Lens-attached matter detector for determining presence or absence of attached matter, lens-attached matter detection method executed by the same, and vehicle system having the same

• US 9,224,051 B2
• Filed: 07/01/2013
• Issued: 12/29/2015
• Est. Priority Date: 07/03/2012
• Status: Active Grant

• Current Assignee: Clarion Co., Ltd. (Faurecia SA)
• Original Assignee: Clarion Co., Ltd. (Faurecia SA)
• Inventors: Irie, Kota, Muramatsu, Shoji, Kato, Kenji, Kiyohara, Masahiro
• Primary Examiner(s): Johns, Andrew W
• Application Number:
  US13/932,787
• Publication Number:
  US 20140010408A1
• Time in Patent Office: 911 Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/30252 Vehicle exterior; Vicinity of vehicle
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06V 20/56 exterior to a vehicle by using sensors mounted on the vehicle

Abstract

A lens-attached matter detector includes an edge extractor configured to create an edge image based on an input image, divide the edge image into a plurality of areas including a plurality of pixels, and extract an area whose edge intensity is a threshold range as an attention area, a brightness distribution extractor configured to obtain a brightness value of the attention area and a brightness value of a circumference area, a brightness change extractor configured to obtain the brightness value of the attention area and the brightness value of the circumference area for a predetermined time interval, and obtain a time series variation in the brightness value of the attention area based on the brightness value of the attention area, and an attached matter determiner configured to determine the presence or absence of attached matter based on the time series variation in the brightness value of the attention area.

Claims

1. A lens-attached matter detector, comprising:

an edge extractor that creates an edge image based on an input image from an imager having a lens, divides the edge image into a plurality of areas including a plurality of pixels, and extracts an area where edge intensity is within a threshold range as an attention area;

a brightness distribution extractor that obtains a brightness value of the attention area and a brightness value of a circumference area;

a brightness change extractor that obtains the brightness value of the attention area and the brightness value of the circumference area obtained by the brightness distribution extractor for a predetermined time interval, and obtains a time series variation in the brightness value of the attention area based on the brightness value of the attention area for the predetermined time interval; and

an attached matter determiner that determines the presence or absence of attached matter based on the time series variation in the brightness value of the attention area.

2. The lens-attached matter detector according to claim 1, further comprising a mask processor that eliminates from the attention area the area except lens-attached matter in the area where edge intensity is within the threshold range extracted by the edge extractor.

3. The lens-attached matter detector according to claim 2, wherein the mask processor eliminates from the attention area an area including an edge extending in a direction toward a vanishing point of the input image.

4. The lens-attached matter detector according to claim 2, wherein the mask processor eliminates from the attention area an area including a track on a road surface, which is generated on the input image.

5. The lens-attached matter detector according to claim 2, wherein the mask processor eliminates from the attention area an area including a shadow of an attached object provided with the imager, which is generated on the input image.

6. The lens-attached matter detector according to claim 2, wherein the mask processor eliminates from the attention area an area including an image of a light source, which is generated on the input image.

7. The lens-attached matter detector according to claim 1, wherein when the brightness value of the attention area obtained by the brightness distribution extractor is smaller than a predetermined value by being compared with the circumferential area, and the time series variation in the brightness value obtained by the brightness change extractor is smaller than a predetermined value, the attached matter determiner determines that attached matter having a low permeability is attached to the attention area.

8. The lens-attached matter detector according to claim 1, wherein when the brightness value of the attention area obtained by the brightness distribution extractor is larger than a predetermined value by being compared with the circumferential area, and the time series variation in the brightness value obtained by the brightness change extractor is larger than a predetermined value, the attached matter determiner determines that attached matter having a high permeability is attached to the attention area.

9. A lens-attached matter detection method which is executed by the lens-attached matter detector according to claim 1, comprising

an edge extraction process that creates an edge image based on an input image, divides the edge image into a plurality of areas including a plurality of pixels, and extracts an area where edge intensity is within a threshold range as an attention area;

a brightness value distribution extraction process that obtains a brightness value of the attention area and a brightness value of a circumference area of the attention area;

a brightness value extraction process that obtains the brightness value of the attention area and the brightness value of the circumference area obtained by the brightness value distribution extraction process for a predetermined time interval, and that obtains a time series variation in the brightness value of the attention area based on the brightness value of the attention area for the predetermined time interval; and

an attached matter determination process that determines the presence or absence of attached matter based on the time series variation in the brightness value of the attention area.

10. The lens-attached matter detection method according to claim 9, further comprising a vehicle information-obtaining process that obtains vehicle information including at least vehicle speed information, wherein

all of the processes are completed when the vehicle information-obtaining process determines that a vehicle speed is a threshold or below based on the vehicle speed information.

11. The lens-attached matter detection method according to claim 9, wherein all of the processes are completed when a gain adjustment value of the input image is a threshold or more.

12. A vehicle system, comprising

a lens- attached matter detector according to claim 1;

an imager provided in a vehicle, having a lens, and that images a circumference of the vehicle; and

at least one application that operates based on detection information of the lens-attached matter detected by the lens-attached matter detector relative to the input image shot by the imager.

13. The vehicle system according to claim 12, wherein the lens-attached matter detector detects the lens-attached matter relative to only an area for use in the application in the input image shot by the lens.