Classifications

• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T7/00—Image analysis
  • G06T7/20—Analysis of motion
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
  • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
  • G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
  • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
  • G06F3/03—Arrangements for converting the position or the displacement of a member into a coded form
  • G06F3/0304—Detection arrangements using opto-electronic means
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
  • G06V10/00—Arrangements for image or video recognition or understanding
  • G06V10/10—Image acquisition
  • G06V10/12—Details of acquisition arrangements; Constructional details thereof
  • G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
  • G06V10/147—Details of sensors, e.g. sensor lenses
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/10—Image acquisition modality
  • G06T2207/10028—Range image; Depth image; 3D point clouds
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/10—Image acquisition modality
  • G06T2207/10048—Infrared image
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/30—Subject of image; Context of image processing
  • G06T2207/30248—Vehicle exterior or interior
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
  • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
  • G06V40/20—Movements or behaviour, e.g. gesture recognition
  • G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language

Definitions

• the present invention generally relates to the field of processing data acquired by a sensor.
• It relates more particularly to a data processing method and an associated embedded system.
• a sensor generally a matrix sensor
• acquires a set of pixels which are then processed in order to obtain information relating to the environment facing the sensor.
• Such solutions are used in particular in the field of vehicles, for example to build a map of the environment located at the front of the vehicle or to recognize gestures made by the driver of the vehicle.
• processing algorithms used are, however, more and more complex and thus consuming material resources (microprocessor, memory, etc.).
• the invention proposes a method of data processing comprising the following steps:
• the processing algorithm is thus applied to only a subset of the acquired pixels, which limits the volume of the manipulated data. Thanks to the location of the object, the treatments performed focus on the area of the environment for which the application of the processing algorithm is of interest.
• Other non-limiting and advantageous features of the data processing method taken individually or in any technically possible combination, are as follows:
• the sensor is matrix
• the first set of pixels is a matrix of pixels
• the step of determining the second set of pixels comprises a step of determining an area comprising said position and / or a step of extracting the pixels of the first set located in said area;
• the senor delivers a plurality of first sets of pixels
• the method comprises a step of determining, for each of said first sets of pixels, a second set of corresponding pixels by extracting the pixels of the first set concerned located in said area (the same area being thus used for the different first sets) ;
• the senor is part of a time of flight sensor further comprising an electromagnetic radiation emitting element
• said first sets are acquired by the sensor respectively in correspondence with distinct phases of the (modulated) signal emitted by the emitting element of electromagnetic radiation;
• said position is determined by analysis of an image representative of the environment (such as at least a part of the first set of pixels or an image obtained by means of the aforementioned processing algorithm);
• said position is determined by tracking the object (typically by evaluating a speed of the object in the image between two successive images and estimating the new position of the object on the basis of this speed and a position of the object in the previous image).
• the invention also proposes an embedded system comprising a processing unit and an environment-facing sensor, wherein the sensor is designed to acquire a first set of pixels and wherein the processing unit comprises a module designed to determine a position of an object of the environment, a module designed to determine a second set of pixels included in the first set of pixels (the second set of pixels covering said position and being distinct from the first set), and a module designed to apply a pixel processing algorithm of the second set only.
• the processing unit comprises a module designed to determine a position of an object of the environment, a module designed to determine a second set of pixels included in the first set of pixels (the second set of pixels covering said position and being distinct from the first set), and a module designed to apply a pixel processing algorithm of the second set only.
• FIG. 1 represents an onboard system including a flight time sensor
• FIG. 2 represents the functional modules of a processing unit of the onboard system of FIG. 1;
• FIG. 3 represents pixel matrices processed by a processing method using the invention.
• FIG. 1 schematically shows a system embedded in a vehicle, here a motor vehicle, and comprising a flight time sensor 1 and a processing unit 10.
• the flight time sensor 1, here a three-dimensional camera based on the principle of flight time comprises at least one element emitting electromagnetic radiation 2 (typically one or more electroluminescent diode (s) emitting in the infrared) and a matrix sensor 4 (such as an image sensor sensitive to radiation emitted by the emitter element 2, here the infrared) defining a set pixels.
• electromagnetic radiation 2 typically one or more electroluminescent diode (s) emitting in the infrared
• a matrix sensor 4 such as an image sensor sensitive to radiation emitted by the emitter element 2, here the infrared
• the radiation E emitted by the emitter element 2 (generally through an optical transmission system not shown in FIG. 1) is reflected towards the matrix sensor 4 (radiation referenced R in FIG. 1) by the first object encountered on the path E. radiation
• an optical system (such as a lens) is placed facing the matrix sensor 4 so that each pixel of the matrix sensor 4 receives the reflected signal R coming from a particular direction of the solid angle analyzed by the time of flight sensor 1.
• a processing unit 10 controls the emission of the radiation E by the transmitter element 2, and then analyzes the signals measured by the matrix sensor 4 (represented here in the form of a plurality of matrices M k of pixels) so as to determine in particular a matrix D comprising the distances d 1 of the first object encountered for a plurality of directions of space facing the flight time sensor 1.
• This matrix D of the distances d 1 (as well as possibly other information, such as a luminance matrix L) is transmitted by the processing unit 10 to another electronic system (not shown) for use by the latter.
• the flight time sensor 1 can be placed at the front of the vehicle in order to build a map of the front environment of the vehicle and / or to detect an obstacle and / or to evaluate the speed of another vehicle located at the front (by deriving a distance of the vehicle evaluated on the basis of some of the distances di).
• the flight time sensor can be placed in the passenger compartment of the vehicle (for example facing the driver of the vehicle) and the distances di determined by the processing unit 10 can be used within a gesture recognition algorithm.
• the processing unit 10 controls the acquisition by the matrix sensor 4 of a plurality of matrices M k of pixels, at times respectively corresponding to a plurality of distinct phases (two by two) of the modulated signal.
• FIG. 2 shows the functional modules of the processing unit
• Each functional module 12, 14, 1 6 represented in FIG. 2 corresponds to a particular functionality implemented by the processing unit 10.
• Several (if not all) functional modules can however in practice be implemented by the same entity. physical, here a processor of the processing unit 10, on which executes program instructions stored in a memory associated with the processor (each functional module then being implemented in this case by the execution of a particular game instructions stored in this memory).
• the processing unit 10 thus comprises a location module 12, a selection module 14 and an analysis module 1 6.
• the location module 12 is designed to locate (roughly) an object present in the solid angle observed by the matrix sensor 4.
• the location of the object is achieved by analyzing an image representative of the environment facing the matrix sensor 4 in the field of view of the matrix sensor 4.
• This representative image may be in practice one of the matrices M k of pixels produced by the matrix sensor 4 or the luminance matrix L produced as indicated below by the analysis module 1 6, possibly (in both cases) after a downsampling step to reduce the volume of processed data.
• This first possibility can be used in particular if no object has been detected beforehand.
• the location of the object is achieved by tracking the object and estimating the current position of the object (typically based on the position of the object at the previous iteration and an estimated displacement of the object of an iteration to the next iteration).
• the location module 12 When the location module 12 locates an object, the location module 12 transmits POS position information to the selection module 14.
• the POS position information comprises, for example, coordinates of the detected object expressed with respect to the matrix of pixels acquired by the matrix sensor 4. These coordinates make it possible to define the position and / or the extent of the object concerned in the solid angle observed by the matrix sensor 4.
• the selection module 14 also receives as input the pixel matrices M k acquired by the matrix sensor 4.
• the selection module 14 receives no position information POS and then transmits the pixel matrices M k received at the input to the analysis module 1 6 (without modification).
• the location module 12 transmits the POS position information to the selection module 14 as already indicated.
• the selection module 12 is then designed to:
• zone Z includes all the pixels located 10 pixels or less (vertically or horizontally) from the position defined by the POS position information (including the defined pixel or the pixels defined by the POS position information). .
• Each set S k of pixels is therefore a subset of the set of pixels of a corresponding matrix M k of pixels.
• the same zone Z is used to extract the pixels of the different matrices M k of pixels.
• the analysis module 1 6 is designed for applying at least one processing algorithm to the sets S k of pixels received at the input.
• the analysis module 1 6 implements a first processing algorithm making it possible to obtain the matrix D of the distances di j on the basis of sets S k of pixels received at the input (when the module of selection 14 carries out the aforementioned extraction, or on the basis of matrices M k of pixels when no object is detected by the location module 12 and that the selection module 14 then simply transmits these matrices M k of pixels).
• the analysis module 1 6 here also implements a second processing algorithm, which in turn makes it possible to obtain the above-mentioned luminance matrix L on the basis of sets S k of pixels received at input (or, as previously, on the base of matrices M k of pixels).
• the processing algorithms are applied to only part of the pixels (the sets S k ), which makes it possible to reduce the volume of the treatments performed.
• the selection module 14 extracts pixels from the matrices M k of pixels as indicated above and transmits to the analysis module 1 6 sets S k of pixels, the matrix D of the distances d 1 j and / or the matrix L of luminance are usually partial (ie do not include data for the entire field of view of the matrix sensor 4).
• the matrix sensor 4 acquires at step E2 at least one matrix of pixels, here a plurality of matrix M k of pixels as explained above.
• the locating unit 12 detects an object O in the environment facing the matrix sensor 4 (step E4) and determines its POS position, for example by analyzing one of the matrices M k of pixels. Alternatively, the detection of the object O could be performed by analyzing a luminance matrix L calculated by the processing unit 10.
• the selection module 14 determines in step E6 a zone Z covering the object O (and for example larger than the object O to take into account an error margin).
• the selection module 14 For each matrix M k of pixels, the selection module 14 extracts the pixels located in the zone Z in order to obtain a set S k of corresponding pixels (step E8).
• the analysis module 1 6 can thus apply at step E10 at least one processing algorithm to sets S k of pixels (and to these pixels only), here in order to obtain a (partial) matrix D of distances dij or a (partial) matrix of luminance L.

Landscapes

• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• General Physics & Mathematics (AREA)
• Physics & Mathematics (AREA)
• General Engineering & Computer Science (AREA)
• Multimedia (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Health & Medical Sciences (AREA)
• General Health & Medical Sciences (AREA)
• Vascular Medicine (AREA)
• Human Computer Interaction (AREA)
• Optical Radar Systems And Details Thereof (AREA)
• Testing Or Calibration Of Command Recording Devices (AREA)
• Image Processing (AREA)
• Image Analysis (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=61027878

Family Applications (1)

Country Status (3)

Families Citing this family (1)

Family Cites Families (2)

• 2017
  • 2017-10-19 FR FR1759875A patent/FR3074950B1/fr active Active
• 2018
  • 2018-10-17 WO PCT/EP2018/078465 patent/WO2019077010A1/fr not_active Ceased
  • 2018-10-17 EP EP18785645.5A patent/EP3698321A1/de active Pending

Non-Patent Citations (4)

Also Published As

Similar Documents

Legal Events