JPH0552933A - System and device for measuring position of traveling object - Google Patents

Abstract

PURPOSE:To more accurately and more reliably measure the position of a traveling object. CONSTITUTION:The picture (e) of a traveling object is extracted as the pre- process of correspondence retrieval. An original picture (c) obtained by image sensing section (a) and (b) contains the picture (d) of a still object, etc., and the picture (e) of the traveling object is extracted by finding the difference between the original picture (c) and such a comparison picture as a past picture, background picture, etc. The extracted picture (e) of the traveling object is used for the correspondence retrieval. Since the correspondence retrieval can be performed only on the picture (e) of the traveling object, the position measuring speed of the traveling object can be increased. When three or more image sensing sections are used by performing calibration as the sections (a) and (b) and a lost part filling up process is performed on the lost part of the picture (e) or the lost part is supplied from the original picture, false matching can be prevented and the reliability of the position measurement can be improved.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving object position measuring system and apparatus for measuring the distance of a moving object based on image information.

[0002]

2. Description of the Related Art Moving image processing is a technique for capturing a moving object with a camera or the like and obtaining information about the moving object from image information obtained in time series. For example, in the field of vehicle monitoring in a parking lot, particularly in the field of vehicle movement monitoring, such technology can be effectively used for automation of monitoring work or for the convenience of users.

In moving image processing, a method of measuring the distance of a moving object is also known. For example, a plurality of (for example, two) image pickup means such as cameras are arranged at intervals so that their visual fields overlap. When the image information obtained by the plurality of image pickup means is viewed stereoscopically, it is possible to measure the distance to the object included in the overlapping visual fields. Distance measurement is
The above example is effective for knowing the position of the vehicle existing in the parking lot.

[0004]

Conventionally, in order to speed up the distance measurement, a correspondence search method for collating images obtained by a plurality of image pickup means and a method for dividing an image area have been used. It was being done. However, the conventional distance measurement method is inefficient when it is desired to know the distance of only a moving object. This is because all the objects included in the overlapping visual fields of the image pickup means must be subjected to the distance measurement processing. Therefore, it is necessary to perform arithmetic processing even for an originally unnecessary portion, which speeds up arithmetic operation,
As a result, it hinders quick distance measurement of a moving object.

The present invention has been made to solve the above problems, and it is possible to perform correspondence search only on a moving object and to measure one moving object at a higher speed. The purpose is to

[0006]

In order to achieve such an object, the position measuring method for a moving object according to the present invention provides a method of measuring a still object image with respect to a plurality of images obtained by a plurality of image pickup units arranged at intervals. It is characterized in that an image showing a moving object is extracted by obtaining a difference from the included comparison image, and only the extracted image of the moving object is subjected to corresponding search to measure the position of the moving object.

Further, the moving object position measuring method according to a second aspect is characterized in that the comparison image is an image previously obtained by the image pickup unit or a known background image.

Further, the moving object position measuring method according to any one of claims 3 to 5 has means for preventing false matching, and in particular, claim 3 moves while testing three or more imaging units. The object is imaged, the claim 4 is that the correspondence search is performed after performing a filling process for compensating for the missing portion caused by the difference in the extracted image of the moving object, and the claim 5 is the case of the extracted moving object. A feature of the present invention is that the correspondence search is performed after the missing portion caused by the difference is supplemented by the original image obtained by the imaging unit.

The moving object position measuring apparatus according to the present invention obtains a difference between a plurality of image pickup units arranged at intervals and a plurality of images obtained by the image pickup unit and a comparison image including a still object image. And a matching unit that finds the position of the moving object by correspondingly searching the images of the moving object extracted by the difference processing unit.

[0010]

First, the principle of the position measuring method of the present invention will be described with reference to FIG. In the present invention, first, the difference between the image (original image) c obtained by the imaging units a and b and the comparison image is obtained. It is assumed that the original image c includes an image d of a stationary object and an image e of a moving object as shown in FIG. Therefore, the calculated difference includes the image e of the moving object. A plurality of differences correspond to the plurality of image pickup units a and b,
That is, the difference between the comparison image and the image obtained by the image pickup unit a belonging to the time series f, and the difference between the image obtained by the image pickup unit b belonging to the time series g and the comparison image are obtained. Further, the image e of the moving object obtained in this way is subjected to a correspondence search with a stereo pair indicated by h1, h2, ...
The position of the moving object is determined. That is, only the image e of the moving object is the target of the correspondence search, and the position of the moving object can be measured at high speed.

Further, in claim 2, the comparison image is
It is an image obtained by the imaging units a and b in the past or a known background image. Each of these images is an image including the image d of the stationary object, and the above-described operation can be obtained.

Further, in claims 3 to 5, false matching is prevented. The false matching means that the image information necessary for the correspondence search is missing, and in the case of claim 1, it may occur due to the difference. According to the third aspect, such false matching is prevented by using three or more image capturing units as the image capturing units a and b while examining the image capturing units and capturing a moving object. According to the fourth aspect, the missing portion caused by the difference is compensated by the hole filling processing such as the expansion processing and the false matching is prevented. In the fifth aspect, the missing portion caused by the difference is compensated by the original image c obtained by the image pickup units a and b, and false matching is prevented.

In the position measuring device of the present invention, the position measuring method described above is executed. That is, the image c captures an image c related to stereoscopic vision. next,
The difference processing unit obtains the difference between the image c obtained by the imaging unit and the comparison image, and extracts the image e of the moving object. The image e of the moving object extracted by the difference processing unit is the target of the correspondence search, and the matching unit
The position of the moving object is determined.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings.

FIG. 2 shows the configuration of a moving object position measuring apparatus according to an embodiment of the present invention applied to a vehicle monitoring system for a parking lot. The device shown in this figure includes an input unit 10, a control unit 20, and an output unit 30.

The input unit 10 includes cameras 10A and 10B. The cameras 10A and 10B are arranged so as to stereoscopically view the object. In this embodiment, three or more cameras are tested to prevent false matching. The control unit 20 also controls the difference processing units 22A and 22A.
B, a matching section 24, and an overall processing section 26. The difference processing units 22A and 22B take in the image information obtained by the imaging of the corresponding cameras 10A and 10B, respectively, and perform the processing described below to extract the image of the moving object. The matching unit 24 searches for correspondence between the images extracted by the difference processing units 22A and 22B,
Find the distance of a moving object. The overall processing unit 26 obtains the position of the moving object based on the obtained distance, or controls the output unit 30. The output unit 30 is configured to operate under the control of the overall processing unit and has a display unit 32 and a voice synthesis unit 3.
4, the gate opening / closing part 36 is provided. The display unit 32 displays a position of a moving object (a vehicle in a parking lot monitoring system), a message to the moving object, and the like. The voice synthesizing unit 34 synthesizes a voice message instead of the display or at the same time as the display, and causes the loudspeaker 38 to output the voice. The gate opening / closing unit 36 drives a mechanism necessary for moving a moving object (opens / closes a gate for loading / unloading a vehicle in a parking lot monitoring system).

The feature of this embodiment is that the moving object is extracted as a pre-process before the stereo image is searched for the correspondence. 3 and 4 conceptually illustrate the contents of this preprocessing.

First, the example in FIG. 3 is an example in which the comparison image is a past image and the optimum threshold method is used. In this example, the difference between the original image obtained by the camera 10A and the past image (the image previously obtained by the camera 10A) is obtained. The obtained difference is selected by the threshold value determined by the optimum threshold method, and as a result, only the image that truly shows the moving object is extracted. In addition, in this extraction, an image expansion process is performed to fill in the missing pixels.
This process is performed by the difference processing unit 2 to prevent false matching.
2A.

At the same time, the same processing is performed by the difference processing section 22B on the original image obtained by the camera 10B. Since the cameras 10A and 10B are arranged in stereo, the position of the moving object can be determined by performing correspondence search (matching) using the obtained two types of moving object images. This process is executed by the matching unit 24 in accordance with, for example, the face-area method. When matching,
The missing part of the moving object image is extracted from each original image to prevent false matching. Next, the example of FIG. 4 is an example in which the comparison image is the background image, unlike FIG. The background image is known, and the difference processing units 22A and 2A
2B has this as information. Since the comparative image is the background image, it is not necessary to perform expansion processing.
The extracted moving object image is subjected to correspondence search by, for example, a differential method.

In the above description, the parking lot application is taken as an example, but other applications may be used. Further, the threshold value determining method is not limited to the optimum threshold value method, and may be the fixed value method or the floating threshold value method. Further, as the correspondence search method, a surface correspondence method, a rough special differential correspondence method, a dense special differential correspondence method, a moving point programming method, a differential method, etc.
Either can be adopted.

[0021]

As described above, according to the present invention,
It is only necessary to perform the correspondence search only on the image of the moving object, the position of the moving object can be measured at high speed, and a position measuring device useful for various applications such as a parking lot can be realized.

Further, according to the present invention, false matching is prevented and the reliability of the position measurement of the moving object is improved by using the test of the image pickup unit, the filling processing, and the extraction of the missing portion from the original image.

[Brief description of drawings]

FIG. 1 is a diagram showing an operation principle of the present invention.

FIG. 2 is a diagram showing a device configuration when a position measuring device according to an embodiment of the present invention is applied to a vehicle monitoring system for a parking lot.

FIG. 3 is a diagram showing an example of a moving object extraction process in this embodiment.

FIG. 4 is a diagram showing another example of a moving object extraction process in this embodiment.

[Explanation of symbols]

 a, b Imaging unit c Original image d Image of stationary object e Image of moving object f, g Time series h1, h2 Stereo pair 10 Input unit 10A, 10B Camera 20 Control unit 22A, 22B Difference processing unit 24 Matching unit 30 Output unit

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Katsuyuki Taniyama 5-1-1 Shimorenjaku, Mitaka City, Tokyo Japan Radio Co., Ltd.

Claims (6)

[Claims] 1. An image of a moving object is extracted by calculating a difference between a plurality of images obtained by a plurality of image pickup units arranged at intervals and a comparison image including an image of a still object, and the extracted image of the moving object. A position measuring method of a moving object, characterized in that only the corresponding search is performed to measure the position of the moving object. 2. The moving object position measuring method according to claim 1, wherein the comparison image is an image previously acquired by the image capturing unit or a known background image. .. 3. The position measuring method for a moving object according to claim 1, wherein the moving object is imaged while testing three or more imaging units. 4. The moving object position measuring method according to claim 1, wherein a correspondence search is performed on the extracted image of the moving object after performing a filling process for compensating for a missing portion caused by the difference. Position measurement method for objects. 5. The moving object position measuring method according to claim 1, wherein a correspondence search is performed after the missing portion caused by the difference is supplemented by the original image obtained by the image pickup unit in the extracted image of the moving object. Position measurement method for moving objects. 6. A plurality of image pickup units arranged at intervals, and a difference processing unit which extracts a moving object image by obtaining a difference between a plurality of images obtained by the image pickup unit and a comparison image including a still object image. A moving object position measuring device, comprising: a matching unit that searches for an image of the moving object extracted by the difference processing unit and finds the position of the moving object.