PROCESS AND DEVICE FOR DETECTION OF DROPS IN A DIGITAL IMAGE AND COMPUTER PROGRAM FOR EXECUTING THIS METHOD

• US 20130044910A1
• Filed: 02/08/2011
• Published: 02/21/2013
• Est. Priority Date: 02/09/2010
• Status: Active Grant

• Current Assignee: Institut francais des sciences et technologies des transports, de l'amenagement et des reseaux (IFSTTAR)
• Original Assignee: Institut francais des sciences et technologies des transports, de l'amenagement et des reseaux (IFSTTAR)
• Inventors: Aubert, Didier, Cord, Aurelien
• Granted Patent: US 9,058,643 B2
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: B60S 1/0844 including a camera
• CPC Class Codes: G06T 2207/30248 Vehicle exterior or interior
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/12 Edge-based segmentation

Abstract

A method for detection of drops on a transparent wall through which digital images (I11-I15,I111-I115,I211-I215,I311-I315) are acquired by means of an image acquisition system, comprising the following steps:

• b) a gradient image is established from the digital image (I16) by assigning to each pixel the value of the gradient for the pixel;
• c) the gradient image is filtered by assigning to each pixel a value obtained by rank filtering;
• d) a mask representing the position of the detected drops in the digital image is established by activating the pixels for which the difference between the value of the pixel in the filtered image and the value of the pixel in the gradient image exceeds a predetermined value.

This method enables robust detection of drops in the image or the sequence of images.

Program and device for detection of drops for executing this method.

Claims

1. A method for detection of drops appearing in a digital image acquired through a transparent wall by means of an image acquisition system, the method comprising the following steps:

b) a gradient image is established from said digital image, by assigning to each pixel the value of the gradient at the position of the pixel;

c) said gradient image is filtered to obtain a filtered image, by assigning to each pixel a value obtained by applying a rank filter to a neighbourhood of the pixel;

d) a mask is established representing the position of the detected drops in said digital image, in which mask the pixels for which the difference between the value of the pixel in the filtered image and the value of the pixel in the gradient image exceeds a predetermined value are activated.

2. The method as claimed in claim 1, further comprising the following step:

e) the mask is completed by activating therein non-activated pixels, located between neighbouring activated pixels.

3. The method as claimed in claim 1, further comprising the following step:

f) the mask is refined by deactivating therein groups of activated connected pixels fulfilling at least one predetermined criterion of elimination, relating to a property of the group of connected pixels.

4. The method as claimed in claim 3, in which a criterion of elimination taken into account at step f) is fulfilled if and only if the number of pixels of the group of activated connected pixels is less than a first predetermined value and/or greater than a second predetermined value.

5. The method as claimed in claim 3, in which a criterion of elimination taken into account at step f) is fulfilled if and only if the ratio between the height and the width of the group of activated connected pixels is less than a third predetermined value and/or greater than a fourth predetermined value.

6. The method as claimed in claim 3, in which a criterion of elimination taken into account at step f) is fulfilled if and only if the eccentricity of the group of activated connected pixels is less than a predetermined value.

7. A method for detection of drops from a plurality of digital images, following which in an initial step a) said plurality of digital images is averaged so as to obtain a mean image which forms the digital image on the basis of which the method as claimed in claim 1 is carried out.

8. The method for detection of drops as claimed in claim 1, further comprising the following step:

g) drops in the mask obtained are identified, each drop corresponding to a group of activated connected pixels.

9. A method for detection of drops appearing in a plurality of digital images acquired through a transparent wall by means of an image acquisition system, the method comprising the following steps:

a plurality of masks is established successively from said digital images by the method defined by claim 8;

during a step h) a confirmed mask is established containing the drops appearing in at least one predetermined proportion of masks obtained in this way.

10. The method for detection of drops as claimed in claim 8, comprising further the following step:

i) a control signal is sent to a machine, said signal being function of the number or of the density of drops identified at step g), especially in the case where the machine is a window wiper system.

11. A computer program comprising instructions for executing a method for detecting drops as claimed in claim 1.

12. A device for detection of drops, comprising an image acquisition system and a calculator connected to the latter for receiving images acquired through a transparent wall by the image acquisition system, the calculator being suitable for executing the method as claimed in claim 1.

13. The method for detection of drops as claimed in claim 7, further comprising the following step:

g) drops in the mask obtained are identified, each drop corresponding to a group of activated connected pixels.

14. A computer program comprising instructions for executing a method for detecting drops as claimed in claim 7.

15. A device for detection of drops, comprising an image acquisition system and a calculator connected to the latter for receiving images acquired through a transparent wall by the image acquisition system, the calculator being suitable for executing the method as claimed in claim 7.