A full-stokes polarization camera system

Abstract

The present invention discloses a polarization camera system that integrates a polarization sensor with a homogeneous dispersive retarder to detect and measure simultaneously the four components of the Stokes vector of polarized light. Unlike existing methods that require complex setups or expensive components, this approach uses a single sensor setup, combining the division of focal plane method with wavelength sensitivity inherent in many camera sensors. The system (1) comprises: a polarization sensor (2) having a polarizer array (2a) and a color channel filter (2b); a photodiode array (2c) configured to measure at least four intensities; a light entrance (4); a microprocessor (5); a first homogeneous dispersive retarder (3). The first homogeneous dispersive retarder is configured to introduce wavelength-dependent retardation δ(λ) to the incident polarized light, so that [δ(λi) ≠ δ(λj)]; and, the microprocessor is configured to calculate a complete Stokes vector of incident polarized light from the measured intensities.

Description

[0001]A FULL-STOKES POLARIZATION CAMERA SYSTEM DESCRIPTION FIELD AND OBJECT OF THE INVENTION The present invention discloses a full-Stokes polarization camera system that integrates a polarization sensor with a particular homogeneous dispersive retarder to detect and measure simultaneously the four components of the Stokes vector of polarized light. A retarder, in the context of the present invention, refers to a component that introduces a controlled delay or phase shift to linearly polarized light passing through it. This delay varies with the wavelength of light, making it dispersive. In the context of the present invention, homogeneous dispersive retarders are utilized, which means they introduce the same delay across its entire surface and vary the delay based on the wavelength of light. Unlike existing complex setups that require multiple cameras and / or micro-retarders arrays, this novel approach uses a homogenous retarder to obtain full information of the polarization state at multiple pixel locations taking advantage of the wavelength sensitivity that is inherent to the photodiodes of camera sensors. This method avoids the fabrication and alignment difficulties of micro-retarders arrays, or the need of dividing the beam into using multiple cameras working simultaneously. Polarization imaging has found diverse applications in fields such as remote sensing, material analysis, and biomedical imaging. BACKGROUND OF THE INVENTION The Stokes vector is a four-dimensional parameter that captures the total intensity, linear polarization, circular polarization, and ellipticity of polarized light. The ability to measure the full Stokes vector enables a more detailed analysis of polarization phenomena and expands the range of potential applications for polarization imaging. The Stokes vector can be described as I0, I90 represent the intensity of horizontal and vertical directions, I45, I135 represent the intensity of 45◦and 135◦directions, Il and Ir represent the intensity of light with left circular and right circular polarization (i.e. related to the direction of circular rotation of the lightelectromagnetic field). In a common camera, only the first term of the Stokes vector, denotedas Itot representing total intensity, can be directly measured. No polarization information ismeasured. In contrast, a commercial polarization camera enables the direct measurement of the firstthree terms of the Stokes vector (S0, S1, S2). This is achieved through the integration of alinear micropolarizer array, comprising various orientations (four orientations in Fig. 1),placed in front of each photodiode pixel, as depicted in Fig. 1. In this system, I0, I90, I45, I135are readable intensities. The fourth term about information of circular polarization (S3) ismissing in this commercial configuration, lacking the ability to measure the complete Stokesvector. Existing approaches to image the four components of the Stokes vector simultaneouslyusing cameras rely on complex and sometimes expensive setups involving multiplepolarization elements and / or detectors. Two main approaches can be distinguished:“Division of Amplitude Polarization Imaging” and “Division of focal plane”.The “Division of Amplitude Polarization Imaging” consists of using multiple cameras, eachwith its own polarization element, so that the incident light is divided into different amplitudechannels using beam splitters, instead of using a single camera sensor. In case polarizationcameras are used as detectors, two polarization cameras are needed, with at least one ofthem equipped with its own retarder, to capture the full Stokes vector. This is the mostextended method used in the literature. The main disadvantage is that two cameras needto be used and the sensors need to be combined or synchronized to gather comprehensivepolarization information in parallel. A prior art document disclosing this knowledge is:Xingzhou Tu, Oliver J. Spires, Xiaobo Tian, Neal Brock, Rongguang Liang, and Stanley Pau, "Division of amplitude RGB full-Stokes camera using micro-polarizer arrays," Opt. Express 25, 33160-33175 (2017).10.1364 / OE.25.033160 The “Division of focal plane” is based on commercially available polarization sensors. Thesetypes of sensors (Fig. 1) already use a division of focal plane method, but they are notcapable of measuring the whole Stokes vector. So far, two approaches have beensuggested in the state of the art to measure the full Stokes vector (none of them availablecommercially):^ With Microretarder Arrays. A microretarder array layer is fabricated and placed ontop of the polarizer array to extend its capability to measure all four Stokesparameters. Disadvantage: A microretarder array (made of a birefringent reactivemesogen or liquid crystal polymer) is difficult to fabricate, and to position and alignwith the array of photodiodes. A prior art document regarding this: Tu, Xingzhou &McEldowney, Scott & Zou, Yang & Smith, Matthew & Guido, Christopher & Brock, Neal & Miller, Sawyer & Jiang, Linan & Pau, Stanley. (2020). Division of focal plane RGB full-Stokes imaging polarimeter. Applied Optics. 59.10.1364 / AO.391027 andalso Myhre, G., Hsu, W.-L., Peinado, A., LaCasse, C., Brock, N., Chipman, R. A., & Pau, S. (2012). Liquid crystal polymer full-stokes division of focal plane polarimeter. In Optics Express (Vol. 20, Issue 25, p. 27393). Optica Publishing Group. https: / / doi.org / 10.1364 / oe.20.027393. The retarder array needs to be at least binary as shown in Patent “Miniaturized full-Stokes vector polarization imagingdevice based on binary digital coding birefringent Crystal”, CN108007575A. ^Polarization arrays based on metasurfaces. Metasurfaces sensitive to polarizationare designed to split and focus light to various pixels on an image sensor for different polarization bases. The main disadvantage is that a metasurface is verydifficult and costly to fabricate. Additionally, metasurfaces produce loses (leading tolow efficiency in light detection). A prior art documents regarding this knowledge arethe Patent US20220214219A1, entitled: “Metasurface Mask for Full-Stokes Divisionof Focal Plane Polarization of Cameras” and the Patent US11385104B2, entitled:“On-chip polarization detection and polarimetric imaging”. In addition to these approaches, lensless polarization cameras using a phase mask have been also recently proposed. However, this approach relies on computational power forimage reconstruction and besides being largely susceptible to imaging artifacts, it cannotprovide real-time imaging. A prior art document regarding this: Baek, N., Lee, Y., Kim, T., Jung, J., & Lee, S. A. (2022). Lensless polarization camera for single-shot full-Stokes imaging. In APL Photonics (Vol. 7, Issue 11). AIP Publishing. https: / / doi.org / 10.1063 / 5.0120465. It is desirable to have images of the S3parameter of the Stokes vector and not just S0to S2, because it makes the polarization camera much more informative. For example, in remote sensing applications S3 can be used to distinguish surfaces, since metals reflect light with elliptical polarization (S3≠0) while insulators, such as plastics, provides linear reflection(S3=0). The measurement of S3 is also very important for any imaging system that aims todetect tissues or materials with birefringence, since when light propagates through thesematerials it acquires an ellipticity (S3≠0). These birefringent materials are very diverse, fromstudying fibrillar biological tissues, detecting contaminating microplastics or checking thestress in Silicon substrates that are used in solar panels or for the manufacture of microchips. An added value of having the complete Stokes vector is that the degree of polarization(DoP) of the incident light can be evaluated. ^^^ = ^^ DESCRIPTION OF THE INVENTION The present invention overcomes the limitations of conventional polarization sensors by introducing a homogeneous dispersive retarder (waveplate) positioned before a polarization sensor, enabling the detection of diverse polarization states beyond linear polarization. The dispersive retarder introduces wavelength-dependent retardation, δ(λ), between two orthogonal components of linear polarization, allowing the sensor's color channels tocapture different polarization states, thereby measuring simultaneously all components ofthe Stokes vector of incident polarized light. Essentially this method combines the division of the focal plane method offered by the polarization sensor with the sensitivity to wavelength offered by many camera sensors (e.g. cameras with wavelength filters embedded in the sensor or color cameras). In systems using microretarder arrays, the complexity arises because these arrays most usually introduce multiple axis orientations (e.g., several azimuthal orientations θ) acrossthe pixels of the sensor. This requires precise fabrication, alignment and calibration of theretarder arrays, making the system more complex to manage. However, in the system ofthe present invention, this complication is avoided because it only requires a homogenous(i.e uniform) retarder with a single orientation (only one θ). The reason is that thehomogeneous dispersive retarder introduces a wavelength-dependent phase shift uniformlyacross the entire sensor, so there is no need for an array where different pixels havedifferent associated retarder axes.In essence, the system of the present invention takes advantage of the wavelengthsensitivity provided by the color channel filter. Since different wavelengths experience different retardations (i.e., δ(λi) ≠ δ(λj)), this allows the system to solve the full polarizationstate (the complete Stokes vector) without needing multiple orientations for the retarder orother retardance non-uniformities. Thus, the proposed system simplifies both hardwarecomplexity and calibration efforts by reducing the complexity of the retarder. This makes thesolution provided by the present invention more efficient, less error-prone, and easier toimplement than prior art solutions using microretarder arrays. The homogeneous dispersive retarder is actually the simplest and most economical type of retarder that exists. It can be obtained simply with a homogeneous sheet of a transparent crystal such as quartz or sapphire (polymer-based retarders would also be suitable). Thesize of the retarder only needs to be large enough to cover the entire sensor and itspositioning does not require any special alignment. The delay introduced by this sheet is^ where λ is the wavelength, Δn is the difference in refraction indices (ordinaryand extraordinary) of the material and d is the thickness of the sheet.The dispersive retarder is characterized by its ability to introduce wavelength-dependent phase delays, enabling the differentiation of polarization states across various color channels of the sensor. As incident light passes through the dispersive retarder, the different color channels measure the polarization states with varying retardations, allowing for the determination of the full Stokes vector.Thus, the present invention discloses a full-Stokes polarization camera system comprising:^ a polarization sensor, which in turn comprises: a polarizer array having at leasttwo orientations; a color channel filter configured to detect at least twowavelengths (λi, λj); and, a photodiode array configured to measure at least fourintensities, one intensity for each wavelength and orientation combination;^ a light entrance for incident polarized light;^ a microprocessor connected to the photodiode array;^ a first homogeneous dispersive retarder positioned between the polarizationsensor and the light entrance, and having an axis orientation -θ-;wherein the first homogeneous dispersive retarder is configured to introduce wavelength-dependent retardation δ(λ) to the incident polarized light, so that [δ(λi) ≠ δ(λj)]; and, wherein the microprocessor is configured to calculate the complete (i.e., the four components) Stokes vector of incident polarized light from the measured intensities. In a particular embodiment of the invention, the photodiode array is configured to measure eight intensities when the polarizer array has four orientations and the color channel filter is configured to detect two wavelengths (λi, λj). In a particular embodiment of the invention, the color channel filter is a Red-Green-Blue“RGB” sensor, so that when the polarizer array has four orientations, the color channel filteris configured to detect the wavelengths of the red, green and blue colors, the system is configured to provide color images.In a particular embodiment of the invention, when the polarizer array has at least twoorientations, the system further comprises a second homogeneous dispersive retarderconfigured to introduce wavelength-dependent retardation δ’(λ) so that [δ’(λi) ≠ δ’(λj)]. Thesecond homogeneous dispersive retarder is placed between the first homogeneous dispersive retarder and the polarization sensor, and having an axis orientation -θ’-, so that [δ(λi)= δ’(λi); δ(λj)= δ’(λj); θ≠θ’] or [δ(λi) ≠ δ’(λi); δ(λj) ≠ δ’(λj); θ≠θ’]. In this embodiment the photodiode array is configured to measure four intensities. In a particular embodiment of the invention, the first homogeneous dispersive retarder and / or the second homogeneous dispersive retarder provide a retardance difference from 1º to 179º degrees. Preferably, the first homogeneous dispersive retarder and / or thesecond homogeneous dispersive retarder provide a retardance difference of 90 degrees:|δ(λi) - δ(λj) | = 90o, which offers optimal conditioning of the system.In a particular embodiment of the invention, the microprocessor is further configured tocarry out a calibration procedure consisting of calculating the orientation of the retarder’saxis (θ, θ’) and the retardance (δ, δ’) at each wavelength (δ(λi), δ(λj)), for known polarization of the incident polarized light. In a particular embodiment of the invention, the microprocessor is further configured tocalculate the complete Stokes vector from the measured intensities.e.g.: where wherein I0, I90represent the intensity of horizontal and vertical directions, I45, I135represent the intensity of 45◦and 135◦directions, Iland Irrepresent the intensity of left circular and right circular light.New and advantageously, the present invention combines one or two homogeneousdispersive retarders with the color (or wavelength) sensitivity of a polarization sensor tocalculate the Stokes vector. The polarization sensor of the present invention at leastcomprises a polarizer array, a color channel filter and a photodiode in any of the embodiments shown in figures 2 to 4. The calculation of the Stokes vector is explained in detail below. A critical innovative aspect of the present invention is that it avoids the problematic use of microretarder arrays described in the prior art and uses instead a homogeneous dispersive retarder. The use of a color (or wavelength) dependenthomogeneous dispersive retarder (or two homogeneous dispersive retarders in someparticular embodiments) in combination with the color (or wavelength) sensitive polarizationsensor provided in the present invention allows to calculate the full Stokes vector. Themathematical calculation of the Stokes vector is explained in detail below. Different mathematics / equations are required to solve the Stokes vector when using a homogeneous dispersive retarder as compared to a microretarder array. It is shown that is impossible to find a mathematical solution for Stokes when using a microretarder array. Description of the Stokes vector The Stokes vector is a vector containing 4 real parameters that will be measured: S0 describes the total intensity, S1 and S2 are the linear polarization components and S3 the circular polarization component. Description of a linear retarder In the context of polarization optics, a retarder is an optical component that modifies the polarization state of light by changing the Stokes vector (S) into a different one (S’): S'^ M R SWhere MRis the Mueller matrix of a retarder. For a linear retarder with an azimuth angle θ and retardance δ, the Mueller matrix is given by: Retarders are made by transparent materials (most usually crystals and polymers). In thesematerials δ depends on the wavelength (dispersion).In a homogenous retarder both θ and δ are uniform on the entire surface of the retarder. Ina microretarder array, θ and / or δ change periodically across the surface, which is achievedby advanced micro- or nano-fabrication procedures.Working principle of the full-Stokes camera A polarization camera works by connecting measured intensities for specific polarizations (I) with the incoming stokes vector (S) I^ WS (3)Where W is an instrumental matrix that describes the polarization behavior of the camera. In a standard polarization camera (containing pixels with 0º, 90º, 45º and 135º polarization orientations). In this case Eq. (3) becomes: In order to obtain the S vector, this equation needs to be inverted (S^ W^ 1 I ), but thisinversion is not possible because W here is a singular matrix (det (W)=0).If a homogeneous linear retarder is placed in front of the polarization camera (disregarding color detection). Eq. (4) is transformed to: For θ=0º, equation (5) becomes into: The inversion of these equations (5,5’) remains impossible also here because W still is asingular matrix (det (W)=0). If color (wavelength) detection is added the above equation needs to be modified because,due to the dispersive homogenous retarder, each color will have a different associatedretardance In case two wavelengths are used, the equation will be: where superscripts or subscripts 1 and 2 are associated with two different colors (orwavelengths) detected by the polarization camera. For θ=0º, equation (6) becomes into: These equations (6, 6’) can be used to determine the complete Stokes vector from the Ivector. As here W is not square matrix, the pseudoinverse (W+) is used: S^ W ^ IFull-Stokes detection is enabled because different color channels will experience different retardances, allowing the system to analyze distinct polarization states in each color channel, thereby mathematically addressing the inversion problem that cannot be resolved when only a single, homogeneous retardance is considered.The advantages of the homogenous dispersive retarder of the present invention overmethods using retarder arrays or metasurfaces are:^ Simpler Setup. The homogeneous retarder in combination with a polarization sensoris simpler and more cost-effective than microretarder arrays or metasurfaces, which require complex designs and intricate nanofabrication processes. ^High Compatibility with Existing Technology. The present invention can be easilyimplemented using commercially available color polarization cameras, making it accessible and avoiding the need for custom-built or specialized equipment. It also does not increase the bulk or power consumption of the camera. ^No Need for Nanofabrication: Sensors based on retarder arrays or metasurfacesrequire complex, highly specialized, and expensive nanofabrication processes. In contrast, the homogeneous retarder approach utilizes readily available components, such as crystal plates or polymer films, completely bypassing the need for advanced fabrication techniques. ^No Crosstalk or Alignment Issues: Microretarder arrays require precise alignmentand size matching between the retarder array and the photodiode array to preventcrosstalk between adjacent pixels, which will cause pixel noise and hence a lowerimage quality. This makes implementation complex and highly dependent on the specific type and size of the photodiode array. In contrast, the homogeneous retarder provides uniform retardation across the entire surface, eliminatingalignment challenges and ensuring compatibility with any photodiode array. BRIEF DESCRIPTION OF THE FIGURESFigure 1 shows a polarization camera system. It is the commercially available state of theart. Figure 2 shows a full-Stokes polarization camera system according to a first embodiment of the present invention. Figure 3 shows a full-Stokes polarization camera system according to a second embodiment of the present invention.Figure 4 shows a full-Stokes polarization camera system according to a third embodimentof the present invention.DESCRIPTION OF PREFERRED EMBODIMENTS OF THE INVENTIONReferences used in the figures: 1.- Full-Stokes polarization camera system;2.- Polarization sensor:^ polarizer array (2a);^ color channel filter (2b)^ photodiode array (2c)3.- first homogeneous dispersive retarder; 3’.- second homogeneous dispersive retarder;4.- light entrance for incident polarized light; 5.- microprocessor.The first embodiment is given in Fig.2 and it is mainly based on a binary color channel filterand a polarizer array having four orientations. Two different wavelengths or colors areenough for the Full Stokes vector measurement. The homogeneous dispersive retarder 3can be easily accomplished using crystal plates (also known as waveplates) as crystals arenaturally dispersive materials, so that for the wavelengths of interest λ1 and λ2 the retardanceare different [δ(λ1) ≠ δ(λ2)]. The conditions of the measurement are optimal when |δ(λ1) -δ(λ2) | = 90o because it leads to maximum separation of polarization states between the twowavelengths λ1 and λ2, enhancing the system's ability to capture a wide range of polarizationinformation. However, measurements are also possible if the difference in retardancebetween the two wavelengths is less or more than 90 degrees, i.e., between 1 and 179degrees. The full-Stokes polarization camera system 1 as shown in FIG.2 comprises the polarization sensor 2, the light entrance 4 for incident polarized light, the microprocessor 5 and the firsthomogeneous dispersive retarder 3. The polarization sensor 2 comprises: the polarizerarray 2a having four orientations, the color channel filter 2b configured to detect two wavelengths “λ1” and “λ2” and the photodiode array 2c configured to measure eightintensities since two wavelengths pass through four orientations. The first homogeneousdispersive retarder 3 is positioned between the polarization sensor 2 and the light entrance4. The first homogeneous dispersive retarder 3 has an axis orientation -θ-, and it isconfigured to introduce wavelength-dependent retardation δ(λ) to the incident polarizedlight, so that [δ(λ1) ≠ δ(λ2)]. The microprocessor 5 is connected to the photodiode array 2c,and it is also configured to calculate a complete Stokes vector of the incident polarized lightpassing through the light entrance 4 from the eight measured intensities. Calibration procedures are crucial to ensure the accuracy and consistency of polarizationmeasurements across the wavelength sensitive sensor. This mainly involves thedetermination of the axis orientation -θ- of the homogeneous dispersive retarder 3 and theretardance at each wavelength δ(λ1) and δ(λ2), for known polarization inputs of the incidentpolarized light. To calculate the complete Stokes vector of the incident polarized light, the microprocessor5 employs mathematical algorithms, such as inversion techniques. These mathematicalalgorithms analytically calculate, in real time, the values of S0, S1, S2 and S3 from measured intensities provided that the values of the θ, δ(λ1) and δ(λ2) are known (they are determined from calibration) The above calibration procedures and Stokes vector calculation, also apply to the following second and third embodiments as well as any other part of the present disclosure. The second embodiment is given in Fig.3 and it is mainly based on a RGB color channel filter and a polarizer array having four orientations. Bayer filter arrays are commonly used in digital cameras to capture color information. This array typically consists of red, green, and blue (RGB) color channel filters arranged in a specific pattern over the camera sensor (photodiode array). Each pixel on the sensor is filtered to detect either red, green, or blue light. This implementation enables the full-Stokes polarization camera system 1 to capture both color and full polarization information simultaneously, providing a fused output that allows for comprehensive analysis and visualization of polarized scenes. The configuration of the color polarization camera is illustrated in Fig. 3. In the embodiment of Fig. 3, two color channels (e,g, the Green (G) and Red (R), or the Blue (B) and Green (G)) are enough to extract the full polarization information, so that the Stokes vector of the incident light at the light entrance 4 is shown at the camera frame rate. The remaining color channel is only used to capture the color image. Alternatively, all three-color channels can be used for polarization measurement. This leads to twelve intensities since three wavelengths (colors) pass through four orientations. The full-Stokes polarization camera system 1 as shown in FIG.3 comprises the polarization sensor 2, the light entrance 4 for incident polarized light, the microprocessor 5 and the first homogeneous dispersive retarder 3. The polarization sensor 2 comprises: the polarizer array 2a having four orientations, the color channel filter 2b configured to detect three wavelengths “λ1” (Red – 750 nm), “λ2” (Green 550nm) and “λ3” (Blue 470nm) and the photodiode array 2c configured to measure eight intensities since two wavelengths pass through four orientations. The first homogeneous dispersive retarder 3 is positioned between the polarization sensor 2 and the light entrance 4. The first homogeneous dispersive retarder 3 has an axis orientation -θ-, and it is configured to introduce wavelength-dependent retardation δ(λ) to the incident polarized light, so that [δ(λ1) ≠ δ(λ2)]. The microprocessor 5 is connected to the photodiode array 2c, and it is also configured to calculate a complete Stokes vector of the incident polarized light passing through the light entrance 4 from the eight measured intensities.The third embodiment is given in Fig.4 and it is mainly based on a two-colors channel filterand a polarizer array having two orientations. This implementation is operative to determinethe full Stokes vector when the orientation of the two homogenous retarders 3 and 3’ thatare superposed, are oriented at different angles (θ and θ’). For this implementation it isalways necessary that θ≠θ’. This superposition of two misoriented retarders is the mostflexible and general form of retarder (sometimes it is called an “elliptical retarder”). Note that two homogenous retarders 3 and 3’ in Fig.4 can be chosen as two identical misalignedretarders (δ(λ1)= δ’(λ1), δ(λ2)= δ’(λ2) and θ≠θ’ ) or they can have different retardances (δ(λ1)≠ δ’(λ1), δ(λ2) ≠ δ’(λ2) and θ≠θ’).The full-Stokes polarization camera system 1 as shown in FIG.4 comprises the polarizationsensor 2, the light entrance 4 for incident polarized light, the microprocessor 5, the firsthomogeneous dispersive retarder 3 and the second homogeneous dispersive retarder 3’.The polarization sensor 2 comprises: the polarizer array 2a having two orientations, thecolor channel filter 2b configured to detect three wavelengths “λ1” and “λ2” and thephotodiode array 2c configured to measure four intensities since two wavelengths passthrough four orientations. The first homogeneous dispersive retarder 3 and the second homogeneous dispersive retarder 3’ are facing each other and positioned between the polarization sensor 2 and the light entrance 4. The first homogeneous dispersive retarder 3has an axis orientation -θ-, and it is configured to introduce wavelength-dependentretardation δ(λ) to the incident polarized light, so that [δ(λ1) ≠ δ(λ2)]. The secondhomogeneous dispersive retarder 3’ is configured to introduce wavelength-dependentretardation δ’(λ) so that [δ’(λ1) ≠ δ’(λ2)], and it also has an axis orientation -θ’-. The firsthomogeneous dispersive retarder 3 and the second homogeneous dispersive retarder 3’ are further configured to provide retardances so that [δ(λ1)= δ’(λ1); δ(λ2)= δ’(λ2); θ≠θ’] or [δ(λ1) ≠ δ’(λ1); δ(λ2) ≠ δ’(λ2); θ≠θ’]. The microprocessor 5 is connected to the photodiode array 2c, and it is also configured to calculate a complete Stokes vector of the incidentpolarized light passing through the light entrance 4 from the four measured intensities (sincetwo wavelengths pass through two orientations). This is the minimum number of detectedintensities that allow the full Stokes vector measurement.

Claims

CLAIMS1.- A full-Stokes polarization camera system (1) comprises:^ a polarization sensor (2), which in turn comprises:o a polarizer array (2a) having at least two orientations;o a color channel filter (2b) configured to detect at least two wavelengths (λi,λj); and, oa photodiode array (2c) configured to measure at least four intensities, oneintensity for each wavelength and orientation combination;^ a light entrance (4) for incident polarized light;^ a microprocessor (5) connected to the photodiode array (2c);^ a first homogeneous dispersive retarder (3) positioned between the polarizationsensor (2) and the light entrance (4), and having an axis orientation -θ-;wherein the first homogeneous dispersive retarder (3) is configured to introducewavelength-dependent retardation δ(λ) to the incident polarized light, so that [δ(λi) ≠ δ(λj)];and, wherein the microprocessor (5) is configured to calculate a complete Stokes vector ofincident polarized light from the measured intensities.2.- A full-Stokes polarization camera system (1), according to claim 1, wherein when thepolarizer array (2a) has four orientations and the color channel filter (2b) is configured to detect two wavelengths (λi, λj), the photodiode array (2c) is configured to measure eight intensities. 3.- A full-Stokes polarization camera system (1), according to claim 1, wherein when the polarizer array (2a) has four orientations and the color channel filter (2b) is a Red-Green- Blue “RGB” sensor, the color channel filter (2b) is configured to detect the wavelengths of the red, green and blue colors, so that the system is configured to provide color images. 4.- A full-Stokes polarization camera system (1), according to claim 1, wherein when the polarizer array (2a) has at least two orientations, the system further comprises a secondhomogeneous dispersive retarder (3’) configured to introduce wavelength-dependentretardation δ’(λ) so that [δ’(λi) ≠ δ’(λj)], wherein the second homogeneous dispersiveretarder (3’) is placed between the first homogeneous dispersive retarder (3) and thepolarization sensor (2), and having an axis orientation -θ’-, so that [δ(λi)= δ’(λi); δ(λj)= δ’(λj); θ≠θ’] or [δ(λi) ≠ δ’(λi); δ(λj) ≠ δ’(λj); θ≠θ’]. 5.- A full-Stokes polarization camera system (1), according to any of claims 1 to 4, wherein the first homogeneous dispersive retarder (3) and / or the second homogeneous dispersive retarder (3’) provide a retardance from 1º to 179º degrees. 6.- A full-Stokes polarization camera system (1), according to claim 5, wherein the first homogeneous dispersive retarder (3) and / or the second homogeneous dispersive retarder (3’) provide a retardance difference of 90 degrees: |δ(λi) - δ(λj) | = 90o. 7.- A full-Stokes polarization camera system (1), according to any of claims 1 to 4, wherein the microprocessor (5) is further configured to carry out a calibration procedure consisting of calculating an orientation of the retarder’s axis (θ, θ’) and the retardance (δ, δ’) at each wavelength (δ(λi), δ(λj)), for known polarization of the incident polarized light. 8.- A full-Stokes polarization camera system (1), according to any of the preceding claims 1, wherein the microprocessor (5) is further configured to calculate the Stokes vector from the measured intensities:wherein I0, I90 represent the intensity of horizontal and vertical directions, I45, I135 represent the intensity of 45◦and 135◦directions, Il and Ir represent the intensity of light with left circular and right circular polarization.

Images