Image analysis – Applications – Vehicle or traffic control
Reexamination Certificate
2000-05-04
2002-08-06
Chang, Jon (Department: 2623)
Image analysis
Applications
Vehicle or traffic control
C382S106000, C340S937000
Reexamination Certificate
active
06430303
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image processing apparatus for accurately extracting one or a plurality of objects utilizing a thresholded differential image processing technique when a plurality of stationary objects and a plurality of moving objects are contained together in an image of a time sequence of images.
More specifically, the present invention relates to an image processing apparatus, which allows both a background image and an image including at least one stationary object or at least one moving object (having a speed not more than a predetermined speed) to be extracted, and also allows a difference-calculation process to be carried out between the images.
In such an image processing system, it is possible to distinguish a stationary object or a moving object in the images and it is also possible to analyze the movement of each of the moving objects.
Further, the present invention relates to an image processing apparatus, in which a plurality of markers are provided in a background where the objects move, and these markers are extracted by utilizing an image processing technique similar to the case in which the moving objects are extracted, and further it is discriminated whether or not the thus extracted markers are in the steady state.
If the markers are in the steady state, portions where the moving objects and the markers overlap each other can be determined. Therefore, the number of the markers (the size of markers displayed in each image), which are in the steady state and exist between two moving objects, can be calculated to obtain a distance between two moving objects.
In general, supervisory systems using the above-mentioned image processing technique can be utilized in various places. Each of these supervisory systems serves to rapidly locate an accident, a disaster, and the like. Recently, such supervisory systems are likely to be utilized for preventing such accidents, disasters, and the like, in addition to a function of merely detecting the existence of an accident, etc.
To meet this need, it is necessary to extract or identify an object which moves with an abnormal motion that will cause such an accident, a disaster, and the like. Therefore, an efficient technique is needed for rapidly and accurately detecting a moving object which demonstrates such an abnormal motion.
More specifically, it is required for the supervisory system to detect and analyze the movement of each of a plurality of moving objects contained in a series of images. Further, it is also necessary for the supervisory system to rapidly calculate a distance between the two moving objects with a high degree of accuracy.
2. Description of the Related Art
Some techniques for analyzing the movement of each of a plurality of moving objects by utilizing an image processing apparatus are typically disclosed in Japanese Unexamined Patent Publication (Kokai) No. 5-159057 and No. 5-159058.
In each of these techniques, first, regions where moving objects may be positioned are extracted using a predetermined assumption. Next, a specified moving object is distinguished from the other objects, on the basis of various characteristics, e.g., the size of each of the regions, and the central position of each region. Subsequently, in accordance with a change of the position of the moving object with a lapse of time, the movement of the moving object can be analyzed.
For example, when an analysis of the motion of a man is to be performed, a given portion of an image which is to be analyzed, is extracted from the image. Next, with respect to the extracted portion, i.e., an object to be processed, various characteristics, e.g., the position of projections and the location of central positions, are calculated, and used to distinguish the object from the other portions. Further, the process is executed with respect to a plurality of images in a time series, i.e., continuous motion type images.
According to the above-mentioned technique, to ensure obtaining adequate attributes, e.g., a speed of the object, it is necessary to analyze all the areas where the same original object can exist in the time series, and to identify the objects as the same original object.
More specifically, if a plurality of objects respectively existing in a plurality of the time series images are not accurately correlated with each other, by analyzing all the areas where the object can exist, it is difficult to calculate the speed of the object with a sufficiently high accuracy.
In the case where only one moving object exists, a process for correlating a plurality of objects in the continuous images with each other is relatively simple. In this case, it is possible to easily obtain the attributes, e.g., a speed of the moving object, using changes in the time base.
However, especially in the case where a large number of moving objects exist in one image, a process for correlating a plurality of objects in the time series images with each other for all the moving objects becomes difficult.
Further, when a plurality of stationary objects exist, as well as a plurality of moving objects, it becomes extremely difficult to rapidly complete such a correlation process for all of the stationary and moving objects using real time processing with a frame rate processing determined by a frequency of a video signal.
Furthermore, when a plurality of moving objects respectively move with a speed different from each other, it becomes almost impossible to complete the correlation process for all of the moving objects using real time processing determined by the frequency of a video signal (a video frame rate).
SUMMARY OF THE INVENTION
In view of the above-described problems existing in the prior art, the main object of the present invention is to provide an image processing apparatus which allows one or a plurality of objects to be rapidly and accurately extracted and analyzed, in a case where a large number of moving objects exist in a time series image.
A further object of the present invention is to provide an image processing apparatus which allows one or a plurality of objects to be rapidly and accurately extracted and analyzed, even in the case where a plurality of stationary objects exist, as well as a plurality of moving objects.
A still further object of the present invention is to provide an image processing apparatus which allows the movement of each of a plurality of moving objects to be rapidly and accurately extracted and analyzed, even in the case where the plurality of moving objects respectively move with different speeds.
A still further object of the present invention is to provide an image processing apparatus which allows all of the stationary and moving objects to be correlated with each other during real time processing with a processing rate determined by a frequency of a video signal, in the case where a plurality of stationary objects exist, as well as a plurality of moving objects, and also in a case where the plurality of moving objects respectively move with different speeds.
A still further object is to provide an image processing apparatus which allows a distance between two moving objects to be calculated, so that an abnormal object motion that may bring about an accident, a disaster, and the like, can be rapidly detected.
To attain these objects, the image processing apparatus according to the present invention includes an image-input unit which inputs an image including a background and a plurality of the objects; a background image extract unit which extracts the background; a first average background extract unit which extracts an image that includes one or a plurality of stationary objects or moving objects each having a speed not higher than a predetermined first speed and also includes the background; and a second average background extract unit which extracts an image that includes the stationary objects or moving objects each having a speed not higher than a predetermined second speed and also includes the background.
Further, the image processing app
Egawa Hiroichi
Naoi Satoshi
Shiohara Morito
Chang Jon
Fujitsu Limited
Staas & Halsey , LLP
LandOfFree
Image processing apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2893618