Object presence detection method and device

• US 6,911,642 B2
• Filed: 02/16/2001
• Issued: 06/28/2005
• Est. Priority Date: 02/18/2000
• Status: Expired due to Term

• Current Assignee: Ficosa North America Corporation (Panasonic Holdings Corporation)
• Original Assignee: Ficosa North America Corporation (Panasonic Holdings Corporation)
• Inventors: Daura Luna, Francesc, Bande Martinez, Daniel, Tecchiolli, Giampietro, Sartori, Alvise, Tarrago Pujol, Jose Maria, Crespi, Bruno
• Primary Examiner(s): Allen, Stephone B.
• Application Number:
  US10/203,503
• Publication Number:
  US 20030141762A1
• Time in Patent Office: 1,593 Days
• Field of Search: 250/208.1, 250/221, 250/222.1, 340/435, 340/436, 340901-904, 340555-557, 367/93, 367/909
• US Class Current: 250/208.2
• CPC Class Codes: G01S 11/12 using electromagnetic waves other than radio waves
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
• CPC Class Codes: G08G 1/167 Driving aids for lane monitoring, lane changing, e.g. blind spot detection

Abstract

An object presence detection device, of the type mounted to a motor vehicle having at least one blind spot, where the detection device is for detecting an object situated in the blind spot, comprises: a receiver for detecting electromagnetic waves, comprising a focussing device, and a light detector converting said received electromagnetic waves into electrical signals; an electronic circuit converting the electrical signals into digitized signals; a logic circuit analyzing the digitized signals to analyze the presence of objects in the blind spot which are moving relative to the vehicle, and emitting variable output signals depending on the result of the analysis; indicator members activated by the output signals, suitable to be perceived by the driver.

Claims

1. A method of object presence detection, by using a detection device of the type mounted to a motor vehicle, said vehicle having at least one blind spot, where said detection device is for detecting an object such as an approaching vehicle situated in said blind spot, comprising:

(a) a receiver (7) for detecting electromagnetic waves, said receiver (7) comprising a focusing device, and a light detector including a set of active photosensor elements that converts said electromagnetic waves into electrical signals, said light detector defining an image surface,

(b) an electronic circuit (9) converting said electrical signals into digitized signals,

(c) a logic circuit (15) analyzing said digitized signals to analyze the presence of objects in said at least one blind spot which are moving relative to said vehicle, and generating variable output alarm signals depending on the result of said analysis, and

(d) indicator members activated by said output signals, said detection device being adapted to distinguish a vehicle from other objects, wherein said distinguishing is effected by recognizing arrises, clearly marked on at least a first portion of said image surface of interest in vertical and/or horizontal directions on a road surface and the follow-up of the movement of said arrises through various successive images to calculate a relative speed of a detected object providing said arrises.

2. A method, according to claim 1, wherein on a second portion of said image surface a different technique is used based on a phase difference to obtain an estimate of an optical flow in one direction.

3. A method according to claim 1, wherein said recognition of arrises is performed by said logic circuit (15) carrying out a mathematical convolution over at least said first portion of said image surface of the digitised signal thereafter distinguishing a vehicle from other objects.

4. A method according to claim 3, wherein said mathematical convolution over at least said first portion of said image surface is carried cut by means of a convolution kernel appropriate for a movement detection.

5. A method according to claim 3, wherein said convolution is calculated in time no greater than 100 ms.

6. A method, according to claim 1, wherein:

said portion of the image of interest is defined by a mask controlled by software and in that an algorithm calculates the arrises of the image and their directions, said vertical and horizontal arrises being normalized and integrated along the vertical and horizontal axes respectively;

a normalization factor is adjusted dynamically based on average density of said arrises, using unidirectional projection of said arrises on corresponding coordinates axis to trace a trajectory according to said axis so that approaching objects generate projections points with positive speeds, and said projection points are separated from other points that are stationary or have relative movements with the aid of directions filters finally identifying and selecting the resulting trajectories.

7. A method according to claim 6, wherein:

once the trajectory of an approaching vehicle has been identified by means of the displacement of its corresponding unidirectional projections a speed of said vehicle is estimated and supervised so that it is possible to detect situations in which this approaching vehicle reduces speed and circulates at same speed as said vehicle fitted with said device, staying at a short distance or vehicles circulating in different lanes at practically a same speed, and

a specific type of alarm signal is emitted until substantial changes are observed in an environment of said image of an approaching vehicle.

8. A method according to claim 6 wherein:

said mark controlled from a software that defines said image surface of interest is positioned in such a way that a car located in the overtaking lane of a straight road and at a great distance is located in the left upper end of the mask, and the position of the mask with regard to the image can be adjusted to get a fine adjustment of the field of vision.

9. A method according to claim 2, wherein in order to estimate an optical flow in said one direction of an axis of a road a pair of successive images is processed and visual elements that are not moving forward are filtered and eliminated and in that said procedure is optional and may not be used.

10. A method according to claim 7 wherein in case of a left turn of the motor vehicle constant displacement of visual elements contained in different images are detected in an upper portion of successive images by means of a correlations technique and coherence of said signal during a number of successive images is used as an indication that a vehicle fitted with the device is turning so that as the detection field moves to a position closer to said vehicle more restrictive requirements are imposed before activating said alarm signals.

11. A method according to claim 7 and further additionally detecting whether said vehicle fitted with said device has initiated actions meaning an approach to said object.

12. A method according to claim 11 wherein said indicative actions comprise the activation of an indicator.

13. A method according to claim 1 wherein said indicative actions comprise turning a steering wheel.

14. A method according to claim 1 wherein said alarm signals include luminous signals of at least two colours, where each colour indicates a different warming level.

15. A method according to claim 1 wherein it further includes means for detecting drowsiness of the driver using said detection device and emitting an alarm signal depending on the relative position of the vehicle fitted with the detection device to the lines marking the lanes on the road surface.

16. A method according to claim 1, wherein said set of photosensor elements are integrated in a digital camera mounted in a rearview mirror of a vehicle and in that said camera is positioned in such a way that:

a left vertical edge of said image surface is close to a side edge of a vehicle fitted with said device;

an upper edge of said image surface is lightly above the horizon line, by approximately one eighth of the image,

said camera is slightly tilted in the clockwise direction so that it aligns said image surface along an axis of a road.

17. An object presence detection device to put into practice a method according to claim 1, said detection device being of the type mounted to a motor vehicle, said vehicle having at least one blind spot, where said detection device is for detecting an object such an approaching vehicle situated in said blind spot and comprising:

(a) a receiver (7) for detecting electromagnetic waves, said receiver (7) comprising a focusing device, and a light detector including a set of active photosensor elements that converts said electromagnetic waves into electrical signals, said light detector defining an image surface,

(b) an electronic circuit (9) converting said electrical signals into digitized signals,

(c) a logic circuit (15) analyzing said digitized signals to analyze the presence of objects in said at least one blind spot which are moving relative to said vehicle, and generating variable output alarm signals depending on the result of said analysis, and

(d) indicator members activated by said output signals, said detection device being adapted to distinguish a vehicle from other objects, wherein said active photosensor elements are integrated in a digital camera, and said digital camera is positioned in such a way that;

a left vertical edge of said image surface is close to a side edge of a vehicle fitted with said device;

an upper edge of said image surface is slightly above the horizon line, by approximately one eighth of the image,

said camera is slightly tilted in the clockwise direction so that it aligns said image surface along an axis of a road.

18. A detection device as claimed in claim 17, wherein said digital camera is housed in a vehicle external rear-view mirror unit.

19. A detection device according to claim 18, including a digital camera in each of said external rear-view mirrors of said vehicle.

20. A detection device as claimed in claim 18, wherein, a radius of action of said detection device is greater than 15 m.

21. A detection device, according to claim 20 wherein said detection device has a radius of action greater than 20 m.

22. A detection device, according to claim 17, wherein said electronic circuit (9) is of a CMOS, DMOS, MOS, Si-Ge or BiCMOS technology.

23. A detection device according to claim 17, wherein said light detector is CMOS, DMOS, MOS, Si-Ge or BiCMOS technology.

24. A detection device according to claim 17, wherein said electronic circuit (9) is SOI (silicon on insulator) technology.

25. A detection device according to claim 17, wherein said light detector is SOI (silicon on insulator) technology.

26. A detection device according to claim 17, wherein it comprises a mask controlled by software which allows to perform a partial selection of said active sensor elements so that said electronic circuit converts into digitised signals only electrical signals from said partial selected active photosensor elements.

27. A detection device according to claim 26 wherein:

said light detector is constituted by active sensor elements having a dynamic range equal or superior to six decades, i.e. between a minimum detection threshold value and a saturation threshold there is a range of six decades, with the light intensity being expressed in lux, on one same image surface; and

one said sensor element has a dynamic range equal or superior to the six decades, as defined above, between two consecutive images.

28. A detection device according to claim 27, wherein aid light detector comprises a two dimensional matrix of said sensor elements defining a number of rows of sensor elements, where each of said rows extends in a first direction and is parallel to the remaining rows, and where the group of rows forms said image surface, and each of said rows is provided with at the most 320 active sensor elements for processing the image and said light detector includes at the most 256 active rows of sensor elements for processing for processing the image.

29. A detection device according to claim 26 wherein said logic circuit (15) comprises a specialized electronic circuit including (a) a Von Neumann type sequential central processing unit (CPU), (b) a parallel coprocessor (TOT), specialized in the calculation of a convolution over at least a portion of said image surface of the digitised signal, said coprocessor including at least 32 partial accumulator multipliers with a high calculation speed for calculating said convolution directly on the image surface at such a calculation speed that the convolution is completed before a new image acquisition is initiated, and (c) a local RAM (SRAM).