Television – Special applications – Observation of or from a specific location
Reexamination Certificate
2000-12-22
2004-11-16
Kelley, Chris (Department: 2613)
Television
Special applications
Observation of or from a specific location
Reexamination Certificate
active
06819353
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a method for monitoring a monitored area, comprising the steps of recording an image of the monitored area and of comparing the recorded image with a reference image in order to determine if there is an alarm situation. The invention further relates to a monitoring unit and a computer program.
TECHNICAL BACKGROUND
The use of cameras for various types of monitoring, both in industry and in private dwellings, is already known. The aim of the monitoring is often to prevent burglary and theft.
The camera takes pictures of a monitored area. The monitored area is limited by the optics of the camera. The recorded images are analyzed for an evaluation of their content to be made. If the camera monitors, for example, premises in which people are not permitted, an analysis is made of the recorded image and an evaluation made whether there is an extraneous object within the monitored area or not.
In a monitoring context the term scene is used. The scene changes, for example, is something moves physically within the monitored area, for example if a person enters the monitored area. On the other hand, the scene does not change if the light in the monitored area changes. If a change of scene is detected, it can be determined that there is an alarm situation.
The analysis of the recorded images can be carried out manually, by a person looking at a recorded image and making a visual evaluation about whether there is an intruder in the monitored area or not. If an intruder is discovered on a recorded image, the monitoring personnel who carries out the evaluation can, for example, telephone the police for further measures to be taken. This is an expensive form of monitoring, as it requires a person to go through the recorded images. Even if the person who carries out the evaluation of the recorded images can evaluate and monitor several monitored areas at the same time, it is time-consuming and very monotonous and stressful work, as an “event” that is not permitted must not be missed in the recorded images. In addition to an intruder entering a building, an event that is not permitted can be an incorrectly manufactured bottle in a production line.
The analysis of the recorded images is therefore carried out to an increasing extent by means of digital image processing. In order to discover, for example, an intruder in a recorded image, the recorded image is compared with a reference image, which can be a background image that was recorded on a previous occasion when there was no intruder in the monitored area. The reference image can also be the immediately preceding recorded image. If there is a difference between the two compared images this can mean that there is an intruder in the monitored area. The difference can also be due to changes in the light in the monitored area, which for instance can arise when the sun goes behind a cloud, when it starts to get dark or when a light is switched on. One way of reducing the problem caused by changes in the light is to detect which changes concern changes in the light and which concern changes in the scene. This is solved by using different criteria for which changes in intensity are the result of changes in the light and which are the result of changes in the scene. This is a difficult problem to solve. The algorithms that are used are very complicated and take a long time to process.
Another problem of the prior-art technique is if changes in the light give rise to changes in intensity that do not follow the said criteria for changes in intensity. This occurs, for example, when there are objects in the scene that cast shadows. In this case, the changes in the light will be one way in the shadow and another way alongside.
SUMMARY OF THE INVENTION
An object of the present invention is therefore to provide a method that avoids the above-mentioned problems.
This object and other objects which will be apparent from the following description are achieved, according to a first aspect of the present invention, by a method, which is of the type described by way of introduction and is characterized by the step of creating the reference image by combining at least two previously stored images of a set of images, so that the lighting conditions in the reference image is adjusted to the lighting conditions in the recorded image.
Instead of first comparing the recorded image and the reference image to determine whether the image has changed and then deciding whether the change is an alarm situation or a change in the light, the changes in the light are now considered before the comparison is made. This is carried out by having previously stored images which can be combined and in this way can recreate the lighting conditions in the reference image. A new reference image can be created before each comparison with a recorded image. It is also possible to use the preceding reference image when the next image of the monitored area is recorded and is to be compared with a reference image. For example, a new reference image can be recreated every couple of minutes, or every fifth, tenth or hundredth time an image is recorded and is to be compared with a reference image. An advantage of the present invention is that an optimal reference image is obtained in a simple and quick way, which reference image takes the lighting conditions into account. This means that the step of comparing the recorded image and the reference image becomes simpler and requires less complex algorithms than must be used, for example, for the classification of areas according to changes in the light and changes in the scene. This also has the advantage that less memory is required. This is particularly important when there is limited processing capacity, which is the case in a system with distributed intelligence. Distributed intelligence means that an analysis of the recorded image is carried out in a monitoring unit at the monitored area to determine whether there is an alarm situation.
The result of the comparison is also more reliable as the risk of a false detection is reduced when a reference image according to the invention is used.
The monitoring can, for example, be carried out in premises where valuable documents are stored and in which people are not permitted, and aims at detecting whether a person is in the room or not. In order to be able to detect a person, an image is recorded and the area which the image covers is called the monitored area. The recorded image is compared with a reference image in order to discover whether a person has entered the monitored area. The reference image is an image that represents the background of the monitored area at the time of recording the image. The background of the monitored area can be declared as the monitored area when no object that can cause an alarm is in the monitored area, that is to say when the monitored area is undisturbed. The reference image is created by combining images which are stored in a set of images, so that the reference image is adjusted to the prevalent lighting conditions in the recorded image. The image information in the reference image is adjusted to the image information in the recorded image. The image information comprises the light intensity in the pixels of the image. The images in the set of images do not need to correspond individually to an actual light level for the scene. What is important is that it is possible to create a reference image for as many different types of light level as possible from a limited number of images. It can be possible to update the set of images when the lighting conditions in the recorded image cannot be recreated in the reference image from the set of images.
By combining images in a set of images, it is also possible to take into account shadows from fixed objects, which shadows are included in the reference image.
Another advantage is that the monitoring system is more robust and less sensitive to interference when the reference image is adjusted immediately to the prevalent lighting conditions and does not detect these sepa
Birch & Stewart Kolasch & Birch, LLP
Bugg George A
Kelley Chris
Wespot AB
LandOfFree
Multiple backgrounds does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multiple backgrounds, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multiple backgrounds will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3315867