Image analysis – Applications – Vehicle or traffic control
Reexamination Certificate
2001-05-14
2004-10-05
Patel, Jayanti K. (Department: 2625)
Image analysis
Applications
Vehicle or traffic control
C340S910000, C340S917000, C340S918000, C701S117000, C701S118000
Reexamination Certificate
active
06801638
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention concerns a device and a method for recognizing traffic signs.
Traffic on streets, railways, navigable waterways and airfields is regulated-by traffic signs. The traffic regulations referenced by the traffic signs frequently have a geographic limitation which extends over segments of the traffic pattern, after which the vehicle operator need no longer consider the traffic sign. Such segments are marked, for example, by a traffic sign which indicates the beginning of the applicability of the regulation, and a further traffic sign which indicates the lifting of the applicability of the regulation. For the vehicle operator it would be very helpful to have a memory aid, which would provide him with information regarding the content of the traffic sign (to which obedience is being required), and particularly those for which a long period of time has lapsed since the sign was passed.
DESCRIPTION OF THE RELATED ART
In the previous state of the art devices and methods have been proposed, which have the objective of completely extracting the information content of the traffic sign.
For example, in patent document DE 36 19 824 C2 a device is proposed for aiding the memory of a vehicle operator, which detects the speed limit by recognition of the traffic sign and indicates this within the cabin of the vehicle for the duration of its applicability. This device includes an image sensor, an information-processing unit, a memory unit and an indicator unit. By means of the image sensor images are detected or recorded from the traffic lane ahead of the vehicle. By means of the information-processing unit these images are examined in real time for traffic signs and upon recognition of a traffic sign the content thereof, in this example a speed limit, is extracted. The extracted value of the highest permitted speed is stored in the memory unit and indicated on the indicator unit for the duration of validity, that is, until recognition of a traffic sign which removes the restriction.
The disadvantage of this device is comprised in the high error rate for non-recognized traffic signals and its high cost. The actual recognition of a particular traffic sign, that is, the extraction of its complete information content, with a sufficiently high confidence requires high quality color images and a highest possible resolution (Janssen et al., “Hybrid Approach of Traffic Sign Recognition”, Proc. Intelligent Vehicle Symposium, Tokyo, Japan, pp. 390-395, 1993). The evaluation of such images in real time requires a very high computer power. Image sensors and information processing units having such high power are expensive. And even with these the error quotient for various traffic sign recognition systems is in the magnitude of several percent. Such error quotients cannot be tolerated for warranty and/or insurance reasons.
In order to minimize the workload on the computers, it is proposed in the document DE 197 36 774 A1, that after the detection of a traffic sign within a camera segment, this is extracted therefrom and indicated. The indication or display occurs without previously evaluation of the content of the traffic sign and without prior classification. The disadvantage of this process lies in the fact that no workup or processing of the traffic sign image occurs. This has the consequence, that in poor visibility conditions the observer requires increased care and attention to be able to recognize and understand the traffic signs displayed to him.
The task of the present invention is thus comprised on the one hand of finding a process and a device which make it possible to reliably, with as low as possible computer load and thus economically, to recognize traffic signs and to provide a display to an observer as a reminder or aid. Herein the error rate in the framework of the recognition of the object identified as a traffic sign should be kept as low as possible.
SUMMARY OF THE INVENTION
The task is solved by the process or, as the case may be, the appropriately designed device, in which the image data from the image sensor is analyzed and classified in an information processing unit, and on the basis of the results is recorded in memory as a class-specific image segment and/or a symbolic figure representing an image stored a memory unit, which is then displayed by means of the display unit. Therein, in the framework of the analysis, the information-processing unit it is determined whether the image data of the image sensor contains one or more objects, which, with sufficiently high probability, are traffic signs. Consequently, the so detected objects are forwarded to a further processing and classification step. In a particularly advantageous manner, in the framework of the invention, during the further processing of the image data of the objects, a separation is undertaken into traffic sign specific upper classes and lower classes. Thereby, appropriate classification specific characteristic data are extracted and relayed to a separate classification. Depending on whether the classification specific characteristic data are correctly recognized with high probability during the classification, they are replaced by the corresponding or appropriate symbolic image representing image data stored in the storage unit or by the appropriate original image data stemming from the image sensor. The image data resulting from the dissections are recombined into a synthetic image of a traffic sign and this image is then stored in the memory unit and brought to; the display by means of the display unit.
An important advantage of the invention is comprised therein, that it is no longer necessary to carry out the computer taxing and with high error quotient associated problem of classification of a relatively complex traffic sign which must be solved over its entirety, rather the problem, by separation of the detected objects into traffic sign specific upper classes and lower classes in a preceding classification process, is dissected or redirected to clearer and therewith simpler classification problems. Therewith in like measure the demands on the necessary computing power and the image quality are reduced.
REFERENCES:
patent: 5420794 (1995-05-01), James
patent: 6249375 (2001-06-01), Silhengst et al.
patent: 6285393 (2001-09-01), Shimoura et al.
patent: 6453056 (2002-09-01), Laumeyer et al.
patent: 6459387 (2002-10-01), Kobayashi et al.
patent: 6556692 (2003-04-01), Gavrila
Piccioli G, De Micheli E Parodi P Campani M: “Robust method for road sign detection and recognition”. Image and Vision Computing 14 (1996), pp. 209-223.
“Das Computer-Auge”. Münchner Merkur No. 254, Nov. 4, 1998, p. A1 (in German).
“Computer erkennt Verkehrszeichen”. RFE 2, 1997, p. 7 (in German).
Janssen Reinhard
Lindner Frank
Ulmer Berthold
Daimler-Chrysler AG
Patel Jayanti K.
Pendorf & Cutliff
Tabatabai Abolfazl
LandOfFree
Device and method for recognizing traffic signs does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Device and method for recognizing traffic signs, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Device and method for recognizing traffic signs will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3297880