GB2335977A

GB2335977A - Imaging device using polarimetry

Info

Publication number: GB2335977A
Application number: GB9822680A
Authority: GB
Inventors: Philippe Clemenceau; Sebastien Breugnot; Laurent Collot
Original assignee: Thomson CSF SA
Current assignee: Thales SA
Priority date: 1997-10-17
Filing date: 1998-10-16
Publication date: 1999-10-06
Anticipated expiration: 2018-10-16
Also published as: FR2769980B1; GB9822680D0; FR2769980A1; GB2335977B

Abstract

A polarimetric imaging device for encoding the image of an object in degrees of polarisation comprises means (21) for emitting two differently elliptically polarised beams (2,3) towards the object (4), means (8) for detecting the amount of each beam leaving the object (I1, I2), and signal processing means for determining the degree of polarisation of the object as the ratio Dp = (I1 -I2)/(I1 + I2). The emission means may comprise a laser diode (21), polariser (22), voltage controlled liquid crystal cell (23), and delay plate (24). The returning waves are collected by a telescope (29) and passed through an analyser formed by a delay plate (30) and a polariser (31) to a CCD camera (8). The device is applicable to active imaging eg for the recognition or monitoring of objects.

Description

1 1 2335977 1 "AN IMAGING DEVICW The present invention relates to an

imaging device using polanmetry. It can be applied especially to active imaging, for example in the context of the recognition or mondoring of objects.

In acfive imaging, a flow of photons is sent to k..he scene to be observed by means of a source, for example a laser or a lamp. One of the major advantages of this type of imaging lies in the choice available of several parameters of emission such as, for example, the wavelength, the frequency, the pulse frequency or pulse duration depending on the lo information desired in the scene. In Doppler imaging, the frequency of the radio broadcast wave is compared with that of the source in order to detect vibrations or movements of the scene. In imaging by telemetry, h is the travelling time of the wave sent by the source that is measured in order to encode the image as a function of the distance between the source and the scene.

There is another valuable parameter that makes it possible to improve the information contained in the image obtained. This information is the polarisation of the received wave. The knowledge of this characteristic provides information on the depolarisation induced by the scattering of fight from the different parts of the scene. This depolarisation depends in particular on the roughness of the different parts of the scene. Natural objects such as for example trees or sand create greater depolarisation than man-made objects. In particular, a specular reflection does not cause depolarisation. It is then possible to quantify this depolarising capacity by introducing the degree of polarisation so as to encode the image in terms of polarisation. As a result, an image by polarimetry may be obtained. An image of this kind makes it possible in particular to (( de-camouflage >> the objects of a scene whose albedo is identical to that of the background, i.e. objects that cannot be separated by the simple observation of their reflection coefficient but have a degree of polarisation different from that of the background of the scene. In other words, an image encoded in degrees of polarisation improves the distinguishing of the objects forming a scene.

In order to obtain perfect knowledge of the polarimetrical properties associated with the back-scattering of light from an object, it is necessary to measure a matrix known as the Mueller matrix. The term 2 back-scattering' herein designates the scattering of a wave by the object following the incidence of a.wave on the object. The Mueller matrix depends on the object and especially on the angle of incidence of the beam on the object, the angle of observation and the roughness of the object and its material. The reflection coefficient as well as the degree of polarisation, may be obtained from this matrix. However, to obtain the Mueller matrix which comprises 16 coefficients, it is necessary to carry out 16 measurements. This increases the complexity of an imaging system and hence its cost. Furthermore, the time needed for the computation of the image encoded in lo terms of degrees of polarisation is increased. This is a drawback for imaging at video rates. Furthermore, this large number of measurements gives rise to great uncertainty over the degree of polarisation since there is the additional uncertainty related to the 16 measurements.

The aim of the invention is to reduce the number of measurements needed for the encodinq of an imaae- in terms of dearees of polari-satim. To acoor to twentim there is pro an Imagmg device Wberein, with the image of an object being encoded in terms of degrees of polarisation, the device comprises at least means for the emission of at least two elliptically polarised waves towards the object, a detector of the values of intensity of the waves back-scattered by the object and signal-processing means obtaining the ratio (11 - 12) 1 (11 + 12) where 11, 12 are the detected values of intensity corresponding respectively to the first emitted wave and the second emitted wave, the ratio being the degree of polarisation of the object. 25 The main advantages of the invention are that it enables high spee.d in obtaining an image encoded by polarimetry and improved image contrast, is simple to implement, and is economical. Other features and advantages of the invention shall appear from the following description made with reference to the appended drawings, of which:

- Figure 1 is a block diagram of an exemplary device according to the invention; - Figure 2 is a possible exemplary embodiment of a device according to the invention.

1 3 The Stokes-Mueller formulae enable a description of the state of polarisation of a monochromatic wave. These formulae are characterised by a vector S, known as a Stokes vector, consisting of parameters 1, Q, U and V defined hereinafter. If we consider a wave that is propagated along an axis Oz and a reference 0, x, y, such that the axes Ox and Oy are perpendicular to the axis Oz, the vector E representing the electromagnetic field may be broken down as follows:

Ex(t) = EOxfficos[ot + 3x(t)] (1) Ey(t) = EOyfficos[cot + &y(t)l (2) where Ex(t) and Ey(t) are respectively the components of the vector ú along the axes Ox and Oy.

co is the pulsation of the electromagnetic field and 8x(t) and 8y(t) are the phase shifts of the components Ex(t) and Ey(t). The components of the Stokes-Mueller vector S are defined by the following relationships:

I=E 2 +E 2 OX OY Q=E 2 -E 2 Ox OY U MoxE0Y COS8 V MOXE0Y sin 8 (3) (4) (5) (6) 3 = 8At) - 8y(t) being the phase difference between the two components Ex(t) and Ey(t) of the vector representing the electromagnetic field.

The components 1, 0, U, V form the parameters, known as the Stokes parameters, of the wave emitted towards the scene to be observed.

The Mueller matrix M of an object subjected to back-scattering makes it possible to express the Stokes vector Sr back-scattered as a function of the incident Stokes vector S, giving:

Sr = M.S (7) The degree of polarisation Dp(M) of the matrix M is defined according to the following relationship where mij represents an element of this 4A matrix:

4 i=3j=3 m 2 _ n0 1 E Y D 01 0 p (M)= 10. '= j=' 3.nz (8) DP(M) is a characteristic of the material of the object. Dp(M) = 0% means especially that the material totally depolarises the incident wave and that the back-scattered wave is totally depolarised, i.e. that the phases of the s components of the back-scattered field have become totally incoherent. One example of a totally depolarised wave is natural light.

Dp(M) = 100% signifies the fact that the back-scattered wave is polarised, namely that the phases of the back-scattered field are totally coherent.

Dp(M) < 100% signifies the fact that the back-scattered wave is partially polarised, namely that R is the sum of a depolarised light and an elliptic vibration. In fact, Dp(M) represents the polarised light content of the back-scattered wave.

The obtaining of the Mueller matrix of an object therefore provides perfect knowledge of the polarimetrical properties of the wave backscattered from this object and therefore enables it to be encoded in degrees of polarisation and possibly in terms of reflection coefficient. As indicated here above, sixteen measurements are needed to obtain this matrix. The invention makes it possible to reduce the number of measurements needed for the encodin.q from sixteen to two.

Fi 1 gives a view, in a block diagm, of a v edx the invention. it comprises at least means 1 1or the emission of two waves 2, 3 towards a scene, more particularly towards an object 4 whose image has to be obtained. The choice of the ainitted wave is described hereinafter.

The device furthermore comprises a detector 8 of the values of intensity of the wave back-scattered by the object 4 and sig na]-processing means 9 used to obtain the following ratio Dp:

Dp = (11 - 12) 1 (11 + 12) i where 11, 12 are the detected values of intensity corresponding respectively to the first wave 2 and the second wave 3, the ratio Dp being the degree of polarisation of the object as can be seen in the rest of the description. The precision of the degree of polarisation may, for example, be refined by playing on the use of the reception means 5 and particular means 10 for the analysis of the back-scattered waves.

Hereinafter, a Stokes vector, as defined according to the 5 relationships (3) to (6), will be likened to its associated wave Ea EOY. There could therefore be a question of the emission or reception of a Stokes vector.

By the invention, the Mueller matrices of the objects forming the scenes of the day-to-day environment are considered to be diagonals. For example, the following are Mueller matrices Mp and Mb lo corresponding respectively to paint and to concrete:

MP = 1 -0.03 -0.04 -0.02 1 -0.01 0.02 0.01 -0.05 0.65 0.03 0 -0.02 0.18 0.03 -0.01 -0.06 0 0.63 0 -0.01 -0.04 0.17 0 -0.02 0.03 -0.01 0.6 0.03 0.02 0 0.12 These matrices have already been defined by methods with sixtyfour or sixteen measurements.

The non-diagonal elements of the matrices have values much smaller than those of the diagonal elements. They ar-e substantially equal to the uncertainty and may be interpreted as measurement noise. The Mueller matrix M is therefore considered, for a common object, to have the following form:

M = moo 0 0 0 0 MI 1 0 0 0 M22 0 0 0 M33 - 6 Its degree of polarisation D(p) is then the following:

DP = (M1 'I + M22 + M33)13M00 (9) Furthermore, if we make the additional assumption that the object has a circular symmetry, namely if we consider the Mueller matrix to be the same if the object is rotated about the axis defined by the incident beam, perpendicularly to the components E, and EY of the vector representing the lo electromagnetic field, it turns out that this matrix M can be switched over with any matrix of rotation about the above-mentioned axis. It follows then that only three diagonal elements are free, i.e. that two are equal. These three elements are:

Moo which is the coefficient of reflection of the target object; IVI,, = M22 which is the coefficient of linear depolarisation; M33 which is the circular coefficient of depolarisation.

The matrix M then has the following form:

MOO 0 0 0 M, 0 0 0 M,, 0 0 0 0 M33 M= and its degree of polarisation DP is given by the following relationship:

DP = (2M, 1 + M33) 1 3M00 (10) The emission means 1 comprise an active element enabling the emission of two carefully chosen Stokes vectors. For this purpose, the emission means send, for example successively, two well-chosen elliptical waves 2, 3 towards the object 4. For example, the two Stokes vectors emitted are the following, multiplied by a scalar factor:

7 S, = 1 1/V53 1 / -J1 / -3) and S2 1 / -J3 1 / J5) Uie d that ed the invmtim emp rec and means 10 for the analysis of a wave 6, 7 back-scattered by the object 5 4. The two Stokes vectors emerging from the object 4 are:

S1r = moo mi 11 M221,F3 M33 I-J-30 00 and S2r M, 1 _M22 / r3 -M33 / J3) The means of analysis 10 are for example obtained in such a way that the intensity 11 and 12 coming out of this system are actually the scalar lo products of the following vectors Vs by these Stokes vectors S1r and S2r:

Vs = 0,5 J3- ibe device that embodLies the m ccaprises a detector 8 collecting the intensities 11 and 12 mentioned here above:

I, = 0,5(M00 + 1 / 3(M1 1 + M22 + M33)) (11) 12 = 0,5(M00 1 / 3 (M, 1 + M22 + M33)) (12) The device that embodies the invention comprises signal processing means 9 used to encode the image of the object 4 in intensity 1 and in degrees of polarisation Dp:

8 1 = MOO = 11 + 12 (13) Dp = (M,, + M22 + M33)13M00 = (11 - 12) 1(11 + 12) 114) For the encoding in degrees of polarisation, the factor 0.5 may for example be eliminated from the vector Vs. This factor does not come into play in the encodina in dearees of oolarisation Do.

lbw, it can clewly be wm that a deywe wbich ees the invention provides for an encoding of an image in intensity and in degrees of lo polarisation with only two measurements of values of intensity.

Figure 2 gives an exemplary view of a possible embodiment of a device according to the invention. The emission means 1 comprise, for example, in the direction OZ of propagation of light, respectively a laser diode 21, a polariser 22, a voltage-controlled liquid crystal cell 23 and a delay plate 24. The polariser is, for example, a vertical axis polariser. The laser source is for example a diode with maximum power of 100 mW at 830 nm and vertical polarisation. However it may be replaced by any other type of light source, for example a monochromatic source. A system with two mirrors 25, 26 is used to orient the emitted beam and merge the axis of emission 27 with the axis 28 of a telescope 29 described hereinafter. The active element is the liquid crystal cell 23. This cell is controlled for example by a voltage with a frequency 100 Hz used to send the delay plate 24 a vertical or horizontal rectilinear polarisation of the wave given by the laser source 21 depending on the level of voltage, for example +5 V or -5 V. The delay plate 24 is for example a lambda 18 plate whose fast axis forms an angle of -27.835 with the above-defined axis Oy, the vertical axis, the positive direction chosen being the trigonometric direction.

The reception means 5 comprise for example a telescope 29, for example of the Newton type. The means of analysis comprise for example a delay plate 30 and a polariser 31. The detector 8 is for example a chargecoupled detector 32. The image of the object on the detector 8 is formed by means of the telescope 29. The first mirror 25, along the direction of propagation, is for example parabolic and opened to a quarter of its focal distance f/4, this focal distance being for example 610 nm. The second mirror 26 is for example a plane elliptical mirror with an equivalent diameter 9 of 50 mm. The system of analysis comprises the delay plate 30 located after the telescope, in the direction of propagation. This plate 30 is for example a lambda 18 plate whose fast axis forms an angle of 62.615 with the abovementioned axis Oy. The axis of the polariser 31 which follows the plate 30 is 5 vertical.

After the first polariser 22, that of the transition means, the polarisation being rectilinear and vertical, the corresponding Stokes vector is the following, multiplied by a scalar factor:

Sin = 1 -1 0 W) since EO,. = 0, the vertical direction being the direction of the axis Oy.

After the liquid crystal cell 23, which acts as a lambda 12 plate, if a voltage is applied to it, the two following vectors are obtained according to the voltage level applied to this cell:

Sini = 1 -1 0 W) and Sin2 = (1" 1 0 0) The vectors Sin, and Sin2 correspond respectively to a vertical or a horizontal polarisation carded out by the Uquid crystal cell 23. These two vectors are different from the above vectors S, and S2. In fact, they differ from these vectors only by a rotation of 62.6150, especially in order to have vertical and horizontal polarisations and thus facilitate the implementation.

After the first delay plate 24, that of the emission means, two vectors V, and V2, which are respectively images of the vectors Sin, and Sin2, are obtained. Since Mplatel is the Mueller matrix of the first plate 24, V, and V2 are obtained by the following relationships:

V, = Mplatel-Sinl (15) and V2 = Mplatel.Sin2 (16) Namely, given the matrix Mplatel:

('I (1 1% OY805 -0,138 -0,578, V, = -0,805 0,138,0,578, and V2 = These vectors are back-scattered by the object 4 to give the following vectors:

00 -0.9805M, 1 09138M22 W,578M33 1 Vir = and V2, = 07805 -0,138 -0,578) Mplate2 andMpolar2 being respectively the Mueller matrices of the second plate 30 and the second polariser 31, namely those of the means of analysis, the matrix W of the means of analysis is such that W = mpoiar2. Mplate2 The Stokes vectors Scl, and Sc12 of the waves taken at the detector 8 are then given by the following relationships:

SO = W.V1r (17) andSd2 = W.V2, (18) The intensities 11, 12 reaching the detector 8 are equal to the scalar product of the first line of the matrix W with the vectors Scil, Sc12 at output of the means of analysis 30, 31. The first line of the matrix W of the system of analysis being (0.5; -0.402; 0.065; 0.289), we get:

I, = 0,5M00 + M24M1 1 + 01009M22 + 0,167M33 (19) 12= 0,5M00 - (0,324M1 1 + 0,009M22 + 0,167M33) (20) We therefore find:

11 MOO -- I] + 12 and (I, - 12) / (I, + 12) = (0,648M, 1 + 0,0 1 8M22 -;0,334M33) / MOO In taking M,, = M22, we get:

(11 - 12) 1 (11 + 12) = (2M 11 + M33) 1 3Moo) According to the relationship (10): (11 - 12) 1 (11 + 12) = Dp which is the degree of polarisation.

M,, = M7> expresses. actually the fact that. bv rotation of the system of the device that ees the inventicn, bewe of the reference OXY, 1 the Mueller matrix of the object 4 remains unchanged. This ensures especially that the same degree of polarisation Dp is measured whatever the rotation about the axis of propagation 27 of the beam sent to the object. The encoding in intensity may be equal to the sum of the intensities detected 11 + 12, for example multiplied by a factor.

The signal processing means 9 are for example based on microprocessors and associated memories. They are contained for example in a computer 32 connected to the detector 8. The screen of the computer shows for example an image encoded in degrees of polarisation andlor an image encoded in intensitv, nameiv a displav of the encQdinq in deqrees of poLlrLsatim andlor in intensitj. A de Aiich ees the invention reTaires at least two measurements of intensity. It could nevertheless be envisaged with additional measurements in case of need.

In this exemplary embodiment, the image is encoded in intensitv, =i in dec- of poltian. It is possible to provide for devices ad pe invention wnere ine image is encoded only in degrees of polarisation.

The advantage of the invention especially is that it improves the contrast of an image obtained by polarimetry. Indeed, for the encoding in degrees of polarisation, an encoding in sixteen standard measurements has greater uncertainty than an encoding in two measurements as in the case of the invention. The large number of linear combinations performed on the 12 sixteen intensities measured increases the uncertainty concerning the degree of polarisation Dp. Furthermore, an encoding that requires only two measurements enables greater speed of image processing and m4-r.,/ thus enable a video rate to be followed. The invention is also simple to implement inasmuch as it does not require any complicated assembling operations. As a result of the above advantages, it is clearly economical.

13

Claims

I. Aa -imaging device wherein, with the ima-e of an object being encoded in terms of degrees of polarisation, the device comprises at!east means for the emission of at least two elliptically polarised waves towards the object a detector of the values of intensity of the waves back-scattered by the object and signal processing means obtaining the ratio (11 - 12) 1 (11 + 12) where 11, 12 are the detected values of intensity corresponding lo respectively to the first emitted wave and the second emitted wave the ratio being the degree of polarisation of the object, the Stokes vectors associated with the emitted waves being substantially the following, multiplied by a scalar factor and a matrix of rotation about the axis (0z) of propagation of the waves:

S, = 1 e'i 1/V53 A/V53 and S2 /.,F3 J5) 2. A device accordinta to claim 1, which comprises means for the reception and analysis of the back-scattered waves, made in such a way that the values of intensity 11 12 taken into account by the detector are actually the scalar products of the following vector Vs with the Stokes vectors (S1r, S2r) of the back-scattered waves:

Vs = -JS multiplied by a scalar factor 3. A device according. to any of the preceding claims, wherein the two Stokes vectors are substantially:

1 i 14 C Sini and S,2 0 o 0) 4. A device accordin., to any of the preceding claims, wherein the emission means comprise at least one vertical polarisation light source, one vertical axis polariser and one liquid crystal cell that is voltage controlled to send a vertical or horizontal rectilinear polarisation of the wave given by the light source according to the level of voltage.

5. A device accordinis to claim 4, wherein the emission means comprise a lambda/8 C delay plate whose fast axis forms an lo angle of substantially -27.835 with the vertical axis (0y), the positive direction bein g #ie trigonometric direction.

6. A aevice according to any of the preceding- claims, which comprises means for the reception of the waves back-scattered by the object., these means comprising a telescope, to form the is image of the object on the detector., a system of two mirrors enabling the beam emitted by the emission means to be oriented so as to merge as axi!$ of emissior with the axis of the telescope.

7. A device according to claim 6, wherein the telescope is of the Newton type.

8. A device according to either of the claims 6 or 7, wherein the first mirror along the direction of propagation, is a parabolic mirror.

9. A device according to any of the claims 6 to 8. wherein the second mirror is a plane elliptical mirror.

10. A device accordi-,g to claims 5, which comprises means ot analysis connected between the reception means and the detector of intensities, the means of analysis comphsing a lambda/8 type delay plate whose fast axis forms an angle of substantially -27.8350 with the vertical axis (Ov), the positive direction beinq the trigonometric direction.

11. A device according to any of the preceding claims, which comprises means of analysis connected between the reception means and the detector of intensity analysis comprisina a vertical axis Polariser the means of 12. A"device accordic& to any of the preceding claims, which comprises a codifiat.ion in intensity.

13. A device according to claim 12. wherein the codification in intensity is given by the sum of the two intensities 11 + 12 detected by the deteclor, muffiplied by a factor.

14. A device according to any of the Rreceding claiffis, Yherein the signal processing m"ns are contained in a le computer.

15. A device according to claim 14, wherein the screen of the computer has an image of the object encoded in polarisation andlor in intensfty.

16. An imaging device substantially as described hereinbefore with reference to and as illustrated in Figure 1 or Figure 2 of the accompanying drawings.