DEVICE FOR IDENTIFYING THE NATURE OF A LIQUID FROM A DROP THEREOF

Abstract

The invention relates to a system for identifying a liquid from data of an image of a drop (10) of this liquid on a reflective surface (12), comprising: a) means for estimating, from said image data, the contact angle and the capillary length of the drop placed on this surface; b) means for comparing the contact angle and capillary length estimations with known data of contact angles and capillary lengths of a drop of another liquid and for identifying the nature of the liquid of the drop to be identified. figure 4

Description

Title of the invention: DEVICE FOR IDENTIFYING THE NATURE OF A LIQUID FROM A DROP THEREOF TECHNICAL FIELD AND PRIOR ART

[0001] The invention relates to a new device and a new method for detecting the nature of the liquid which constitutes a drop.

[0002] Methods are known that implement RFID readers, photodiode, broadband units or piezoelectric sensors, but these are generally not accessible to individuals because they require equipment that is most often only found in an analysis laboratory.

[0003] A system, described by S.Yue et al. "Liquid testing with your Smartphone", in Proceedings of the 17th annual International Conference on Mobile Systems Applications and Services, ser.MobiSys' 19, NY, USA; Association for Computing Machinery 2019, page 275 - 286, implements capillary waves generated by the vibromotor of a Smartphone placed above a container containing the liquid. Other mobile applications implement a vibromotor and an accelerometer or a determination of viscosity or the transmission of acoustic waves.

[0004] But, in general, these systems require the implementation of a substantial quantity of liquid, for example the equivalent of at least one cup or even a minimum of several hundred milliliters, for example about 400 ml. This is obviously impossible for certain liquids, in particular bodily fluids, for example in the case of individuals suffering from oligoanuria or another pathology or in the case of blood.

[0005] Another solution, described in J. Chan et al. in "Testing a drop of liquid using Smartphone lidar", Proc. ACM Interact. Mob. Wearable Ubiquitous Technol., vol. 6, no. 1, March 2022, implements a Smartphone. But, increasingly, Smartphone manufacturers are considering eliminating LiDAR sensors from these devices.

[0006] The problem therefore arises of finding a new device and a new method which can implement detection of the nature of a liquid, on the basis of a small quantity or a small volume thereof, for example a volume much less than 100 ml, for example still between 50 μl and 500 μl.

[0007] There is also the problem of finding a new device and a new method which can be implemented by a small device, for example a Smartphone, so that they can be used by natural persons, without having, or before resorting to an analysis laboratory.

[0008] Another problem is that of easier detection of liquids that are counterfeit or adulterated, for example containing water in an unacceptable proportion and / or containing poor quality alcohol. Detection of this type of liquid, based on small quantities, would allow easier and faster testing. Statement of the invention

[0009] The invention firstly relates to a device or system for identifying a liquid from data of an image of a drop of this liquid on a reflective surface, comprising:

[0010] a) means, or programmed means, for estimating, from said image data, and possibly from the image of the drop reflected on said surface, the contact angle and the capillary length of the drop placed on this surface;

[0011] b) means, or programmed means, for identifying the nature of the liquid of the drop to be identified on the basis of the contact angle and capillary length estimates.

[0012] The means for estimating the contact angle and the capillary length of the drop may comprise:

[0013] - means or programmed means for extracting the outline of the image from the drop ;

[0014] - and / or filtering means, or programmed means for filtering, said image data and / or means, or programmed means, of optimization, for example of the Nelder-Mead type;

[0015] - and / or means, or programmed means, for estimating the contact line between the drop of liquid to be identified and said surface; for example, the contact line can be estimated using the image of the drop of liquid to be identified and possibly its image reflected on said surface.

[0016] A device or system according to the invention may comprise:

[0017] - means, or programmed means, for generating information relating to at least one exposure condition of the drop of liquid to be identified, for example a background of the drop and / or at least one lighting condition and / or at least one orientation of a light source and / or an inclination of a surface of a display device relative to a reference direction;

[0018] - and / or means, or programmed means, for estimating the height (H) of the drop of the liquid to be identified and / or the radius of curvature (b) of the drop at its apex; an estimate of the height can be obtained by maximum likelihood or by consensus, or by a consensus algorithm, based on random sampling.

[0019] A device or system according to the invention may comprise:

[0020] - means, or programmed means, for generating at least one reference mark (Ymin, Y max) delimiting an area of ​​the drop image from which image data can be or is selected;

[0021] - and / or at least one camera.

[0022] In a device or system according to the invention, the means for identifying the nature of the liquid of the drop to be identified on the basis of the contact angle and capillary length estimates may comprise:

[0023] - means, or programmed means, for comparing the angle estimates of contact and capillary length with known data of contact angles and capillary lengths of a drop of another liquid;

[0024] - and / or means, or programmed means, of classification or for classifying, implementing a classification algorithm, for example of the K nearest neighbors type.

[0025] A device or system according to the invention may further comprise programmed means for generating a relative capillary length of the drop.

[0026] A device or system according to the invention may further comprise at least one camera for generating said image data.

[0027] A device or system according to the invention may comprise a wireless telephone, or Smartphone, which comprises at least means a) and b).

[0028] A device or system according to the invention may further comprise a reflective surface on which the drop can be positioned, for example the screen of a cordless telephone.

[0029] The invention also relates to a method for identifying a liquid from image data of a drop of this liquid placed on a reflective surface, implementing a device as described above.

[0030] The invention also relates to a method for identifying a liquid from image data of a drop of this liquid placed on a reflective surface, comprising the following steps:

[0031] a) esthner, from said image data, the contact angle and the capillary length of the drop placed on this surface;

[0032] b) identifying the nature of the liquid to be identified on the basis of the contact angle and capillary length estimates, for example comparing the contact angle and capillary length estimates with known data of contact angles and capillary lengths of a drop of another liquid and identifying the nature of the liquid of the drop to be identified.

[0033] Such a method may comprise:

[0034] - an extraction of the contour of the image of the drop;

[0035] - and / or filtering of said image data and / or optimization means, by example of Nelder-Mead type.

[0036] In a method according to the invention, it is possible to estimate or calculate data of a contact line between the drop of liquid to be identified and said surface.

[0037] In a method or device according to the invention, the contact line can be estimated from the contact points of the drop on said surface, these points being for example obtained or estimated using the tangent to the drop and possibly to its reflected image on said surface; the contact points of the drop on said surface can be obtained or estimated:

[0038] - as being the points for which the tangent to the drop of liquid to be identified, and possibly to its image reflected on said surface, is substantially perpendicular to said surface;

[0039] - from, or among, 3 points located on one side of the drop of liquid to be identified, and possibly of its image in said surface, and 3 points located on the other side of this drop and possibly of said image.

[0040] In a method or a device according to the invention, the contact line can be estimated using a sign change of the tangent to the liquid drop to be identified and its reflected image on said surface.

[0041] In a method according to the invention, information can be generated, relating to at least one exposure condition of the liquid drop to be identified, for example a background of the drop and / or to at least one lighting condition and / or to at least one orientation of a light source and / or to an inclination of a surface of a visualization device relative to a reference direction.

[0042] A method according to the invention may comprise:

[0043] - an estimation of the height (H) of the liquid drop to be identified (for example by consensus based on random sampling) and / or of the radius of curvature (b) of the drop at its apex;

[0044] - and / or the generation of at least one marker (Ymin, Ymax) delimiting an area of ​​the image of the drop in which image data is selected.

[0045] In a method according to the invention, step b) may comprise:

[0046] - a comparison of contact angle and capillary length estimates with known data of contact angles and capillary lengths of a drop of another liquid;

[0047] - and / or a classification by a classification algorithm, for example of the K type closest neighbors.

[0048] In a method according to the invention, the drop to be identified is for example a drop of an alcoholic beverage, or of urine or of water or of oil or of honey.

[0049] The drop may have, for example, a volume of between 50 and 500 μl.

[0050] In a method according to the invention, the reflective surface may be the screen of a cordless telephone or Smartphone and / or steps a) and b) may be implemented by a cordless telephone or Smartphone.

[0051] The invention also relates to a computer program, comprising the instructions for implementing a method according to the invention, as described above or in the remainder of the present application.

[0052] The invention can therefore be implemented as a health application in a mobile phone (Smatphone) or a PC peripheral by exploiting the camera of a mobile phone or a “webcam” of a computer or a microcomputer or an electronic tablet. Brief description of the drawings

[0053] - [Fig.l] represents a drop of ethanol (on the left) and water (on the right);

[0054] - [Fig.2] is a representation of an axisymmetric drop on a non-wetting surface;

[0055] -[Fig.3A] and [Fig.3B] represent device processing modules according to a rea lization of the invention,

[0056] - [Fig.4] represents droplets on the screen of a smartphone;

[0057] - Figures 5A and 5B represent operating modes of an embodiment of a device according to the invention;

[0058] - Figures 6A-6B represent 2 water drops with different backgrounds,

[0059] - [Fig.7] represents a water drop on the screen of a Smartphone, and the line of contact of this drop on the screen,

[0060] - [Fig.8A], [Fig.8B] and [Fig.8C] represent different examples of contact points;

[0061] - [Fig.9] represents a difficult case of identification of a contact point;

[0062] - [Fig.10A] represents a drop of milk (on the left) and a drop of water (on the right);

[0063] - [Fig.lOB] represents an image of the points of the contour of a drop and its image, this image including noise in the form of parasitic points,

[0064] - Figures 11 A, 1 IB and 1 IC represent different droplet profiles, disturbed by parasitic data;

[0065] - Figures 12A and 12B show contact angle measurement results of different products ([Fig.l2A]) and different liquids having different alcohol contents ([Fig.l2B]);

[0066] - [Fig. 13] represents the dispersion of the characteristics measured by a method according to the invention, for different liquids,

[0067] - [Fig.l4A], [Fig.l4B] and [Fig.l4C] represent various examples of classification of liquids,

[0068] - Figures 15A and 15B represent results of detection of counterfeits of alcoholic beverages (rum in [Fig.l5A], vodka in [Fig.l5B]);

[0069] - Figures 16A and 16B show urine test results, on the one hand angle contact as a function of the protein level in the urine ([Fig.16A]), on the other hand the use of the capillary length and the contact angle ([Fig.16B]).

[0070] DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS

[0071] One aspect of the invention will be explained in connection with Figures 1 and 2.

[0072] Liquids can be differentiated on the basis of their wettability, i.e. their ability to spread and form a uniform film on a surface.

[0073] This characteristic can be measured by the surface tension and polarities of the different liquids (see, for example, the article by DE Sullivan: "Surface tension and contact angle of a liquid-solid interface", the Journal of Chemical Physics, vol. 74, no. 4, pages 2604-2615, 1981).

[0074] For example, water, which has a fairly high surface tension, tends to form (as illustrated in [Fig.4]) droplets or beads 10 on a hydrophobic surface 12, such as tempered glass, wax or oil. [Fig.l] represents 2 drops 2, 4 on a hydrophobic surface 6: a drop 2 of ethanol (on the left), a drop 4 of water (on the right), the latter having a lower surface tension than ethanol and which thus wets the hydrophobic surface 6 less easily.

[0075] The contact angle 0, shown in [Fig.2] for a drop 10, is an indicator of the wettability of a liquid, which facilitates its identification. This angle is defined between the solid surface 12 on which a drop of the liquid rests and the tangent to the drop at the point where the liquid and the solid meet. A liquid that has a high surface tension has a larger contact angle than a liquid that has a low surface tension ([Fig.l]).

[0076] Young's equation allows us to establish a relationship between the contact angle and the tension at the interface of the system:

[0077] [Math.l] cos(0) = (ySV - ySL) / YLV

[0078] Where Ysv, respectively ySL and yLV represent the surface tensions of solid - vapor, respectively solid - liquid and liquid - vapor.

[0079] So if, for example, different liquids are in the air and rest on the same solid, the contact angle becomes a direct indicator of the surface tension of the liquid yLV. Subsequently, the surface tension is designated by y

[0080] Capillary length, on the other hand, is a length that links the force of gravity and surface tension. For a liquid, a short capillary length indicates that a small drop volume is sufficient for the force of gravity to compensate for the capillary forces: this results in a flattened shape of the drop. Conversely, a long capillary length means a large drop volume for it to deviate from a shape spherical and becomes flattened. The capillary length a is defined by:

[0081] [Math.2] a =

[0082] where y is the surface tension of the liquid, p its density and g the acceleration due to gravity.

[0083] Therefore, the capillary length provides additional information about the liquid.

[0084] An example of a device according to the invention is illustrated in [Fig.3A].

[0085] It comprises means (for example a camera) for recording image data of a drop 10, deposited on a surface, preferably reflective 12, of a liquid to be identified or measured and / or of a reference liquid deposited on this reflective surface. Preferred characteristics of this surface are indicated below.

[0086] The image data of the drop can be processed according to the present invention by electronic means, for example a processor or a microprocessor programmed for this purpose, and allowing:

[0087] - to calculate or estimate the contact angle and the capillary length of the drop;

[0088] - to identify the liquid of the drop, for example among a plurality of other liquids possible or known, for example further using means for comparing the contact angle and capillary length estimation data to known capillary length contact angle data of such other liquids and / or classification means as described below.

[0089] It is possible to extract the capillary length and the contact angle from the drop shape.

[0090] For example, for a liquid drop of axisymmetric shape, as illustrated in [Fig.2], the pressure difference AP, at any point M of the air-liquid interface, can be expressed as a function of the surface tension y of the liquid and the radius of curvature of the drop at point M according to the following expression:

[0091] [Math.3] AP = y((l / rD + (1 / ^2))

[0092] Measuring riet r2 is not easy. At the top of the drop, on the other hand, ri= r2. Calling b the radius of curvature at the top, we can write the previous equation in the form:

[0093] [Math.4] APQ = 2y / b.

[0094] In other words, inside a drop 10 which is symmetrical with respect to a central vertical axis y, the pressure difference AP varies linearly with the elevation. Considering the point O of [Fig.2] as being the origin, AP can be expressed in all point as a function of gravity and the density disparity between the droplet and the surrounding medium. If Ap represents this density difference and y the distance from the origin point O, then we have:

[0095] [Math.5] AP - 2yfb + Ap.gy

[0096] Combining the above equations, we arrive at the Laplace - Young equation:

[0097] [Math.6] 2y[b + Ap.gy — yOJrï + l / r2)

[0098] By expressing the radii of curvature, in x, y coordinates (H being the height in the y direction), we arrive at the Bashforth - Adams equation:

[0099] [Math.7] y(y " ( + v '2) 3 / 2 + y 'x( 1 + y '2) 112 ) = 2y / b + Apgy

[0100] Where y' = dy / dx and y' '=dy2 / dx2. This differential equation describes the behavior of y(x) but has no analytical solution.

[0101] A simpler approach is to introduce the substitution y'(x) = tan (¢), which avoids complex manipulations related to the fact that y is not a simple function of x.

[0102] By parameterizing with the angle O and introducing the capillary length expressed above, the following set (7) of equations can be obtained, which only implements the Cartesian coordinates y and x with the 2 parameters a and b:

[0103] [Math.8] dx _ bxccAfi) d$ a^bxy+lx-bsin^)} dy bxsln(^) dij) cdbxy+2x-bsin{^>)

[0104] The shape of the drop 10 is defined by the set of points x(O) and y(O) which satisfy the above equation, as well as by the contact angle 0 between the drop and the surface. To obtain the theoretical profile of a drop for given values ​​of a and b, this equation can be integrated from the baseline (y=0) to the height y=H using for example a method of the “BDF” type (Backward Differential Formula), or differential backpropagation method, for example described in the article by G.D.Byrne et al. entitled “A polyalgorithm for the numerical solution of ordinary differential equation”, ACM Trans. Math. Softw, vol.l, no.l, pp. 71-96, March 1975.

[0105] In the above developments, b is the radius of curvature at the apex linked to the volume of the drop.

[0106] From an image of the drop, for example as recorded by a camera, for example again by that of a Smartphone, we can extract the profile of the edge of the

[0107]

[0108]

[0109]

[0110] [YES]

[0112]

[0113]

[0114]

[0115]

[0116] drop, which forms a set of points {[xi5 y J The parameters a and b can be obtained by minimizing the following expression (8): [Math.9] * v” II ' at b = a,b |[" i Where x'i, y'; represent the point on the drop profile that corresponds to equation (7) above, with parameters a and b. a* corresponds to the best approximation of the capillary length for the drop, measured in pixels; b* represents the radius of curvature at the vertex O and is related to the volume of the drop; the contact angle 0 is defined as the parametric angle <e>at which the vertical position is equal to the measured height of the drop, i.e. 0 / y(0) = H (see [Fig.2]). According to the invention, it is therefore possible, from data of an image of a drop 10 of a liquid (which data is obtained using a camera, for example a camera of a Smartphone 14, as illustrated in [Fig.5A]), in particular data of a profile of the drop, to calculate or estimate the contact line and to calculate or estimate the following characteristics of the liquid: the contact angle and the capillary length. As explained later, these characteristics can be transformed or processed so as to eliminate the impact of external factors, including in particular the magnification and / or the distance between the camera and the drop. As already explained, these contact angle and capillary length data make it possible to identify a liquid, for example from among a plurality of other possible or known liquids, for example again by comparing the contact angle and capillary length data with those of these other liquids. A device according to the invention or implementing the invention may therefore comprise extraction means, or module, which are programmed to, from image data of the drop: - estimate his profile; - estimate or calculate the contact line and contact angle, then the capillary length. Such a device may include means, described below, for processing the data so as to eliminate the impact of external factors, including in particular magnification and / or the distance between the camera and the drop. Other aspects of these data processing means are described further below. In the case of implementing the invention with a portable device, for example a Smartphone 14, problems may arise which are not encountered when using laboratory equipment, including a tripod and / or light source. of controlled intensity and direction.

[0117] Among these difficulties:

[0118] - the non-parallelism between the display screen (for example of a camera, for example that of a Smartphone) and the plane of the drop, which can lead to distortion effects of the apparent shape of the drop; the plane of the drop is the plane parallel to the display screen or to the camera of the telephone capturing the image of the drop (i.e. "seen from the front of the drop");

[0119] - and / or the identification of the outline of the drop, in particular in the vicinity of the points of contact, perhaps difficult due to uncontrolled factors such as variations in illumination and absorption of light by the liquid;

[0120] - and / or the contact line is clearly visible in the image when the camera is correctly positioned. aligned with both the surface and the drop; however, such alignment, without a fixed device, is very difficult to achieve;

[0121] - and / or the question also arises of the surface to be used for the acquisition of gout images.

[0122] To overcome this type of difficulty, a device or method according to the invention or for implementing the invention may comprise one or more of the following steps or means or modules:

[0123] (1) a step, or means, or a module, 20 for assisting in the acquisition of images: a user deposits a drop 10 on a surface (see for example figures 6A, 6B), for example the surface of the screen 12 of a mobile phone (or Smartphone); the image acquisition means allow an untrained user to obtain a front image 18, of sufficient quality, of the drop; the image thus acquired can be processed by the step or the means of extracting the profile of the drop;

[0124] (2) and / or a step, or means, or a module 22 for extracting a profile from the droplet for example implements a Canny filter (see, for example, J.Canny "A computational approach detection to edge detection", I3E Transactions on Pattern Analysis and Machine intelligence, Vol.PAMI-8, vol.6, p. 179-698, 1986), applied to the image converted to grayscale, in order to extract the edges of the droplet. These edges are then transformed into a set of pixels, set represented by Profile = {M(x,y)}, which represent the profile of the droplet;

[0125] (3) and / or a step, or means, or a module 24 for detecting a line of contact, described in more detail below;

[0126] (4) and / or a step, or means, or a module 26 for extracting the characteristics characteristics 28 of the liquid, already described above.

[0127] (5) and / or a step, or means, or a classification module, which is based on the characteristics of the liquid calculated or estimated; we then implement, for example, a KNN type algorithm (“K nearest neighbors”, or “K nearest neighbors”). neighbors ").

[0128] Preferably, a surface 12 on which a drop is placed for the implementation of the invention has characteristics of chemical inertness, homogeneity and rigidity. If the surface is too rough, measurements of the same contact angle of the same liquid may be different (the more damaged the surface, the more unstable the measurement of the angle is for the same liquid). Similarly, the presence of dust(s) and / or foreign element(s) on a surface that has not been chemically cleaned may affect the contact angle. Therefore, a flat or smooth (non-rough) surface that can be easily cleaned is preferably sought. Finally, the chosen surface preferably has sufficient hydrophobicity to allow the contact angle to be visible (>0) for most liquids.The surface 12 of the screen of a mobile phone, or Smartphone, satisfies these conditions: it is generally made of tempered glass, sufficiently hydrophobic, well suited to the measurement of contact angles. It can be easily cleaned. In addition, its smooth and hydrophobic nature ensures that the contact angles obtained are sufficiently high and can be measured satisfactorily.

[0129] The implementation of the invention may present a problem of maintaining the camera parallel to the plane of the drop (already defined above) during the capture of the image thereof. Indeed, perspective effects can distort the apparent shape of the drop, a user not necessarily being equipped with the material means to eliminate this type of effect, in particular if it is a natural person not having the resources available to an analysis laboratory. Furthermore, it is preferable that there is sufficient contrast between the outline of the drop and the surrounding contact points. Depending on the illumination and the light absorption properties of the liquid, a device or a method, or algorithm, according to the invention can make it possible to resolve the difficulties of identifying the contact points.

[0130] To solve these problems, a device and a method according to the invention can implement various solutions:

[0131] - assistance in taking images without perspective effect, by implementing means such as a gyroscope, for example that of a Smartphone 14 used; gyroscope-type means make it possible to measure the azimuth and / or roll and / or yaw angles, preferably in real time, thus being able to provide an indication 30 ([Fig.5A]) to the user on a correction to be made to obtain a vertical position of the viewing device used;

[0132] - and / or generate one or more markers on the contact zone, for example horizontal lines zontal such as the lines ymin, ymax represented in figures 3A and 5A, this(these) marker(s) indicating and / or delimiting a contact zone and allowing a user to po locate a drop for correct or improved detection of the contact line; this(these marker(s) make it possible to delimit the search area of ​​the contact line; in this case, preferably, a step or means, or module, for extracting contact lines, for example as described below, selects or is programmed to select points only in the area specified or identified by this(these marker(s);

[0133] - and / or means for providing a user with an indication concerning the orientation of a light source and / or the positioning of a drop in front of a white or dark background; Figures 6A and 6B illustrate the impact that a change of background can have, the contact line being generally more visible with a light or white background (as in [Fig.6B]) rather than a dark background (as in [Fig.6A]); the user will be able to modify the light source or the background if the algorithm fails to identify the contact line.

[0134] As regards the extraction of the contact line, according to one aspect of the invention, it is possible to implement the reflection of the drop 10 on the reflective surface 12 on which it is deposited. The points where the profile of the real drop meets that of the reflected drop represent the contact points. By connecting these points along a straight line, the contact line is identified.

[0135] A method, or algorithm, to enable this contact line to be detected can be based on 2 considerations:

[0136] - first of all, the surface 12 on which the drop is deposited is reflective: the side section of the profile of the drop 10 and its image on this surface shows a modification of the sign of the slope of the tangent at the point of contact with the reflecting surface;

[0137] - the change of direction (linked to the change of sign) of the tangent is approximately almost the same for the tangent to the contour of the real drop and for the tangent to the contour of its reflected image; therefore, to detect the point of contact, we can look for the position where the upper slope (of the real drop) and the lower slope (of the image in the reflecting surface) of the tangent to the curve (formed by the contour of the real drop and that of its reflected image), cancel each other out.

[0138] According to a method according to the invention, we search for the index n (the candidate point for the contact line) for which the points n - N and n + N (have slopes equal in absolute value but opposite in sign (N defines the window on which the tangent is calculated; for example we can evaluate the tangent on N = 10 pixels). For example, if we consider the left side of the drop and its image reflected in the surface, the whole being designated by PL, we can define PL(y) = xy for the values ​​of y between ymin and ymax.

[0139] A filter, for example a Savitsky-Golay filter, can be applied to smooth PL(y) and eliminate irregularities. Thus, for each point, the algorithm calculates the upper slope and the lower slope of PL(y) by fitting the lines that pass through the N points above and below the selected point (see lines 5 and 6 of algorithm 1 below). The difference between the 2 slopes is calculated or estimated, this difference being close to 0 at the point of contact (see line 8 of algorithm 1).

[0140] This process can be repeated for both sides of the profile, and the contact line is thus determined by connecting the two contact points.

[0141] Figures 8A, 8B and 8C show 3 different configurations of contact points, the contact angle being, respectively, greater than, less than or approximately equal to 90°. In these figures, as in [Fig.9], Lgi and Lg2 designate the horizontal guide lines, LP designating the profile of the contact plane.

[0142] When the contact angle is < 90° ([Fig.8A]), there is only one point (Pi in [Fig.8A]) which can be the contact point (this is also the case when the angle is very close to 90°); when the contact angle is greater than 90°, as illustrated in [Fig.8C], there are 3 points (Pb P2, P3 in [Fig.8C]) one of which can be the contact point; in this case, the latter is defined as being the one located between the other 2 (see algorithm 1, line 14).

[0143] In other words, the above method makes it possible to identify the point of contact in the case where there are 1 or 3 possible points.

[0144] When the number of possible contact points is different (equal to 2, 4, or more than 3, see for example the case of [Fig.9], with the 7 points Pi P7), a signal can be sent to the user (line 15 of algorithm 1) to indicate to him that the quality of the image is insufficient and invite him to remedy the situation.

[0145] Les étapes décrites ci-dessus peuvent être formulées sous la forme d’un algorithme 1:

[0146] [Math. 10] sfs.atst Side of the drop Fl = {[;<: the point to consîder Ovtput: The estimated contact point Foor nothtng if impossible 1 = h j for n '= K to M do s teft = ; A Sight 5 = / *Cowute the stope of the ftnear fit cf the deta points Left* / ; s / *Compute the sfcpe of the ^teer fit of the data points fl ight® / ; « rf = O then 9 / * tfÿe tr: a y f ■ nd the corstsct poi nt,* ia sf = 1 the» is 'eîurn capcrdsteP UK m if = 3 then « retum mnchdsteF Uk is retum L Algorithm 1

[0147] This algorithm can be coded in the microprocessor of a device for processing data from the camera, for example the microprocessor of a Smartphone.

[0148] More generally, one or more, or even all, of the different steps which are described in the context of the present application may be programmed in a device comprising electronic data processing means, for example a microprocessor; this is for example the microprocessor of a Smartphone, preferably the one with which one or more steps of the drop could have been taken.

[0149] We now describe how we can calculate the capillary length a and the contact angle 0.

[0150] For this, the input data are the drop profile data as well as the contact line. Using the latter, the height H of the drop is measured and the parameters (a, b) are sought which make it possible to adapt the measured profile. For each pair of parameters (a, b) the equation (7) above is integrated from y=0 to H and generates a set of points (x^y'i), these generated points can then be compared with the measured points of the drop profile: the parameters which make it possible to get closest to the points of the measured profile are kept.

[0151] If the illumination conditions are not suitable, the upper point of the drop may be determined inaccurately or with a significant error. In addition, points may be erroneously generated by the algorithm due to poor image quality, which leads to errors in the calculation of the contact angle.

[0152] To solve these problems, we introduce the coordinates O(x0,yo) of the apex point into equation (8) already presented above, which therefore becomes the following equation (9):

[0153] [Math. 11] fl*, b ", jcX v* = argmz«J^ min / aj}^^

[0154] Solving this equation makes it possible to estimate the characteristics a and b and, using y0, to robustly determine the maximum height of the drop.

[0155] To this end, an algorithm can be implemented that allows processing a set of points containing noise; for example, a RANSAC ("Random Sample Consensus") type algorithm can be implemented, as for example described in MA Fischler et al. "Random Sample Consensus: a paradigm for model fitting with applications to image analysis and automated cartography", Commun. ACM, vol. 24, no. 6, pages 381-395, June 1981. The set of points is randomly sub-sampled and equation (9) above is used to obtain the desired characteristics. The points that most closely adhere to the model obtained through the sub-sample, within a predefined margin of error, are selected as potential candidates. The process is repeated with different subsets (different subsamples), and the subset with the maximum number of points retained within the error margin is chosen to calculate the drop characteristics. As illustrated in Figures 11A-1 IC, on which are plotted points la, Ib, le obtained from an image and points lia, Ilb, Ile which result from smoothing according to a certain profile (the one that best fits the observed point):

[0156] - [Fig. 11 A] represents a profile with an apex O (which corresponds to the abscissa point 0) erroneous, leading to an error of 5.54° on the contact angle of a water droplet;

[0157] - [Fig. 1 IB] represents a profile with parasitic points, leading to an error of 12.33° on the same contact angle;

[0158] - [Fig. 1 IC] represents a profile with good smoothing, leading to a very reduced error of 0.05° on the same contact angle.

[0159] To solve equation (9) above, we can implement the following algorithm 2:

[0160] [Math. 12] Out-sct: The $ 2 3 for t. - 1 te ■as yTî '' brapPrs / StePaints: ? :? size (tasters} s fe s feda gc-ad me-det* / ; ssarsî, £ - ëq. 9 ustng as [; <ç ar-d peints           wîth ëq. 7 as / 4 due if scorer < scere the n li score = rsrsrsf ; i You are you ras = t urs-s: ; you retuï nf sa tweî. Algorithm 2

[0161] Generally speaking, in a method or device according to the invention, a 2nd drop can be used which serves as a reference, in order to overcome the fact that the results are obtained in pixels. If L is the liquid tested, and "ref" designates the reference liquid, for example water, the relative capillary length a' of the liquid tested, expressed as a percentage, can be calculated in the following manner:

[0162] [Math. 13] A' — 100* \aL-aref\ / aref

[0163] The reference data may be stored in a memory of a device according to the invention, for example the memory of a Smartphone, and this device comprises programmed means for carrying out a relative calculation such as that above.

[0164] In the context of an implementation on, for example, the Android system, the calculation of the characteristics of the drop from expression (9) above can be carried out using an optimization technique, for example of the Nelder-Mead type (J.A.Nelder and R.Mead: "Simplex Method function minimization", Comput.J, Vol.7, p.308-313, 1965).

[0165] The software can offer 2 operating modes:

[0166] - a so-called “scan” mode (illustrated in [Fig.5A]) which allows the camera to measure the characteristics of the drop; for example, the camera settings are configured for a resolution of 1080 x 1080, with an aspect ratio of 1;

[0167] - and a “surface” mode (illustrated in [Fig.5B]), which makes it possible to represent a shape circular on a screen in order to guide the user to position a drop according to its size (for example according to 4 different sizes: 50, 60, 70 and 80 pixels in radius, which corresponds approximately to drops of 50 to 160 pl in volume) (this mode can be used to indicate where to position the drop).

[0168] The invention can be implemented using 2 Smartphones:

[0169] - one used as a scanner and programmed to perform one of those processing described above (e.g. a “OnePlus8t” Smartphone);

[0170] - and, as a surface on which the drops are deposited, the screen of the 2nd Smartphone (e.g. a Google Pixel 4 Smartphone).

[0171] On the screen of the 2nd Smartphone, a drop of liquid to be measured and possibly a drop of a reference liquid are deposited, for example using a syringe, both of which can have standardized sizes.

[0172] The 2nd smartphone is positioned horizontally on a table. Using the smartphone scanner, a method according to the invention makes it possible to scan the 2 drops on a single image and to obtain the desired parameters (contact angle and relative capillary length). Following each measurement, the surface of the 2nd smartphone can be cleaned using a mixture of water and ethanol, followed by an air drying step.

[0173] An implementation of the invention has been described above with 2 Smartphones, which allows individuals to implement the invention, alone, outside the framework of a laboratory. But it is also possible to implement the invention in a laboratory, for example by using a camera and a reflective surface having the characteristics set out above. The data from the camera can then be supplied to a computer or a microcomputer programmed in order to process them according to a method according to the invention.

[0174] Data is presented below:

[0175] - test, allowing to evaluate the performances of a device and a method according to the invention;

[0176] - and application data of a method and a device according to the invention.

[0177] Test data:

[0178] - Several experiments were conducted with different liquids (cognac, wine Cruz red, ethanol, Gin, Red Label (a whisky), Merlot red wine, Grant whisky, Duval pastis, William Peel whisky, Négrita rum, Label scotch, Calvados Busnel (Sylvain), vodka, Poliakov vodka, urine and water; these different liquids have contact angles that range from 68° to 114°; the average values ​​measured, with the margins of error, are shown in [Fig.l2A]; the error on the contact angle is generally less than 2°, except for Cruz and Merlot, for which it reaches 4°, a difference which is attributed to their contact angles which are around 90°, which makes it difficult to define the contact line exactly; some liquids have very similar contact angles (around 75°) which is attributed to their alcohol content close to 40% (for Grant, William Peel whisky, Label, Rhum Négrita, Calvados Busnel (Sylvain)); but a clear differentiation can be made between these and Poliakov vodka, Duval pastis and Gin, since the alcohol contents of the latter vary significantly (37.5° for Poliakov vodka and Gin, 45° for Duval pastis); .

[0179] - [Fig.l2B] illustrates the relationship between the contact angle and the degree of alcohol in the liquid, showing the ability of a method or device according to the invention to provide data that accurately reflect the alcohol content. In particular, good sensitivity is noted in distinguishing between water (0° alcohol) and a 3.3° ethanol solution.

[0180] - [Fig. 13] represents the contact angle and capillary length data relative for 10 liquids, indicated in this figure (the liquids correspond to the following groups of points: Gi = ethanol, G2 = pastis (45°); G3 = olive oil; G4 = whiskey (40°); G5 = red wine; G6 = coffee; G7 = corn syrup; G8 = sugar syrup; G9 = beer, G80 = water). These data show that differentiating between olive oil and whiskey, based on the contact angle alone, is difficult, this angle being located in the range 75°-80° for both liquids. However, by taking into account the capillary length (on the y-axis), we can clearly differentiate between these two liquids. This observation is supported by the confusion matrices represented in figures 14A, 14B and 14C. Using the estimated contact angle and the capillary length, we can implement a KNN (K nearest neighbors, K=3) type algorithm to differentiate between these liquids.The confusion matrices show an average classification accuracy of 58%, 78% and 84%, respectively, using only capillary length and contact angle, for the 9 different liquid types (in [Fig. 13], water is not counted as it is the reference).

[0181] Application data:

[0182] - Alcohol counterfeit detection test: the invention makes it possible to differentiate between authentic alcohol and counterfeit or adulterated alcohol containing 30% water. Figures 15A and 15B show the contact angle and capillary length calculated by a method according to the invention for counterfeit rum and vodka (or adulterated rum and vodka), as well as for authentic products; these data show that a method and device according to the invention can be implemented to differentiate between authentic beverages and counterfeit or adulterated beverages;

[0183] - urine detection test:

[0184] Kidney disease is more common in individuals with diabetes and / or high blood pressure. When a kidney is damaged, certain substances are secreted that should not normally be present in the urine. Many of these substances reduce the surface tension of urine. This is the case, for example, with proteinuria, which is secreted when albumin passes from the kidney into the urine of diabetics.

[0185] The invention can help detect the level of albumin in the urine of patients suffering from diabetes, thus enabling upstream detection, which can be followed up by more detailed analyses.

[0186] For this type of application, a specific reflective surface can be used, instead of the screen of a Smartphone.

[0187] To test this application, different levels of egg albumin were introduced into healthy human urine samples and their characteristics were measured for different albumin concentrations:

[0188] - [Fig.16A] shows the evolution of the contact angle as a function of the albumin concentration (in mg / l). The data in this figure show that it is possible to detect different levels of albumin in urine using the contact angle of a urine drop.

[0189] - [Fig.16B] represents the contact angle and capillary length data relative on the one hand for healthy urine (point group Ui), on the other hand for urine with different levels of albumin (44 mg / l (point group U2), 100 mg / l (point group U3), 130 mg / l (point group U4)); as can be seen in this figure, it is possible to distinguish between healthy urine and urine with an albumin concentration as low as 44 mg / l.

[0190] The invention can also be applied to drops of other natures, for example to differentiate natural honey from synthetic honey.

[0191] An electronic device implementing a method according to the invention may comprise, in addition to an imaging device for collecting image data, data processing means, for example a processor programmed to implement one or more data processing steps according to the invention, and display means for viewing results of an analysis according to the invention. Such an device is preferably of the portable or mobile or wireless type.

[0192] It is also possible to implement a centralized application on a single server and which uses PCs and / or mobile devices as measurement terminals; in this case, the measurements can be carried out where the patient or sample is but they can be processed in a computing center, remotely from the patient or sample. < / e>

Claims

Claims

1. System (1) for identifying a liquid from data of an image of a drop of this liquid on a reflective surface, comprising: a) means for estimating, from said image data, the contact angle and the capillary length of the drop placed on this surface; b) means for comparing the estimates of contact angle and capillary length with known data of contact angles and capillary lengths of a drop of another liquid and for identifying the nature of the liquid of the drop to be identified.

2. System according to claim 1, comprising means for estimating the line of contact between the drop of liquid to be identified and said surface.

3. System according to claim 2, the contact line being estimated using the image of the drop of liquid to be identified and its image reflected on said surface.

4. System according to claim 3, the contact line being estimated: - using the tangent to the drop of liquid to be identified and to its image reflected on said surface at each of the lateral contact points; - or using the tangent to the drop of liquid to be identified and to its image reflected on said surface at 3 points located on one side of the drop of liquid to be identified and at 3 points located on the other side of this drop.

5. System according to claim 4, the contact line being estimated using a change of sign of the tangent to the drop of liquid to be identified and to its image its image reflected on said surface.

6. System according to one of claims 2 to 5, comprising means for generating information on modification of at least one exposure condition of the drop of liquid to be identified, for example a modification of a background of the drop and / or of at least one lighting condition and / or orientation of a light source and / or inclination of the surface.

7. System according to one of claims 2 to 5, comprising means for estimating the height (H) of the drop of liquid to be identified.

8. System according to claim 7, comprising means for implementing a height estimation by maximum likelihood or consensus on the basis of random sampling.

9. System according to one of claims 1 to 8, further comprising at minus one camera.

10. System according to one of claims 1 to 9, comprising a cordless telephone which comprises at least means a) and b).

11. System according to one of claims 1 to 10, the reflective surface being the screen of a cordless telephone.

12. Method (1) for identifying a liquid from image data of a drop of this liquid placed on a reflective surface, comprising the following steps: a) estimating, from said image data, the contact angle and the capillary length of the drop placed on this surface; b) comparing the contact angle and capillary length estimates with known data of contact angles and capillary lengths of a drop of another liquid and identifying the nature of the liquid of the drop to be identified.

13. A method according to claim 12, the drop to be identified being a drop of an alcoholic beverage, or urine or water or oil or honey.

14. A method according to claim 12 or 13, the reflective surface being the screen of a cordless telephone.

15. Method according to one of claims 12 to 14, steps a) and b) being implemented by a wireless telephone or Smartphone or an electronic tablet or a computer or a microcomputer or a server, for example in a data center.

16. Computer program, comprising instructions for implementing a method according to one of claims 12 or 13.

Images