. Along with this, the time interval of the occurrence is changed. In order to follow this, it is necessary to have a function of appropriately updating the setting to the “time interval” that matches the occurrence frequency of the current event. The third embodiment provides this means. This is a place where the knowledge processing unit 54 actually processes the update rule.
[0091]
When the occurrence of a disturbance is found, it is transmitted to the notification processing unit 7A, which is the next process. In addition, it may be transmitted to the recognition processing unit 6 and processed there. On the other hand, if the condition is not met, it is considered that no disturbance has occurred in 53. In this case, nothing is reported to the next process.
[0092]
Next, the flow of the notification process when a disturbance occurs performed in the notification processing unit 7A will be described. Here, for convenience, it is described that the processing is performed by the notification processing unit 7A. However, as described above, the processing may be performed by the recognition processing unit 6.
[0093]
The notification processing unit 7A receives the notification of the occurrence of the disturbance passed from the disturbance determination unit 8B, and performs the notification processing according to the flow as shown in FIG. The conditional branch indicated by reference numeral 30 is a part for checking whether or not a disturbance has occurred. When the notification of the occurrence of disturbance is received, at 32, the notification process is restricted. On the other hand, if no disturbance is generated, a normal notification process is performed at reference numeral 31.
[0094]
The details of the regulations are as follows.
In Example 1, no alarm is issued during the occurrence of a disturbance.
In Example 2, a part of the report is omitted during the occurrence of the disturbance.
In example 3, the condition of the recognition process is corrected by the apparatus itself during the occurrence of disturbance.
In Example 4, information indicating that a disturbance has occurred is notified to a host PC or the like.
[0095]
Furthermore, with specific examples, how the process of the present invention solves the problems of the conventional apparatus will be described with reference to FIGS. 12 and 13 are diagrams showing how the apparatus works in time.
[0096]
“Time” that has elapsed from the left, “Event” of the water pattern that has occurred and “Whether the device has reported it” or not. This is a “deterrent effect”.
[0097]
Water ripples are generated a total of 9 times in 0, 50, 60, 80, 100, 120, 150, 210, and 270 seconds, and the conventional apparatus issues notifications (1) to (9) respectively. Suppose it was a thing. Depending on circumstances, in the description of FIG. 12 and FIG.
[0098]
First, FIG. 12 shows an operation performed by the method in the first embodiment. For example, “within a certain time interval” is fixed as “within 40 seconds”, and the operation is performed without updating. For the sake of simplicity, “a particular place” is not discussed here.
[0099]
The device that has issued the alarm (1) with the water pattern of 0 seconds suppresses the alarm for 40 seconds thereafter, but the suppression effect disappears after 40 seconds. However, in the report (2) by the water pattern that occurred in 50 seconds, the report (3) is suppressed, and the report (4) is suppressed by the (suppressed) report (3). Suppression progresses in order, and finally it was not possible to suppress completely, but it was issued four times, (1), (2), (8), (9).
[0100]
In the first embodiment, as described above, since the suppression time is always fixed, the notification cannot be suppressed if the interval exceeds that. The reason why it was not possible to prevent unnecessary detection at 210 and 270 seconds was because “a certain time interval” was fixed as “within 40 seconds” as an example, but if this was set to 100 seconds, the problem was not solved. . In this case, in order to suppress the notification for 100 seconds, there may be a case where the target (person, car) to be detected is determined as a disturbance and the notification is suppressed. The opposite negative effect will be noticeable.
[0101]
In the third embodiment, this “certain time interval” is appropriately updated. In updating, the following rules are installed as an example.
A) The initial value of “time interval” is 0, and the upper limit of “time interval” is 100 seconds.
B) If there is a report and the time from the previous report is less than the current “time interval”, the report is suppressed.
C) If there is a report and the time from the previous report is more than the current “time interval”, the report is not suppressed. The “time interval” is added by 20 seconds. However, it will not be added beyond the upper limit.
D) If there is no report (including a suppressed report) for a time equivalent to the upper limit of “time interval”, the current “time interval” is subtracted by 20 seconds.
[0102]
FIG. 13 shows a state in which the operation is performed according to this rule. The device that has issued the alert (1) with a 0 second waterprint suppresses the alert for 20 seconds thereafter, but the restraining effect disappears after 20 seconds. However, in the bulletin (2) that occurred in 50 seconds, the time interval is updated to 40 seconds (rule C), and the (3) is suppressed, and the (suppressed) notification (3) In this way, the deterrence proceeds in the manner that the alert (4) is deterred.
[0103]
The issue (8) could not be suppressed because the time interval was 40 seconds, but next, the issue (9) was suppressed because the time interval increased to 60 seconds. Eventually, it was reported three times, (1), (2), and (8). However, unnecessary detection can be suppressed even if there is a subsequent water pattern. This point is different from the first embodiment.
[0104]
In the third embodiment, since the time interval is updated in this way, when a report is generated, the time interval is increased accordingly and the update is continued until unnecessary detection disappears. As a result, reporting is suppressed at any pace according to the actual situation. The object of the present invention is that “unnecessary detection can be reduced for a change area that continues to occur at a certain interval, such as water droplets due to rain”. It is possible to flexibly follow the interval. Note that the time interval increased to 60 seconds is appropriately reduced by the rule D and becomes 0 at 630 seconds.
[0105]
In this case, what is particularly important is the meaning of “update of time interval”. Although it is not possible to prevent the occurrence of events exceeding 40 seconds at a fixed time interval, it will occur repeatedly when it is updated according to the rules as in this Embodiment 3, and continues to occur at certain intervals, such as raindrops. As for the notification by the “change area”, etc., as a result of the time interval being updated, it becomes possible to finally catch up and suppress the notification by “following the time interval that fluctuates due to changes in rainfall, etc.” . As a result, it is possible to accurately suppress repetitive reporting that occurs at any time interval.
[0106]
Due to the occurrence of the disturbance determined by the above process, although the “water pattern” has been determined to be the target of the report retroactively, the notification processing unit 7A restricts the notification, and unnecessary detection by the “water pattern”. The occurrence of is suppressed.
[0107]
As described above, according to the third embodiment, when a change region that occurs at a specific time interval appears in a specific place, it is determined that a disturbance has occurred and the notification is restricted. Efficiently "can reduce unnecessary detection for a change area that continues to occur at a certain interval, such as water droplets due to rain" and time information on the change area that occurred last time at a specific time interval From time to time, it will be updated as close as possible to the frequency of occurrence of events, so it will be possible to "follow flexibly the time interval that fluctuates due to changes in rainfall, etc." Contrary to this, it has the effect of suppressing the harmful effect of mistaking the report object as a disturbance.
[0112]
【The invention's effect】
Monitoring the image processing apparatus according to the invention this is, as described above, a change region extraction means for extracting difference data by comparing operation on the comparison image data and current image data, the difference data is a predetermined condition A recognition processing means for recognizing it as a notification target when it matches, a disturbance determination means for determining that a disturbance has occurred when the difference data appears at a predetermined position and at a predetermined time interval, and the disturbance has occurred And a notification processing means for restricting the processing of the notification target, wherein the difference data is a change area, and the disturbance determination means is configured so that a circumscribed rectangle of the change area is at a predetermined position and a predetermined time interval. If it appears in the case, it is determined that a disturbance has occurred, so it is possible to reduce unnecessary detection efficiently with a simple process, and at the same time, it is possible to suppress the harmful effects of misidentifying the report target as a disturbance. Achieve the cormorant effect.
[0113]
In addition, as described above, the monitoring image processing apparatus according to the present invention has the effect that unnecessary detection can be reduced because the contents of the restriction are omitted from the notification.
[0114]
In addition, as described above, the monitoring image processing apparatus according to the present invention is unnecessary because the predetermined position is a predetermined area in the current image data in which the probability of occurrence of disturbance is higher than other areas. There is an effect that detection can be reduced.
[0115]
Further, as described above, the monitoring image processing apparatus according to the present invention regularizes the predetermined time interval, so that unnecessary detection can be efficiently reduced, and at the same time, the subject to be notified is disturbed. There is an effect that it is possible to accurately suppress the harmful effects of misunderstanding.
[0116]
Further, as described above, the monitoring image processing apparatus according to the present invention includes an A / D conversion unit that quantizes an input video, and current image data that prepares the current image data from an output of the A / D conversion unit. Since it further comprises a calculation storage means and a comparison image data calculation storage means for creating the comparison image data from the output of the A / D conversion means, unnecessary detection is corrected internally before the notification. There is an effect that it is possible to cope with it so as not to reach.
[Brief description of the drawings]
FIG. 1 is a diagram showing an overall configuration of a first embodiment of the present invention.
FIG. 2 is a diagram for explaining an operation of a disturbance determination unit according to the first embodiment of the present invention.
FIG. 3 is a diagram for explaining the operation of an alarm processing unit according to the first embodiment of the present invention.
FIG. 4 is a diagram showing a specific example of the first embodiment of the present invention.
FIG. 5 is a diagram showing an overall configuration of a second embodiment of the present invention.
FIG. 6 is a diagram for explaining the operation of a disturbance determination unit according to the second embodiment of the present invention.
FIG. 7 is a diagram for explaining the operation of an alarm processing unit according to the second embodiment of the present invention.
FIG. 8 shows a specific example of the second embodiment of the present invention.
FIG. 9 is a diagram showing an overall configuration of a third embodiment of the present invention.
FIG. 10 is a diagram for explaining the operation of a disturbance determination unit according to Embodiment 3 of the present invention.
FIG. 11 is a diagram for explaining the operation of an alarm processing unit according to the third embodiment of the present invention.
FIG. 12 shows a specific example of the third embodiment of the present invention.
FIG. 13 shows a specific example of the third embodiment of the present invention.
FIG. 14 is a diagram illustrating an overall configuration of a conventional monitoring image processing apparatus.
FIG. 15 is a diagram for explaining the operation of a recognition processing unit of a conventional monitoring image processing apparatus.
FIG. 16 is a diagram illustrating a specific example of a conventional monitoring image processing apparatus.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Video input part, 2 A / D conversion part, 3 Current image data calculation accumulation | storage part, 4 Comparison image data calculation accumulation | storage part, 5 Change area extraction part, 6 Recognition processing part, 7A Notification processing part, 8, 8A, 8B Disturbance determination unit.

Claims (5)

A change area extracting means for comparing the current image data and the comparison image data to extract difference data;
Recognizing processing means for recognizing as a notification target when the difference data meets a predetermined condition;
Disturbance determination means for determining that a disturbance has occurred when the difference data appears at a predetermined position and at a predetermined time interval;
A notification processing means for restricting the processing of the notification target when the disturbance occurs,
The difference data is a change area,
The monitoring image processing apparatus according to claim 1, wherein the disturbance determining unit determines that a disturbance has occurred when a circumscribed rectangle of the change area appears at a predetermined position and at a predetermined time interval . The monitoring image processing apparatus according to claim 1 , wherein the content of the regulation is a report omission . 2. The monitoring image processing apparatus according to claim 1 , wherein the predetermined position is a predetermined area having a higher probability of occurrence of disturbance in the current image data than other areas . The monitoring image processing apparatus according to claim 1 , wherein the predetermined time interval is regular . A / D conversion means for quantizing the input video;
Current image data calculation and storage means for preparing the current image data from the output of the A / D conversion means;
2. The monitoring image processing apparatus according to claim 1 , further comprising comparison image data calculation and storage means for creating the comparison image data from the output of the A / D conversion means .

Images