FR3055997A1 - SYSTEM FOR DETERMINING AT LEAST ONE CHARACTERISTIC RELATING TO THE CONTOUR OF A SUBJECT CONTAINED IN AT LEAST ONE DIGITAL IMAGE - Google Patents

Abstract

The invention relates to a system for determining at least one characteristic relating to the contour of a subject (S) contained in at least one digital image of interest. This determination system comprises: a) at least one background structure (1) comprising at least one retro-reflective part (12), b) a photographic assembly comprising suitable imaging means and light means providing artificial lighting of the scene; and c) image processing means adapted to determine at least one feature relating to the contour of said subject from at least one digital image of interest corresponding to a scene with audit exposure. artificial lighting and wherein said subject (S) is positioned in front of said retroreflective portion (12).

Description

Technical field to which the invention relates

The present invention relates to the field of digital image processing.

It relates more precisely to a system for determining at least one characteristic relating to the outline of a subject, for example an animal (in particular a fish), contained in at least one digital image of interest.

Technological background

In various fields, it may be advantageous to be able to determine, in a simple and rapid manner, at least one characteristic relating to the outline of a subject (for example a dimensional characteristic of the latter).

This is the case, for example, in the area of fishing, where tools are needed to record and record the catches of fishermen in a fully automated manner.

One possible solution is to take a digital image of this subject, followed by automatic image processing of this digital image.

However, this solution is in practice particularly complex and delicate to implement, in particular when the shooting is carried out in an environment in which the ambient lighting, even the perspective, is not controllable.

Automatic image processing must indeed be able to determine the outline of the subject, despite the drop shadows which are generated by the ambient lighting of the scene and which are likely to complicate the interpretation.

This automatic image processing must also be able to correct the perspective distortion of the digital image, so as to be able to determine reliable data as to the characteristics sought.

There is therefore a need for a technical solution suitable for determining at least one characteristic relating to the outline of a subject contained in at least one digital image, regardless of the ambient lighting of the scene, or even from the perspective of this. digital image.

Object of the invention

In order to remedy the aforementioned drawbacks of the prior art, the present invention provides a system for determining at least one characteristic relating to the outline of a subject contained in at least one digital image of interest.

This determination system includes:

a) at least one background structure comprising a front face having at least one retro-reflecting part in front of which said subject is intended to be positioned,

b) a photographic unit comprising:

means for taking images comprising a photographic objective, for taking digital images each containing a scene situated opposite said photographic objective,

- light means, adapted to provide artificial lighting of said scene and arranged so that light rays emitted by said light means are capable of being reflected towards said photographic objective by the retro-reflecting part situated opposite said photographic objective, and

means for coordinating the operation of said image taking means and said light means, designed for capturing digital images containing said scene captured with or without exposure to said artificial lighting, and

c) image processing means designed to determine at least one characteristic relating to the contour of said subject from at least one digital image of interest corresponding to a scene with exposure to said artificial lighting and in which said subject is positioned in front of said retroreflective part.

In practice, for taking digital images of interest, the subject is placed on the retro-reflecting part of the background structure.

When taking an image, the light rays emitted by the light means are reflected towards the photographic objective by the retroreflective part located opposite.

This mechanism then makes it possible to obtain digital images in which the possible shadows cast around the subject are reduced (or even eliminated).

The outline of the subject is thus particularly sharp and easy to determine in digital images of interest by an automatic image processing process.

Other non-limiting and advantageous characteristics of the system according to the invention, taken individually or in any technically possible combination, are the following:

the processing means comprise a module for transforming said at least one digital image of interest into a binary image composed of first pixels originating from pixels of said at least one digital image of interest, formed by the visible portion of the retro-reflecting part which is located outside the space occupied by said subject, and from second pixels originating from pixels of said at least one digital image of interest, other than said visible portion of the retro-reflecting part; in this context, the transformation module advantageously cooperates with, on the one hand, a sampling module to collect samples of brightness values on the front face contained in said at least one digital image of interest: at least a first sample collected on a visible portion of the retroreflective part of said front face, and at least a second sample collected on a non-retro-reflecting part of said front face and, on the other hand, a threshold calculation module, for determining, from said samples, at least one brightness threshold value used by the transformation module to distinguish said first and second pixels from the binary image; still in this context, the processing means advantageously comprise a module for detecting said subject in the binary image, which detection module advantageously consists of a module designed to detect the group of second pixels in said binary image whose number is the most Student ;

the processing means comprise: a module for determining the outline of said subject contained in said at least one digital image of interest, a module for determining at least one of the dimensions of the outline of said subject, preferably its greatest length, and optionally a module for identifying said subject, for example its species in the case of an animal subject;

the processing means comprise a module for correcting the perspective deformation of said at least one digital image of interest; preferably, the front face of the background structure comprises at least one geometric reference, and the correction module is designed to take account of said at least one geometric reference; more preferably, said at least one geometric reference comprises at least four geometric references corresponding to the four corners of a quadrilateral; always preferably, the processing means comprise a module for real-time detection of said at least one geometric mark within the scene, and a module for authorizing digital image capture by the image capture means when said at less a geometric coordinate system is detected;

the processing means comprise at least one computer program comprising program code means and intended to be executed;

- The system also includes a support member provided with means for supporting the photographic assembly and means for supporting the background structure, to fix the position of the photographic assembly opposite said background structure;

- the photographic unit and the processing means are grouped in a portable electronic device, for example electronic tablet or smart phone;

- the light means emit light rays in the visible and / or infrared range (preferably near infrared);

- the background structure is impermeable to liquids, in particular water and lipids.

According to a particular embodiment, the processing means comprise at least one analysis mode designed to determine at least one characteristic relating to the outline of said subject, chosen from:

- a first mode of analysis, starting from a single digital image of interest captured with exposure to said artificial lighting,

- a second mode of analysis, starting from a comparison of at least two digital images of interest, captured one with exposure to said artificial lighting and the other without exposure to said artificial lighting.

In this case, said system advantageously comprises means for determining the ambient brightness; and the processing means include means for selecting the first, or the second, analysis mode taking into account said ambient brightness.

The present invention also relates to a background structure for taking images, which background structure comprises a front face comprising at least one retro-reflecting part in front of which a subject is intended to be positioned, which face frontal comprises at least one geometrical reference adapted to an operation of correction of a perspective deformation.

The invention also provides a method for determining at least one characteristic relating to the outline of a subject from at least one digital image of interest, by implementing a system according to the invention.

The process includes the following steps:

- preparation of the scene of interest by positioning said subject in front of the retro-reflecting part of the front face of said background structure,

- capture at least one digital image of interest, with exposure to artificial lighting, and

- processing of said at least one digital image of interest, to determine at least one characteristic relating to the outline of said subject.

Other non-limiting and advantageous characteristics of the process according to the invention, taken individually or in any technically possible combination, are the following:

the processing step comprises an operation for correcting the perspective deformation of said at least one digital image, an operation for transforming said at least one digital image into a binary image, and an operation for determining the outline of the subject contained in said binary image;

- the processing step includes an operation to determine the size and / or the species of a subject of interest belonging to the fish family.

Detailed description of an exemplary embodiment

The description which follows with reference to the accompanying drawings, given by way of nonlimiting examples, will make it clear what the invention consists of and how it can be carried out.

In the accompanying drawings:

- Figure 1 shows, front view, a background structure belonging to the system according to the invention;

- Figures 2 and 3 show a portable electronic device, forming a photographic assembly and image processing means, also belonging to the system according to the invention;

- Figure 4 is a block diagram illustrating the main operations during the implementation of the system according to the invention;

- Figure 5 shows a digital image of interest, corresponding to a scene in which the subject is positioned in front of the retroreflective part of the background structure according to Figure 1;

- Figure 6 shows the digital image of interest according to Figure 5, following a cropping operation and an operation to correct perspective distortion;

- Figure 7 shows the digital image of Figure 6, after a transformation operation into a binary image.

In general, the present invention relates to a system for determining at least one characteristic relating to the outline of a subject contained in at least one digital image of interest.

Topic

The system according to the invention can be implemented in different fields, for the study of subjects which consist, for example, of objects or animals.

By “animal” is meant in particular small animals (for example less than 200 cm, even less than 150 cm, or even even less than 120 cm), in particular fish.

Digital image

The term “digital image” designates a photograph which is acquired, processed and stored in binary form.

In this case, such a digital image is:

- acquired by analog-digital converters located in a photographic unit (digital camera or camcorder);

- processed using graphic tools, in particular so as to extract characteristics relating to the outline of the subject contained in this image;

- stored on a computer medium (USB key, SSD, hard disk, CDROM, etc.).

By "digital image" is meant in particular a two-dimensional image or "2D image", which consists of points called pixels.

Characteristic relating to the outline of the subject

The system according to the invention allows the determination of at least one characteristic relating to the outline of the subject.

The “outline” of the subject corresponds, in the digital image of interest, to a succession of juxtaposed pixels which delimit the periphery of the subject with respect to the background.

In this case, this outline corresponds to the line of demarcation between subject S and the underlying retro-reflecting part described below.

By “characteristic relating to the contour”, one advantageously includes:

- the actual outline of the subject,

- at least one of the dimensional characteristics of the outline of the subject, preferably its greatest length, and

- the identification of the subject, for example its species in the case of an animal subject.

Determination system

To determine these characteristics relating to the contour, the determination system according to the invention comprises:

- at least one background structure 1 (shown diagrammatically in FIG. 1), and

- at least one portable electronic device 2 (shown schematically in Figures 2 and 3), which comprises a photographic assembly 3 (Figure 3) and image processing means 4 (Figure 3).

Background structure

The background structure 1 is intended to be placed in the background of the subject S when taking digital images.

This background structure 1 advantageously consists of a panel, here rectangular.

The background structure 1 is advantageously made of a flexible material; it can also be made of a foldable rigid material, or non-foldable.

The front face 11 of this background structure 1, also rectangular, has several elements:

a retro-reflecting part 12 in front of / on which the subject S is intended to be positioned, and

- A set of geometric markers 13, useful in particular for correcting perspective distortion and for determining the dimensional characteristics of the subject S captured.

The retro-reflecting part 12 advantageously consists of a coating partially covering the front face 11.

This retro-reflecting part 12, when illuminated by a light beam, returns the light rays in the direction of the light source, whatever the initial angle of incidence.

The notion of "retro-reflecting" further qualifies a surface which is capable of reflecting light radiation without change of direction (the reflected light ray has the same direction as the incident light radiation).

Such a retro-reflecting part 12 is for example constituted by a prismatic reflector or by glass microbeads.

The retro-reflecting part 12 occupies an area which is greater than the size of the subject S intended to be studied. In this way, this retro-reflecting part 12 surrounds the entire outline of the subject S reported.

The retro-reflecting part 12 here has a generally rectangular shape, the corners of which are cut.

The geometric marks 13 are chosen from the geometric marks identifiable in a digital image by a computer program.

These geometric marks 13 surround the retro-reflecting part 12. They are advantageously made of a non-retro-reflecting material.

These geometric marks 13 are formed here by the four corners 131 of a peripheral frame 132 (quadrilateral) surrounding the retro-reflecting part 12.

The arrangement of the geometric markers 13, in particular the distance separating the corners 131, is known by the image processing means 4 with a view to the correction of perspective.

In general, this background structure 1 is advantageously produced by an assembly of materials which are impermeable and resistant to liquids, in particular to water and lipids.

For example :

- the background structure 1 has the following dimensions: 400 to 450 mm x 250 to 300 mm;

- The retro-reflecting part 12 has the following dimensions: 300 to 350 mm x 200 to 250 mm;

- The peripheral frame 132 has the following dimensions: 350 to 400 mm x 250 to 300 mm.

Portable electronic device

The portable electronic device 2, shown in Figures 2 and 3, consists for example of an electronic tablet or a smart phone (still commonly referred to as "smartphone").

This portable electronic equipment 2 is able to capture digital images containing the subject S, and to analyze the latter so as to determine at least one characteristic relating to the contour of this subject S in these digital images.

To this end, the portable electronic device 2 comprises:

a photographic assembly 3, to perform a function of the digital photographic camera (or digital camera) type,

- image processing means 4, for processing the images captured by the photographic assembly 3,

a man / machine interface 5 (FIG. 2), advantageously a touch screen on which the digital images captured by the photographic assembly 3 and the data collected by the image processing means 4 can be displayed,

- telecommunication means (not shown), for access to a wide area network (for example the Internet network),

- memory means (not shown), containing in particular computer programs belonging to the image processing means 4,

- computing means (not shown), of the microprocessor type for example, which execute the instructions and process the data of the computer programs.

Photographic set

The photographic set 3 includes:

- image taking means 31, for taking digital images of a scene,

- light means 32, adapted to provide artificial lighting of said scene, and

- coordination means 33, to coordinate the operation of the image taking means 31 and light means 32.

The image taking means 31 comprise in particular a photographic lens 311 associated in particular with a photographic sensor (CCD or CMOS sensor, for example) and with an analog-digital converter (not shown).

A photographic objective 311 is a converging optical system formed by at least one lens, and possibly at least one mirror, giving real images on the photographic sensor.

These image taking means 31 are thus able to capture the scene located opposite the photographic objective 311.

The light means 32 are adapted to provide artificial lighting of the captured scene.

These light means 32 advantageously emit light rays in the visible spectrum range (advantageously a wavelength between 380 nm and 780 nm) and / or infrared (preferably near infrared, namely a wavelength between 780 nm and 3 pm).

These light means 32 are advantageously arranged near the photographic lens 311.

These light means 32 are further oriented with an emission axis parallel to the axis of the photographic objective 311, so that its emitted light rays are reflected towards the photographic objective 311 by the retro-reflecting part 12 located at look.

The coordination means 33 are for their part designed to control the operation of the light means 32 parallel to the image taking means 31.

These coordination means 33 are thus designed for capturing digital images according to two modes of exposure:

- with exposure to said artificial lighting for obtaining a so-called “exposed” digital image, when the light means 32 are active during the shooting by the image taking means 31, or

- Without exposure to artificial lighting to obtain a so-called "unexposed" digital image, when the light means 32 are inactive when the image is taken by the image taking means 31.

The photographic assembly 3 also advantageously comprises means 34 for determining the ambient brightness (also called ambient light detector / sensor).

Means of treatment

The processing means 4 comprise at least one computer program which is recorded on the memory means of the portable electronic device 2.

The at least one computer program comprises several modules. Each module advantageously consists of program code instructions (algorithm) designed to perform a step / operation of a process, when it is executed by a computer.

The computer program advantageously consists of a computer program which is downloaded from an online store (mobile application) or which is accessible from a web server (mobile website).

According to the invention, the different modules are designed for the determination of at least one characteristic relating to the contour of the subject S from at least one digital image of exposed interest.

By “exposed digital image of interest” is meant a scene in which the subject S is positioned in front of the retro-reflecting part 12 of the background structure 1 and the image of which is taken with exposure to artificial lighting.

To this end, the processing means 4 advantageously comprise the following combination of modules:

a benchmark detection module, adapted to detect, in real time, the geometric benchmarks 13 within the scene to be captured,

an authorization module, to authorize a digital image capture by the image capture means 31 when the geometric markers 13 are detected in real time by the detection module,

a capture module, for capturing two digital images of interest of the same scene by means of image taking 31, one with exposure to artificial lighting and the other without exposure to artificial lighting,

- a retrieval module, for retrieving the position of the geometric markers 13 in the captured digital images of interest,

- a module for correcting the perspective of digital images of interest, taking into account data from the recovery module,

- a cropping module, to eliminate the pixels surrounding the front face 11 of the background structure 1,

- a sampling module, to collect samples of brightness values on control zones in the digital images of interest,

a threshold calculation module, for calculating at least one threshold value from said samples of control zones, said at least one threshold value being adapted to identify a visible portion 121 of the retroreflective part 12,

- a transformation module, for transforming a digital image of interest into a binary image taking into account said at least one threshold value determined by the threshold calculation module,

a subject detection module for detecting the subject in said binary image,

- a contour module, to determine the contour of the subject in said binary image,

- an orientation calculation module, to calculate the orientation of the subject in said binary image,

- a clipping module, to create a clipped image of the subject from digital images of interest,

- a rotation module, to rotate said cropped image taking into account the orientation calculation module,

- a module for determining at least one of the dimensions of the outline of said subject, preferably its greatest length, and

- optionally a module for identifying said subject, for example its species in the case of an animal subject, when said modules are executed by said portable electronic device 2.

In general, the modules of the processing means 4 are designed to extract the characteristics relating to the outline of the subject S via at least one mode of analysis chosen from:

- a first mode of analysis, in which the characteristics are obtained from only the digital image of interest exposed (with exposure to artificial lighting), and

- a second mode of analysis, in which the characteristics are obtained by comparison of the two digital images of interest of the same scene, one exposed and the other unexposed.

The second mode of analysis has the advantage of allowing a determination of the outline of the subject in the digital images of interest, by an automatic image processing method, independently of the ambient brightness.

In the case of processing means 4 integrating the two analysis modes, these processing means 4 then advantageously include means for selecting the first or the second analysis mode, taking into account the ambient brightness.

In practice, in the case of low ambient light, the processing means 4 advantageously execute the first analysis mode; alternatively, in the case of high ambient light, the processing means 4 advantageously execute the second analysis mode.

In the different analysis modes, to facilitate image processing, the computer program advantageously transforms the digital image of interest into a binary image (shown diagrammatically in FIG. 7).

In this case, such a "binary image" is composed of two types of pixels:

- the first pixels P1 which come from the pixels L1 of the digital image of interest corresponding to the visible portion 121 of the retroreflective part 12 (located outside the size of the subject S), and

- second pixels P2 which come from pixels L2 of the digital image of interest other than the visible portion 121 of the retroreflective part 12.

The second pixels P2 of the binary image (FIG. 7) here comprise at least two groups:

a first group P21, corresponding to subject S, and

- a second group P22, corresponding to the peripheral frame 132.

To perform this transformation, the sampling module collects samples of the brightness value in two distinct zones of the front face 11 contained in said at least digital image of interest:

at least a first sample is collected on the visible portion 121 of the retro-reflecting part 12 of the front face 11, and

at least one second sample is collected on a non-retroreflective part of the front face 11, for example within the peripheral frame 132.

By “brightness value”, it is advantageously meant a value for the position of the color between white and black, coded on a scale ranging from 0 to 255 respectively.

The threshold calculation module then determines said at least one threshold value suitable for distinguishing the aforementioned first and second pixels P1, P2 from the digital image of interest.

The threshold value of a pixel is advantageously a function of the original brightness value of said pixel (that is to say without exposure to artificial lighting); and this threshold value advantageously corresponds to a minimum difference in brightness from said original brightness value. In general and without limitation, the threshold value of a pixel decreases with the increase of the original brightness value of said pixel.

In a particular embodiment, said at least one threshold value is advantageously obtained by means of an affine function y = a.x + b, in which:

- x is the original brightness value of a pixel, and

- y is the threshold value, corresponding to the increase in the brightness value of said pixel with and without exposure to artificial lighting.

In the first analysis mode, the calculation module determines said at least one threshold value from:

- a brightness value for the first sample collected (retro-reflecting part) with exposure to artificial lighting, and

- a brightness value for the second sample collected (non-retro-reflecting part) with exposure to artificial lighting.

In the second analysis mode, the calculation module determines said at least one threshold value from:

- a RefIDelta value, corresponding to the difference in brightness for the first sample collected (retro-reflecting part) with and without exposure to artificial lighting (respectively RefILight and RefIDark), and

- a BlackDelta value, corresponding to the difference in brightness for the second sample (non-retro-reflecting part), with and without exposure to artificial lighting (respectively BlackLight and BlackDark).

The calculation module then applies a formula F1 for the non-retro-reflecting part:

y = a1. x + b1 in which a1 = (0 - BlackDelta) / (255 - BlackDark), b1 = BlackDelta - (a1 x BlackDark)

The calculation module also executes a formula F2 for the retro-reflecting part 12:

y = a2. x + b2 in which a2 = (0 - RefIDelta) / (255 - RefIDark), b1 = RefIDelta - (a2 x RefIDark)

In this second analysis mode, the refining function of the threshold values is between these two lines of formulas F1 and F2.

For example, the threshold value for a pixel is obtained from the following formula F3:

y = a3. x + b3 in which a3 = (a1 + a2) / 2 b3 = 0 - (a3 x 255)

The ambient brightness can also be used by the processing means 4 to adjust a margin of error of said threshold values (for example, in the case of high ambient brightness, the margin of error is increased).

The transformation module includes an algorithm which transforms each pixel of the digital image of interest into a first pixel P1 or a second pixel P2 mentioned above, this taking into account the calculated threshold value, namely:

the first pixels L1 of the digital image of interest, the brightness value of which is greater than said at least threshold value (advantageously a function of its original brightness value), are transformed into first pixels El, and

- The second pixels L2 of the digital image of interest, whose brightness value is less than said at least threshold value (advantageously depending on its original brightness value), are transformed into second pixels P2.

Among the various modules of the processing means 4, the detection module is designed to detect the group P21 of second pixels P2 in the binary image whose number is the highest.

This detection module executes for example an algorithm consisting in adding the adjoining pixels P2.

Alternatively, the detection module can also be designed to detect several subjects S cooperating with the background structure 1.

In this case, the detection module is designed to detect each group P21 of second pixels P2 in the binary image whose number of pixels is greater than or equal to a predetermined minimum number.

The contour module advantageously consists of a module for determining the polygon surrounding the group P21 of second pixels P2.

Such a module advantageously consists of an algorithm of the Canny filter type (or Canny or Canny Edge detector).

The orientation calculation module consists, for example, of determining the longitudinal axis of the subject S, for example via the orientation of the smallest contour rectangle (Minimum bounding box).

The module for determining at least one of the dimensions of the outline of said subject, preferably its greatest length, is based for example on a Delaunay triangulation.

The module for identifying said subject consists, for example, in relying on dimensional ratios of its outline (length of its generatrix, segments perpendicular to this generatrix, etc.).

Implementation of the system according to the invention

The implementation of the system according to the invention is detailed below in relation to Figures 4 to 7.

The process begins with the preparation of the scene for the capture of at least one digital image of interest (step E1).

For this, the subject S is positioned in front of the retro-reflecting part 12 of the front face 11 of the background structure 1 (FIG. 5).

Preferably, the background structure 1 is placed on a receiving surface, advantageously horizontal, with its front face 11 facing upwards; the subject S is itself deposited on the retro-reflecting part 12 of the front face 11, also horizontal.

More precisely, the subject S is positioned so as to be surrounded by the retro-reflecting part 12, especially its periphery.

The retro-reflecting part 12 is then divided into two portions:

- a so-called “masked” portion, located behind subject S, and

- A portion 121 called "visible", located outside the size of said subject S and visually accessible.

The user can then position the portable electronic device 2 in front of the background structure 1, so that the image taking means 31 are able to capture at least one digital image of interest of the scene.

In practice, the portable electronic device 2 is advantageously positioned so as to capture an image of the scene in vertical (or total) diving, or at least almost in vertical diving.

By "vertical dive" is meant taking a picture directly above the background structure 1 with its associated subject S.

The processing means 4 allow a certain perspective when taking an image, with for example an angle of at most 45 ° relative to an axis extending perpendicular to the front face 11.

To guide the user in positioning the portable electronic device 2, the scene captured by the image taking means 31 is transmitted directly to the touch screen 5. And the detection module searches for the presence in real time, or the absence of geometric markers 13 within the scene captured by the image taking means 31; the detection of the geometric markers 13 is signaled on the touch screen 5.

In an embodiment not shown, the system may also include a carrier member provided with:

- means for supporting the portable electronic device 2, and

- means for supporting the background structure 1.

This embodiment makes it possible to fix the position of the photographic assembly 3 opposite the background structure 1, and thus to know directly the perspective of the image taking with a view to its correction.

When the geometric markers 13 are detected, the authorization module then authorizes the capture of the digital images of interest.

The user can then trigger the capture of the digital images of interest by means of the capture module which advantageously performs the following successive steps:

a first image taking (step E2) in which the light means 32 are inactive, in order to capture a digital image of unexposed interest (without exposure to artificial lighting),

a determination of the ambient brightness by means 34 provided for this purpose (step E3),

an activation of the light means 32 (step E4),

- a determination of the development time for a second image capture (step E5), taking into account the determined ambient brightness, then

- a second image taking (step E6), after said focusing time, in which the light means 32 are active, for capturing a digital image of exposed interest (with exposure to artificial lighting).

In practice, the first digital image of interest (unexposed) is likely to contain drop shadows B arising / coming from ambient lighting (Figures 5 and 6).

The drop shadows B are notably generated by the subject S; these drop shadows B then extend from said subject S, over at least part of the visible portion 121 of the retro-reflecting part 12, and extend its outline Ç.

On the other hand, in the second digital image of interest (exposed), the phenomenon of drop shadows is attenuated, or even eliminated, at the level of the visible portion 121 of the retro-reflecting part 12. In fact, the light rays emitted by the light means 32 are reflected towards the photographic objective 311 by the visible portion 121 of the retro-reflecting part 12 situated opposite said photographic objective 311, which corrects the phenomena of cast shadows.

The processing means 4 then automatically transform the digital images of interest and analyze them to extract the characteristics relating to the outline Ç of the subject S in presence (or even of the subjects in presence).

For this, first of all, the digital images of interest are advantageously subjected to an operation for recovering the geometric markers (step E7) then to an operation for correcting the perspective distortion (step E8), by means of the modules recovery and correction provided for this purpose.

This operation is advantageously implemented on each digital image of interest, this being based on the geometric markers 13 present and detected by the recovery module.

The actual position of the geometric markers 13 being known, the perspective deformation correction module reverses the deformation.

Once the deformation has been corrected, the processing means 4 will be able to calculate the real distance between any point of the resulting image. This distortion correction is also useful to allow a comparison of the pixels between at least two digital images of interest.

Preferably, the digital images of interest are also cropped by the cropping module, so as to eliminate the pixels surrounding the front face 11 of the background structure 1.

For this, the trimming module is based here on the detection of the peripheral frame 132 carried by the front face 11.

Such a digital image of interest, following the operations of trimming and correcting the perspective distortion, is illustrated in FIG. 6.

From these digital images of interest, the processing means 4 determine said at least one threshold value (via the sampling module and the threshold calculation module - step E9) and transform the digital image of interest in a binary image (through the transformation module - step E10).

As previously developed, a threshold value for a pixel, advantageously determined by the execution of the second analysis mode, is advantageously a function of the original brightness value of said pixel.

The determined threshold values then allow an automatic transformation of each pixel of the digital image of interest into a first pixel P1 or into a second pixel P2 mentioned above by means of the transformation module.

As illustrated in FIG. 7, a binary image is obtained in particular in which first pixels P1 (corresponding to the visible portion 121 of the retro-reflecting part 12) surround a group P21 of second pixels P2 (corresponding to the subject S ).

The outline Ç of the subject S is then very clear and easy to determine by image processing. The drop shadows B, which may complicate this image processing, are removed.

The processing means 4 then execute successively:

- the detection module, to detect the subject S in the binary image (step E11),

the contour module, to determine the contour Ç of the detected subject S (step E12),

- the orientation calculation module, to calculate the orientation of the polygon (step E13),

- the clipping module to create the clipped image of the subject S, as well as the rotation module to rotate said clipped image as horizontally as possible (step E14).

The outline Ç of the subject S being precisely determined, the processing means 4 can measure the subject S and analyze its shape.

The processing means 4 can therefore execute the subject contour analysis modules:

- the module for determining at least one of the dimensions of the outline of said subject, preferably its greatest length, and possibly

- the module to identify said subject, for example its species in the case of an animal subject.

The data from the processing means 4 (cropped image, dimensions, identification, etc.) can be displayed on the portable electronic device 2 and recorded in a database (local and / or remote)

Application of the system to a fish type subject

In a particular embodiment, the system according to the invention is suitable for fishermen, for recording and listing the fish caught.

The processing means 4 then allow at least the collection of the size, or even of the species, of the fish caught.

By "size" is meant at least one of the lengths of various parts of the anatomy of the fish, chosen for example from standard length, total length or fork length.

By "standard" length of a fish is meant the measurement of the distance from the tip of the fish's snout to the base of the caudal fin.

The "total" length defines the distance from the tip of the fish's snout to the end of the caudal fin.

The "fork" length measures the distance from the end of the mouth to the fork of the caudal fin.

The module for identifying the subject advantageously consists of a module for determining the species of fish caught.

The characteristics obtained (relating to the subject) are then advantageously displayed on the screen of the portable electronic device 2 and recorded in a database (local and / or remote).

In this particular area of fishing, the processing means 4 advantageously combine the digital images captured with additional data relating to the conditions of image taking, for example:

- the watercourse concerned,

- weather,

- the commune.

This additional data is advantageously synchronized in a remote database.

In general, the system according to the invention thus makes it possible to obtain digital images in which the possible shadows cast around the subject are reduced (or even eliminated), whatever the ambient lighting; the phenomenon of drop shadows does not interfere in the automatic determination of the outline of the subject.

The outline of the subject is thus particularly clear and easy to determine in the digital images of interest by an automatic image processing process, without requiring any configuration on the part of the user.

The contour obtained can thus be analyzed to determine different characteristics of the subject studied.

Claims (10)

1. System for determining at least one characteristic relating to the outline of a subject (S) contained in at least one digital image of interest, which determination system comprises: a) at least one background structure (1) comprising a front face (11) having at least one retro-reflecting part (12) in front of which said subject is intended to be positioned, b) a photographic assembly (3) comprising: - image taking means (31) comprising a photographic lens (311), for taking digital images each containing a scene located opposite said photographic lens (311), - light means (32), adapted to provide artificial lighting of said scene and arranged so that light rays emitted by said light means (32) are capable of being reflected towards said photographic objective (311) by the retro part reflective (12) located opposite said photographic objective (311), and means (33) for coordinating the operation of said image taking means (31) and said light means (32), designed for capturing digital images containing said captured scene with or without exposure to said artificial lighting, and c) image processing means (4) designed to determine at least one characteristic relating to the contour of said subject (S) from at least one digital image of interest corresponding to a scene with exposure to said artificial lighting and in wherein said subject (S) is positioned in front of said retro-reflecting part (12). 2. System according to claim 1, characterized in that the processing means comprise a module for correcting deformation of the perspective of said at least one digital image of interest. 3. System according to claim 2, characterized in that the front face (11) of the background structure (1) comprises at least one geometric reference (13), and in that the correction module is designed to take account of said at least one geometric reference (13). 4. System according to any one of claims 1 to 3, characterized in that the processing means (4) comprise a module for the transformation of said at least one digital image of interest into a binary image composed: - first pixels (P1) originating from pixels (L1) of said at least one digital image of interest, corresponding to a visible portion (121) of the retro-reflecting part (12) which is located outside the space requirement said subject (S), and - second pixels (P2) originating from pixels (L2) of said at least one digital image of interest, other than said visible portion (121) of the retroreflective part (12). 5. System according to claim 4, characterized in that the transformation module cooperates with: - a sampling module for collecting samples of brightness values on the front face (11) contained in said at least one digital image of interest: at least a first sample collected on a visible portion (121) of the retro part -reflective (12) of said front face (11), and at least one second sample collected on a non-retro-reflecting part of said front face (11), and - a threshold calculation module, for determining, from said samples, at least one brightness threshold value used by the transformation module to distinguish said first and second pixels (P1, P2) from the binary image. 6. System according to any one of claims 1 to 5, characterized in that the processing means comprise: - a module for determining the outline of said subject (S) contained in said at least one digital image of interest, - a module for determining at least one of the dimensions of the outline of said subject (S), preferably its greatest length, and possibly - a module to identify said subject (S), for example its species in the case of an animal subject. 7. System according to any one of claims 1 to 6, characterized in that the processing means (4) comprise at least one analysis mode designed to determine at least one characteristic relating to the contour of said subject (S), chosen among: - a first mode of analysis starting from a single digital image of interest captured with exposure to said artificial lighting, - a second mode of analysis starting from a comparison of at least two digital images of interest, captured one with exposure to said artificial lighting and the other without exposure to said artificial lighting. 8. System according to any one of claims 1 to 7, characterized in that the processing means (4) comprise at least one computer program comprising program code means and intended to be executed by said system. 9. Method for determining at least one characteristic relating to the contour (C) of a subject (S) from at least one digital image of interest, by implementing a system according to the any one of claims 1 to 8, which method comprises the following steps: - preparation of the scene by positioning said subject (S) in front of the retro-reflecting part (12) of the front face (11) of said background structure (1), said scene being capable of presenting shadows (B ) which are projected onto said retro-reflecting part (12), - capture of at least one digital image of interest of said scene, with exposure to artificial lighting, and - processing of said at least one digital image of interest, to determine at least one characteristic relating to the outline (C) of said subject (S). 10. Background structure for taking images, which background structure (1) comprises a front face (11) comprising at least one retro-reflecting part (12) in front of which a subject is intended for be positioned, which front face (12) comprises at least one geometric reference (13) suitable for correcting perspective deformation. 1/3