Method and device for processing recorded image information from a vehicle
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
H04N-007/18
G06K-009/00
B60S-001/08
출원번호
US-0989921
(2008-11-19)
등록번호
US-8854455
(2014-10-07)
우선권정보
DE-10-2008-001-679 (2008-05-09)
국제출원번호
PCT/EP2008/065795
(2008-11-19)
§371/§102 date
20101027
(20101027)
국제공개번호
WO2009/135541
(2009-11-12)
발명자
/ 주소
Haug, Karsten
출원인 / 주소
Robert Bosch GmbH
대리인 / 주소
Kenyon & Kenyon LLP
인용정보
피인용 횟수 :
5인용 특허 :
2
초록▼
A method for improving the driver assistance function, in particular of driver assistance systems based on video images recorded from a vehicle, and a corresponding device for that purpose made up of a camera and processing unit. To improve the function during rain, the passing of a windshield wiper
A method for improving the driver assistance function, in particular of driver assistance systems based on video images recorded from a vehicle, and a corresponding device for that purpose made up of a camera and processing unit. To improve the function during rain, the passing of a windshield wiper through the camera image be used to classify individual images and/or portions of images as being of higher or lower quality, in order to improve the quality of the images from the camera. The images from camera are intended to be used for automatic driver assistance systems.
대표청구항▼
1. A method for processing image information recorded from a vehicle, the method comprising: detecting at least parts of a windshield wiper in the image information;using results of the detecting to identify portions of a recorded image that have recently been swept by the windshield wiper; anddivid
1. A method for processing image information recorded from a vehicle, the method comprising: detecting at least parts of a windshield wiper in the image information;using results of the detecting to identify portions of a recorded image that have recently been swept by the windshield wiper; anddividing the image information into higher-quality and lower-quality image information, wherein the image information is divided into image information whose quality is higher than a predefined threshold value and image information whose quality is lower than the predefined threshold value, wherein the dividing of the image information into higher-quality and lower-quality image information takes into account at least the presence of a moisture on the windshield, and wherein image information corresponding to image portions that were identified as recently swept are assigned a higher quality compared to image information corresponding to image portions that have not recently been swept. 2. The method of claim 1, wherein a phase counter is synchronized with the time periods between two consecutive detections of a windshield wiper, and wherein the image information that is recorded immediately after a first recorded image is swept by the windshield wiper is given the highest quality ranking with the aid of the phase counter and the image information that is recorded immediately before a second recorded image is swept by the windshield wiper detector is given the lowest quality ranking. 3. The method of claim 2, wherein: the image information of an individual image is spatially divided by a detected windshield wiper shadow into higher-quality image information and lower-quality image information; andby alternatingly determining the quality of the image information as a function of the detection of a windshield wiper, the spatial division is carried out when the detection of the windshield wiper in the image is positive and the quality ranking is carried out with the aid of the synchronized phase counter when the detection of a windshield wiper in the image is negative. 4. The method of claim 2, wherein the image information is shifted and then ranked as higher quality which is recorded at a later time than immediately after being swept by the windshield wiper, if data are present indicating that a windshield fan is switched on. 5. The method of claim 2, further comprising: performing a software-implemented PLL synchronization in which the detection of the windshield wiper in the image is compared to the progression of an image counter, wherein a phase difference is found if the progression of the image counter does not coincide in time with the detection of the windshield wiper, and in this case the progression of the image counter being varied by varying the number of images to be counted until the detection of the windshield wiper in the image coincides in time with the progression of the image counter. 6. The method of claim 2, further comprising: performing a synchronization of the quality ranking with angle sensor data for the windshield wiper. 7. The method of claim 2, further comprising: performing a synchronization of the exposure parameters with the detection of the windshield wiper. 8. The method of claim 1, wherein the image information of an individual image is spatially divided by a detected windshield wiper shadow into higher-quality image information and lower-quality image information. 9. The method of claim 1, wherein a windshield wiper is detected by determining a noise level of individual image portions, and wherein the image portions that have a high noise level are classified as of lower quality and the image portions that have a low noise level are classified as of higher quality. 10. The method of claim 1, wherein a windshield wiper is detected by dividing the image information into individual image portions, wherein a brightness histogram is subsequently determined for each individual image portion, and wherein the image portions that have low exposure in comparison to the other image portions are evaluated as being covered by the windshield wiper. 11. The method of claim 1, further comprising: performing an additional plausibility check of the quality ranking for different image areas, wherein at least one of the location of the image areas, the time of the last positive windshield wiper detection, and the time from the synchronization parameters are used for the plausibility check. 12. The method of claim 1, further comprising: applying the image information as input to an analysis algorithm; andapplying a weighting factor to at least one of: individual images prior to the input of the image information into the analysis algorithm,image portions prior to the input of the image information into the analysis algorithm, anda result of the analysis algorithm;wherein the weighting factor varies as a function of image quality. 13. A device for processing image information recorded from a vehicle, comprising: at least one camera; andat least one processing unit for subdividing the image information from the camera into individual image portions and performing a statistical calculation for each individual image unit;wherein the processing unit uses the data of the statistical calculation to determine: whether at least part of the windshield wiper is present in the particular image unit,where at least part of the windshield is present, portions of a recorded image that have recently been swept by the windshield wiper, andwhether the quality of the image information is high or low, in particular whether the quality of the image information is higher or lower than a predefined threshold value,wherein the dividing of the image information into higher-quality and lower-quality image information takes into account at least the presence of a moisture on the windshield, andwherein image information corresponding to image portions that were identified as recently swept are assigned a higher quality compared to image information corresponding to image portions that have not recently been swept. 14. The device of clam 13, wherein the processing unit uses time data additionally for the statistical calculation and uses those time data in the form of a synchronization, which is a PLL synchronization, to improve the detection of the windshield wiper in individual image portions through anticipation. 15. The device of claim 13, wherein the processing unit applies the image information as input to an analysis algorithm, the processing unit applying a weighting factor to at least one of: individual images prior to the input of the image information into the analysis algorithm,image portions prior to the input of the image information into the analysis algorithm, anda result of the analysis algorithm;wherein the weighting factor varies as a function of image quality.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (2)
Kenji Mori JP, Oscillating circuit for producing an output signal synchronous with an input signal.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.