JP2008042423A - Image monitoring apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain a high precision image monitoring apparatus which accurately detects an object even if a background and a detection object have similar luminance, gets rid of missing a report, and reduces an erroneous report. <P>SOLUTION: The apparatus captures an image photographed by a camera, and calculates a feature amount for a variance region which is extracted by a variance region extraction means from an input image. When the calculated feature amount is smaller than an amount of the detection object, an expansion region determination means expands a region around the extracted variance region, carries out the correlation of the background and the input image, and determines a region whose similarity degree is lower than a threshold as an expansion region. In the determined expansion region, a sensitivity-up processing means performs the luminance conversion of the expansion region for the background and the input image, corrects the luminance, an object extraction means extracts an object using the corrected image, and an object detection discrimination processing means detects whether an object exists or not for the object extracted by the object extraction means. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image monitoring apparatus for monitoring, and more particularly to an image monitoring apparatus that realizes accurate detection of a target object.

  Among image monitoring apparatuses in the case where the background and the target brightness are similar, a technique for detecting a change area by performing gain increase on an image obtained by subtracting the background is disclosed.

  Further, a technique for detecting a preset target, setting a correction parameter based on the detection result, and correcting the detection target is disclosed.

JP 2003-274395 A Japanese Patent Laid-Open No. 2004-236110

  In the monitoring image processing apparatus, for example, when the detected change area is a regular notification target but the background and the target have similar brightness, in the situation where the accurate area of the detection target cannot be extracted, the parameter for calculating the feature amount No report is issued because the condition does not match (for example, area value). Therefore, it will be in an undetected state (hereinafter referred to as unreported). Therefore, it is necessary to eliminate misreports and reduce false alarms.

  In addition, when the background and the target have similar brightness, only a part (minimum) of the target is detected, so that the object cannot be detected accurately. Even if the correction parameter is set based on this result, it is not a parameter suitable for object detection, and therefore accurate correction cannot be performed. Therefore, it is necessary to accurately detect the target object.

  In Patent Document 1 of the prior art, a luminance difference between an input image and a background image is calculated to determine whether the background image is a low luminance area or a high luminance area, and further, whether the luminance difference is a positive value or a negative value is determined. The luminance difference is increased by a predetermined gain according to the determination result. An image taken with a camera often generates noise (for example, CCD noise) or disturbance (for example, light or shadow) in the entire image, and these are also amplified, so that false alarms can be reduced, but false alarms increase. There was a problem that.

  Further, in Patent Document 2 of the prior art, it is detected whether a preset target is included in the input image, and a correction parameter is set based on this result to correct the input image. When the background and the target have similar brightness, only a part (minimum) area of the target is detected, and thus the object cannot be detected accurately. Even if the correction parameter is set based on this result, it is not a parameter suitable for object detection, so that there is a problem that accurate correction cannot be performed.

  In view of such problems, the present invention provides an image monitoring apparatus that accurately detects an object even when the background and the detection target have similar luminance, eliminates false alarms, and reduces false alarms. It is aimed.

  In order to solve the above-mentioned problems, the present invention provides an image monitoring apparatus for monitoring an image taken by a camera, a change area extracting means for capturing an image taken by the camera and extracting a change area, and the extracted change. If the area is smaller than the size of the detection target, an expanded area determination unit that expands the area around the extracted change area, a sensitivity correction processing unit that corrects the luminance of the expanded area, and a sensitivity correction by the sensitivity correction processing unit And an object extracting means for detecting an object using the uploaded image.

  Furthermore, in order to solve the above-described problem, the present invention provides an image monitoring apparatus that monitors an image captured by a camera, a change area extraction unit that captures an image captured by the camera and extracts a change area, and the extraction If the changed area is smaller than the size of the detection target, an extended area determining means for expanding the area around the extracted changed area, an illumination lighting control means for illuminating the expanded area with aiming, and the illumination The image processing apparatus includes: an object extraction unit that captures an image irradiated by the lighting control unit with a camera, extracts a change region by the object extraction processing unit, and detects an object in the extracted change region.

  In the image monitoring apparatus of the present invention, the change area extraction means uses an image in which one frame or more is accumulated.

  In the image monitoring apparatus according to the present invention, the extended area determination means extends the area using the calculated size of the object to be detected.

  In the image monitoring apparatus of the present invention, the sensitivity correction processing unit performs luminance conversion using the luminance distribution of the region determined by the extended region determining unit.

  In the image monitoring apparatus according to the present invention, when the object extracted by the object extraction unit is determined to be abnormal by the object detection determination processing unit that determines whether the object is abnormal or not, and the object detection determination processing unit. It is provided with the detection result alerting | reporting means alert | reported to a monitoring person.

  In the image monitoring apparatus according to the present invention, the detection result notifying unit displays an image whose sensitivity is increased by the sensitivity increasing processing unit.

  In the image monitoring apparatus of the present invention, the illumination lighting control means uses an illumination device capable of controlling illumination lighting.

  Further, in the image monitoring apparatus of the present invention, the detection result notifying unit displays an image irradiated with the object illuminated by the illumination lighting control unit.

  According to the present invention, there is an effect that it is possible to realize an image monitoring apparatus that accurately detects an object even when the background and the detection target have similar brightness, eliminates misreporting, and reduces false alarms.

  In addition, it saves the trouble of monitoring the monitor and can monitor only when there is an abnormality. If the detected object (for example, a person) is in an abnormal state, the observer can immediately grasp the state. There is an effect that appropriate measures can be taken.

  An example of an image monitoring apparatus using the present invention will be described with reference to the drawings.

  An embodiment will be described with reference to FIG.

In FIG. 1, when the camera 10 captures a target object (for example, a person) 40 of the monitoring target scene 30 and transmits the video signal 20 from the camera to the image processing device 50, the image processing device 50 displays the video signal 20. When the target object 40 exists, the presence / absence of the abnormality is detected, and when the abnormality occurs, the display device 900 indicates the location and time of occurrence of the abnormality of the target object 40 including the video of the monitoring target scene 30. It has a role to inform. There is an effect that the monitor 1000 can give an appropriate instruction in accordance with the place of occurrence and the situation of occurrence from the display content of the display device 900. In addition, the video of the monitoring target scene 30 is displayed on the display device 900 only when an abnormality occurs in the target object 40. On the other hand, since the video of the monitoring target scene 30 is not displayed on the display device 900 at the normal time, there is an effect that it is possible to efficiently monitor by eliminating the trouble of the monitor watching the monitor. Furthermore, since only one camera 10 may be provided for the monitoring target scene 30, there is an effect that an inexpensive monitoring device can be realized. Further, the image processing apparatus 50 may be configured by a PC capable of performing image processing or an image processing dedicated apparatus.

  The image processing apparatus 50 includes an image input unit 200, a change region extraction unit 300, an extended region determination unit 400, a sensitivity enhancement processing unit 500, an object extraction unit 550, an object detection determination processing unit 600, and a detection result notification unit 850.

In the processing in the image processing apparatus 50, first, when the image input unit 200 inputs the video signal 20 of the camera 10, the change region extraction unit 300 extracts the change region from the background image and the input image, removes the noise portion, and removes each label. Is calculated (hereinafter referred to as label area). Using the label area calculated by the change region extraction unit 300, the extended region determination unit 400 determines whether the calculated label area is less than the size of the target object 40 (less than the threshold value). Assuming that only a part is extracted, the region is expanded around the extracted change region, the background image and the input image are correlated to determine a region whose similarity is equal to or less than the threshold, and the sensitivity increase processing unit 500 The sensitivity of the area determined by the extended area determination unit 400 is increased. When the object extraction unit 550 extracts an object region using the background image and the input image with respect to the region whose sensitivity has been increased, the object detection determination processing unit 600 detects the presence or absence of an object. The presence / absence of an abnormality of the object is determined, and the detection result notification unit 850 is notified.

  When the calculated feature amount is the size of the target object 40 (greater than or equal to the threshold value), the extended region determination unit 400 detects the presence or absence of an object from the region extracted by the change region extraction unit 300, and If there is an object, the presence / absence of the abnormality of the object is determined, and the detection result notification unit 850 is notified.

  These image input unit 200, change region extraction unit 300, extended region determination unit 400, sensitivity enhancement processing unit 500, object extraction unit 550, object detection determination processing unit 600, and detection result notification unit 850 are set at predetermined time intervals, for example, It is repeatedly executed in a short time such as one cycle of processing.

  As a result, even when the background and the detection target have similar brightness, it is possible to detect the object accurately in real time, eliminate the false alarm, and reduce the false alarm.

FIG. 2 is a flowchart illustrating an example of a processing procedure of the image processing apparatus 50. First, using the image input by the image input unit 200, the step 252 converts the input image to n (for example, 1 ≦ n ≦
10) Accumulate input frames and accumulate input images (hereinafter referred to as input images). Step 251 sets the initial input image as the initial background image.

  Next, the background difference image created in step 253 is binarized with a predetermined threshold value in step 254, and it is determined whether or not a change area is extracted using the area of the binary image. The threshold value of whether or not the change area can be extracted in step 254 is a value that excludes noise or the like (in the case of only noise, the label area is 20 or less, for example). When there is no change area in step 254, step 210 accumulates the background image of step 251 and the input image for n (for example, 1 ≦ n ≦ 10) frames to accumulate the background image, and the background image of step 251 is accumulated. Update.

  In the case where there is a change area in step 254, the target area is checked if step 255 checks the label area of the change area and is equal to or greater than a threshold (the target object extraction threshold depends on the shooting conditions of the target object and the camera). Step 599 is performed on the extracted object region.

  If the label area of the extraction region is less than the threshold value in step 254, step 256 is processed assuming that only a part of the target object can be extracted.

  In step 256, an extended region in which an entire part of the extracted target object is included is calculated. The size of the extended area in step 256 is obtained in step 260, and the size is determined from the size of one target object and the shooting conditions of the camera.

  In step 290, whether or not the extended region obtained in step 256 is appropriate is determined based on the normalized correlation. In step 291, when the normalized correlation value calculated in step 290 is greater than or equal to the threshold value, it is assumed that the background image and the input image are similar in the extended region obtained in step 256, that is, the target object does not exist. In step 293, it is checked whether or not all the extended areas have been completed. If not, the process returns to step 256 to change the extended areas and calculate again. If the normalized correlation value is less than the threshold value in step 291, the background image and the input image are not similar in the extended region obtained in step 256, that is, the target object exists, step 270 is the extended region in step 256. Increase sensitivity.

  In step 292, object extraction is performed using the image whose sensitivity has been increased in step 270, and in step 599, the presence or absence of an object is detected.

  The step 290 is determined by the normalized correlation, but instead of using the average luminance value or the average luminance value of the edge.

FIG. 3 is a flowchart showing another processing procedure example of the image processing apparatus 50. In step 270, the sensitivity of background image accumulation and input image accumulation is increased for the extended region in step 256, and step 290 determines whether or not the extended region of the image whose sensitivity has been increased is appropriate based on the normalized correlation.

  In step 291, when the normalized correlation value calculated in step 290 is greater than or equal to the threshold value, it is assumed that the background image and the input image are similar in the extended region obtained in step 256, that is, the target object does not exist. In step 293, it is checked whether or not all the extended areas have been completed. If not, the process returns to step 256 to change the extended areas and calculate again.

  In step 292, object extraction is performed using the image whose sensitivity has been increased in step 270, and in step 599, the presence or absence of an object is detected.

  FIG. 4 is a flowchart illustrating an example of a processing procedure for extracting a change region in the change region extraction unit 300. First, step 301 captures an image and sets it as an input image. In step 302, input images are accumulated (input images) by accumulating n (for example, 1 ≦ n ≦ 10) frames of input images. Step 303 checks whether or not a background image has been created. If a background image has not been created, step 304 is processed. If a background image exists, step 305 is processed. Step 304 creates an initial background image from the input image, and uses the image as a background image. The creation of the initial background image in step 304 uses at least one frame of an input image in a state where the target object 40 does not exist in the shooting scene. In step 305, the background image of step 304 is accumulated for n (for example, 1 ≦ n ≦ 10) frames to perform background image accumulation (background image).

Step 306 creates a difference image of a luminance difference for each pixel from the input image of step 302 and the background image of step 305, and step 307 changes a predetermined threshold (about 8 to 14) or more from the difference image. Extract as Thereafter, step 308 calculates the label area of the change region and checks whether or not it is equal to or greater than a threshold (for example, about 20 for noise exclusion). If it is equal to or greater than the threshold, step 309 is processed. Step 310 is processed with no change. In step 310, the input image without change is accumulated as a new background image, and the background is updated. In step 309, the changed region extracted in step 307 is shaped by performing noise removal (for example, the label area is 20 or less) and the like, and then the neighborhood is integrated by expansion and / or contraction to create an extraction region. This is the object area.

FIG. 5 is an explanatory diagram illustrating an example of a determination procedure performed by the extended region determination unit 400. Since the extracted target object 60 is sufficiently smaller than the size of one target object, only a part is extracted. Therefore, by using a circumscribed rectangle of one target object size calculated by the object size calculation unit 260,
First, an extended region 50a in which a part of the target object 40 extracted in the first time 91 is included is calculated. A similarity is calculated by performing a normalized correlation 80 between the background image 30a and the input image 30 for the extended region 50a. When the similarity is checked, in the case of the extended region 50a, most of the background is the background, the similarity is 81 over a threshold (for example, 0.6), and the region for extracting the object is not appropriate.

  Next, an extended region 50b in which a part of the target object 40 extracted in the second time 92 is included on the center side is calculated. A similarity is calculated by performing a normalized correlation 80 between the background image 30a and the input image 30b for the extended region 50b. When the similarity is checked, in the case of the extended region 50b, there are many backgrounds, the similarity is equal to or greater than a threshold (for example, 0.6) 82, and the region from which the object is extracted is not appropriate.

  Finally, an extended region 50c including a part of the target object 40 extracted in the third time 93 is calculated. A similarity is calculated by performing a normalized correlation 80 between the background image 30a and the input image 30 for the extended region 50c. In the case of the extended region 50c, the similarity is checked, and most of the region is an object region. Therefore, the similarity is 83 below a threshold (for example, 0.6), and the region from which the object is extracted is appropriate.

  The circumscribed rectangle of one target object size calculated by the object size calculation unit 260 depends on the shooting conditions of the target object and the camera.

  Since the extension area is appropriately created near the target object by calculating the optimum size for the target object and the shooting conditions of the camera, there is an effect that the target object can be extracted efficiently and accurately.

  Also, the normalized correlation 80 is calculated by Equation 1.

FIG. 6 is a flowchart illustrating an example of a processing procedure of the extended area determination unit 400. In step 255, the label area of the change area is checked, and if it is equal to or greater than the threshold value, step 599 is performed as an object area from which most of the target object is extracted. If it is less than the threshold, step 294 sets the i-th object size region for the changed region of step 255, and step 290 performs normalized correlation to calculate the similarity. Step 291 checks the degree of similarity, and if it is greater than or equal to the threshold, step 293 is processed assuming that no object exists. In step 293, it is determined whether or not the extended area has ended. If not, the process returns to step 294. If the check in step 291 is less than the threshold value, it is assumed that an object exists, and step 270, step 292, and step 599 are processed.

  FIG. 7 is an explanatory diagram illustrating an example in which the object size calculation unit 260 calculates the size of one target object. First, it is assumed that the whole body of the target object 40 is shown in the shooting scene of the camera 10. Here, the lens focal length f122, the camera depression angle θ124, the camera height H123, and the camera coordinates (x, y) 121 that are parameters of the camera 10 are known. Then, the world coordinate origin (X, Y) can be calculated by Equation 2 using the world coordinate origin 125 immediately below the foot of the camera installation.

Now, the moving body in the whole body scene reflected in the camera can calculate the camera coordinates of the top position T (X, Yt, O) 129 and the foot position S (X, Ys, O) 128 of the person 126 from the extracted silhouette. . Further, the height h127 of the person in world coordinates can be calculated by Equation 3.
(Equation 3)
h = (1-Ys / Yt) H (Formula 3)
From this, it can be seen that the height of the moving object that has appeared in the camera 10 is h, and from the top position T (X, Yt, O) 129 of the extracted moving object to the foot position S (X, Ys, O). The size is the whole body silhouette in the camera 10.

  Here, the size of the detection target is based on a person (average adult) whose height is 160 cm, and the size of the whole body silhouette from the feet to the head, that is, the area (the number of pixels A) is calculated in advance. The determination is made based on the size. For example, when the area of the change region is about 5% to 40% of AA, it is less than the size of the detection control, and when it is about 40% or more, it is determined that most of the detection target is extracted.

  FIG. 8 is an explanatory diagram illustrating a luminance conversion example in which the sensitivity increase processing unit 500 increases the sensitivity. A luminance distribution (histogram) is calculated for the extended region 50c in the input image 30b, and a small luminance (dark luminance) 62 and a large luminance (bright luminance) 63 are calculated in the vicinity of the luminance distribution having the maximum luminance distribution of 61. With respect to the luminance 65 before conversion, the luminance 66 after conversion is converted (sensitivity increased) so as to increase the luminance 63 from the luminance 62.

  Here, the calculated low luminance (dark luminance) 62 and high luminance (bright luminance) 63 may be calculated using the average luminance of the extended region 50c.

FIG. 9 is a flowchart illustrating a processing procedure example of the object detection determination processing unit 600. The object extracted in step 309 is detected as the target object 40 in step 601. Step 602 calculates a circumscribed rectangle of the detected object, and step 603 calculates an optical flow within the circumscribed rectangle. Step 604 classifies the direction and size of the movement vector using the optical flow, and step 605 determines that the number of times that the direction of the movement vector is inconsistent and large (greater than or equal to the threshold value) reaches a predetermined value. Check if you did. If it is determined in step 605 that the predetermined value has been reached, step 606 determines that there is an abnormality, and step 607 stores abnormality information. If it is determined in step 605 that the value is less than the predetermined value, step 609 is determined to be normal. In step 608, it is determined whether or not to continue the process. If the process is continued, the process returns to step 309, and otherwise the process ends.

  Here, in the calculation of the optical flow in the circumscribed rectangle processed in step 603, the circumscribed rectangle is divided into blocks, and the similar pattern is found by searching where the pattern similar to the pattern in each block has moved. In this case, a standard method using the movement amount and direction (vector) as an optical flow may be used.

  FIG. 10 is an example showing the inside of the detection result notification unit 850. The display control unit 860 uses the image or information when an abnormality such as an assault on the target object 40 detected by the object detection determination processing unit 600 is used, and the display data generation unit 870 uses the camera identification information and the occurrence when the abnormality occurs. Generate date and time, error type, etc. On the other hand, the display image generation unit 880 generates an image before the occurrence of the abnormality in addition to the image at the time of occurrence of the abnormality and / or the image at the time of occurrence of the abnormality. The display image synthesizing unit 890 uses the data generated by the display data generating unit 870 and the image generated by the display image generating unit 880 to superimpose the data on the image and synthesize the resultant data to the display device 900 of the monitor television. Display in real time. Further, the image generated by the display image synthesis unit 890 may be stored in a time-lapse video and / or a VTR. Thereafter, looking at the display device 900 displayed in real time, the monitor 1000 can quickly process an instruction when an abnormality occurs. Moreover, you may monitor the image made into the time lapse video etc. off-line.

  FIG. 11 is a diagram illustrating an example of display contents of the display device 900. When an abnormality such as an assault of the target object 40 occurs, the display of the abnormality occurrence 920 and / or the image of the assault scene 940 of the target object 40 and / or the camera identification number and / or the date and time of occurrence 930 of the shooting scene that occurred Is displayed. On the other hand, at the time of occurrence of an abnormality, the progress until the occurrence is unknown only by the image at the moment of occurrence of the abnormality. In addition to the image 980 at the time of occurrence of abnormality ti + 4, the image 950 at time ti + 1 before the occurrence of abnormality and / or the time ti + 2 before occurrence of abnormality. The image 960 and / or the image 970 at time ti + 3 before the occurrence of abnormality are also displayed. Thereby, since the monitoring person 1000 can know the progress in addition to the occurrence of the abnormality of the passenger, there is an effect that it is possible to appropriately instruct the countermeasure person by using the console 1100 and / or the telephone.

  FIG. 12 is a diagram illustrating another example of the image processing apparatus 50. An image processing apparatus 50 that includes the illumination device 11 and monitors a target object 40 with a video signal 20 from a camera.

  The image processing apparatus 50 includes an image input unit 200, a change region extraction unit 300, an extended region determination unit 400, an illumination lighting control unit 540, an object extraction unit 550, an object detection determination processing unit 600, and a detection result notification unit 850.

In the processing in the image processing apparatus 50, first, the processing of the image input unit 200, the processing of the change region extraction unit 300, and the processing of the extended region determination unit 400 are performed. The illumination lighting control unit 540 controls turning on / off of the illumination device 11 using the LED so as to aim and irradiate the region determined by the extended region determination unit 400 and / or the vicinity of the region, and thereby make the target object 40 clear. Take a picture.

  These image input unit 200, change region extraction unit 300, extended region determination unit 400, sensitivity enhancement processing unit 500, object extraction unit 550, object detection determination processing unit 600, and detection result notification unit 850 are set at predetermined time intervals, for example, It is repeatedly executed in a short time such as one cycle of processing.

  FIG. 13 is a flowchart illustrating another processing procedure example of the image processing apparatus 50. In step 280, lighting of the illumination device 11 is controlled so as to aim and irradiate the extended region calculated in step 256 and / or the vicinity of the region. Step 294 calculates the average luminance of the extended area before irradiation and the average luminance after irradiation, and calculates this ratio (for example, R). When the illumination device 11 is appropriately turned on, the image input unit 200 inputs an image. Steps 253, 254, and 255 are performed. In step 253, the luminance of the background image is multiplied by R with respect to the input image captured when the lighting device 11 is turned on, and a luminance difference for each pixel from the simulated lighting image is obtained. Alternatively, a method may be used in which the average luminance of the background image is calculated in advance, the illumination of the background image is turned on, the average luminance is calculated, and this ratio is used.

  In step 255, since the target object 40 is clearly photographed, the label area of the change area becomes equal to or larger than the threshold value, and step 292 and step 599 are processed. Step 599 detects the presence / absence of an object, and if there is an object, determines whether the object is abnormal.

  FIG. 14 is a diagram showing another display content example of the display device 900. An image 980 at the time of occurrence ti + 4 and / or an image 950 at time ti + 1 before the occurrence of abnormality and / or an image 960 at time ti + 2 before the occurrence of abnormality and / or an image 970 at time ti + 3 before the occurrence of abnormality are obtained. In addition to the display, an input image before the irradiation for increasing sensitivity and / or photographing the target object 40 clearly is also displayed. For example, the image 981 before the increase in sensitivity of the image 980 and / or the image 951 before the increase in sensitivity of the image 950 and / or the image 951 before the irradiation and / or the image 961 before the increase in sensitivity of the image 960 and / or the image 970 before the irradiation. An image 971 before sensitivity increase and / or before irradiation is displayed on the display device 900.

  Thereby, the effect of sensitivity improvement and / or irradiation can be confirmed.

  Since only a person can be extracted with high accuracy, the present invention can also be applied to a monitoring system that tracks a person, a car, or the like in an environment where the background and the object are similar.

1 is a diagram illustrating an overall configuration of an image monitoring apparatus using the present invention. It is a flowchart which shows the example of a process sequence of an image processing apparatus. It is a flowchart which shows the other process sequence example of an image processing apparatus. It is a flowchart which shows the example of a process sequence which extracts the change area in a change area extraction part. It is explanatory drawing which shows the example of a determination procedure of an extended area | region determination part. It is a flowchart which shows the process sequence example of an extended area | region determination part. It is explanatory drawing in which an object size calculation part shows the example of size calculation of one target object. It is explanatory drawing which shows the brightness | luminance conversion example which a sensitivity improvement process part raises a sensitivity. It is a flowchart which shows the process sequence example of an object detection determination process part. It is an example which showed the inside of a detection result alerting | reporting part. It is a figure which shows the example of a display content of a display apparatus. It is a figure which shows the other Example of an image monitoring apparatus. It is a flowchart which shows the other process sequence example of an image processing apparatus. It is a figure which shows the other example of display content of a display apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Camera, 11 ... Illuminating device, 20 ... Video signal, 40 ... Target object, 50 ... Image processing apparatus, 200 ... Image input part, 300 ... Change area extraction part, 400 ... Expansion area determination part, 500 ... Sensitivity increase process , 540 ... Illumination lighting control part, 550 ... Object extraction part, 600 ... Object detection determination processing part, 850 ... Detection result notification part, 860 ... Display control part, 870 ... Display data generation part, 890 ... Display image composition part, 900 ... display device, 1000 ... monitor, 1100 ... console.

Claims (9)

In an image monitoring device that monitors images taken by a camera,
Change area extraction means for capturing an image captured by the camera and extracting a change area;
If the extracted change area is less than the size of the detection target, an extension area determination means for extending the area around the extracted change area;
Sensitivity correction processing means for correcting the brightness of the expanded area;
An image monitoring apparatus comprising: an object extracting unit that detects an object using an image whose sensitivity has been increased by the sensitivity correction processing unit. In an image monitoring device that monitors images taken by a camera,
Change area extraction means for capturing an image captured by the camera and extracting a change area;
If the extracted change area is less than the size of the detection target, an extension area determination means for extending the area around the extracted change area;
Illumination lighting control means for irradiating illumination with aiming at the expanded area;
An image monitoring system comprising: an object extraction unit that captures an image irradiated by the illumination lighting control unit with a camera, extracts a change region by the object extraction processing unit, and detects an object in the extracted change region apparatus. In the image monitoring apparatus according to claim 1 or 2,
The image monitoring apparatus according to claim 1, wherein the change area extraction unit uses an image in which one frame or more is accumulated. In the image monitoring apparatus according to claim 1 or 2,
The extended area determination means extends an area using the calculated size of an object to be detected. The image monitoring apparatus according to claim 1.
The image monitoring apparatus according to claim 1, wherein the sensitivity correction processing means performs brightness conversion using the brightness distribution of the area determined by the extended area determining means. In the image monitoring apparatus according to claim 1 or 2,
Object detection determination processing means for determining the presence or absence of abnormality of the object extracted by the object extraction means;
An image monitoring apparatus comprising: a detection result notifying unit that notifies a monitoring person when the object detection determination processing unit determines that an abnormality has occurred. The image monitoring device according to claim 6.
The image monitoring apparatus, wherein the detection result notifying unit displays an image whose sensitivity is increased by the sensitivity increasing processing unit. The image monitoring apparatus according to claim 2,
The illumination monitoring control unit uses an illumination device capable of controlling illumination lighting. The image monitoring device according to claim 6.
The image monitoring apparatus, wherein the detection result notifying means displays an image irradiated with the object illuminated by the illumination lighting control means.

Images