System for acquisition and spectral analysis

Abstract

A system is attained for the acquisition and spectral analysis of images, which comprises hardware and software elements such as: - an optical sensor with an acquisition capacity at least comprised between 300 nm and 1000 nm; - at least a multi-filter apparatus comprising in turn: - at least a photographic lens; - at least Bayer type filters; - at least a first filter; - at least a second filter per; - at least a focusing system for each of said first and second filters. - at least a Xenon photographic flash with an average spectral radiance at least equal to 2 μW / (cm2 sr nm) for wavelengths at least comprised between 300 nm and 1000 nm; - a calibration target; - a software apparatus which executes spectral analyses of the images acquired by means of the optical sensor with the multi-filter apparatus and the photographic flash.

Description

[0001] LEIBO. / 33e2025

[0002] “System for acquisition and spectral analysis”

[0003] Description

[0004] Field of the art

[0005] The invention refers to the field of photography and visual studies. Specifically, the invention provides a system for an in-depth spectral analysis that uses optical sensors, even common ones.

[0006] Prior art

[0007] Up to present, many research and analysis activities require increasingly sophisticated, less invasive systems and methods for studying images. Such activities range from those of study, restoration and preservation of art works in the cultural heritage world, to those of medical diagnostics and screening (e.g. dermatological), to those plant analysis in agricultural - among many other activities.

[0008] In the example of cultural heritage, particular attention is given to the preparatory steps of the actual restoration activity, whose which examine the effective state of preservation of the work, the materials employed, the state of preservation of the supports, possible restoration interventions, specific volume or surface pathologies, and the effects of the environment over time. In this important and delicate phase, the ideal solution would be to reach a perfect and complete understanding of all the materials used in the original work, as well as all the substances that were added over time due to aging or artificially. But such knowledge would eventually require some kind of invasive intervention on the paintings, through, for example, the extraction of micro-samples followed by chemical analysis, mass spectrometry analysis, scanning microscope analysis. It is obvious that one rarely has the possibility to take such micro-samples, thus the improvement or innovation in the field of non-invasive analysis represents a step forward in diagnosis.

[0009] The visible spectrum band is only one among those used in the various activities, including LEIBO. / 33e2025 those mentioned above. The imaging in the UV (UltraViolet) and IR (InfraRed) bands has been used since the days of analogic photography, with the use of special films that are selectively sensitive to those bands. A film of particular interest was that so-called “infrared false colors”. This is a film with three layers which, rather than capturing the RGB channels, captures the IR, Red and Green bands. Once developed, the film shows the intensity IR as red intensity, the original red as green and the original green as blue. Most of expert diagnostic technicians usually compare the images with real colors to those with false colors in order to have, from the color displacement, a good idea of the nature of the pigments used, or at least several of these. The advent of digital photography, together with the quality and reliability of modern cameras, opened the way to increasingly sophisticated and accurate examinations. Moreover, many cameras are equipped with a silicon sensor (generally CMOS - Complementary Metal-Oxide-Semiconductor) which in native mode has a broader spectral sensitivity than the visible band. This offers the world of diagnostics a very flexible instrument to use for investigated a broad spectrum. Today in reality it is rather common practice to take pictures with digital cameras not only in the visible range but also in ultraviolet (UVA) and in near infrared (NIR), as well as in mixed bands such as UV / Vis with fluorescence images and in Vis / NIR luminescence images. In addition to the availability of digital cameras modified for UV and IR, the availability of image modification programs in general and the manipulation of digital images has allowed constructing an IR photo with false colors by executing two pictures, one in the visible range and one in the IR, reorganizing the RGB channels with the sequence IR-R-G. The same concept was then applied to UV, attaining an acquisition of two images (UV+Vis) and moving the channels in order to obtain the sequence G-B-UV (UV with false colors). With regard to the false colors (both IR and UV), however, while the film provided a corresponding constancy of the ratio of exposure between the three channels (intrinsic with the manner of production of the film itself) and necessarily a perfect recording of the three channels, the acquisition of two images in two separate time instants does not per se ensure perfect superimposition of the images and perfect exposure equilibrium. The double occurrence in the visible and in the IR (or UV) and their recombination does not LEIBO. / 33e2025 ensure any of these things, thus providing results whose quality depends on the precision of the operator and on the software post-processing instruments.

[0010] In addition, up to this point in the development of the photographic research techniques, the analysis has always been done visually, by comparing at most three bands simultaneously, due to the intrinsic limit of the human eye. On the other hand, the availability of sensors that are increasingly more sophisticated in terms of available megapixels, linear behavior and high reliability, and rather high numerical dynamics (now easily at 14 bit, for over 16000 levels per color), together with increasingly higher processing capacities, at low cost and a greater memory capacity, suggest that it is necessary to significantly improve quality and move in two directions: a standardization of the representation of this broader available band; the introduction of innovative algorithms in order to attain images that are perfectly recorded, and exploit the new wealth of bands simultaneously to the greatest possible extent.

[0011] In the context of imaging and of spectral or multispectral hypercolorimetry, there are numerous patents which provide solutions to problems of various nature.

[0012] One example is the object of the patent US 10228283B2 by NAHT SUMIT et al. The invention provides a spectral imaging system that includes a spectrometer and an optical imaging system. The spectrometer is usable for generating spectral signatures of objects from a scene. The optical imaging system is usable for generating six or more responses from the same scene. Each of the six or more responses represents a different spectral content of the objects in the scene. The outputs generated by the optical imaging system can be used for generating a hypercube by using spectral reconstruction techniques. In one embodiment, the spectral imaging system of the NAHT SUMIT invention could be implemented as a part of a cellphone. Another example is the object of the patent application CN113125010A by YU ZHENMING et al. The invention refers to an apparatus for hyperspectral image acquisition. The apparatus comprises a lens, a first relay lens, a second relay lens, a third relay lens, a beam splitter, a prism, a reflecting mask and a sensor of the camera, in which the image lens, the first relay lens, the beam splitter, the prism, the second relay lens and the reflecting mask are arranged in sequence in the direction of transmission of the incident light, and the third relay lens and LEIBO. / 33e2025 the sensor of the camera are arranged in sequence in the direction of reflection of the reflected light that passes through the beam splitter; when the device is in a condition for acquiring the image of the scene, the reflecting mask reflects the incident light on the prism, the prism executes the reverse dispersion on the incident light and directs the incident light onto the beam splitter, and the beam splitter reflects the light on the sensor of the camera through the third relay lens, thus the sensor of the camera acquires a compressed observation image and a two-dimensional image.

[0013] The inventions reported up to now merely as a non-limiting example are representative of the inventive-technological picture that is available today.

[0014] The abovementioned inventions have technical problems tied to the fact that they do not have portable apparatuses which allow analysis on the field also in video format (not only photographs) and which in any case allow having errors smaller than 2% in detecting the spectral reflectance. Many of the systems that are known today make use of a division into a great number of infrared bands, however statistical regressions (or other processing) make the data lose its “purity”, then allowing displaying only the statistical processing and therefore not having a direct contact with the data acquired with the apparatus.

[0015] Still many systems, like those of the hyperspectral cameras, make use of a pixel by pixel scanning of an image in order to be able to suitably characterize the emission spectrum thereof, nevertheless resulting in extremely long acquisition times and in limited fields of application. The present invention is adapted to provide a new paradigm in the digital photography field, continuing on the evolutive development and coming to enlarge the possibilities, offered even by a commercial reflect camera.

[0016] The present invention aims to provide a system for the acquisition and spectral analysis of images adapted to be easily installable on an existing optical sensor (such as those of a camera, by operating simple modifications and using a suitable hardware and software apparatus) and to allow not only an immediate spectral analysis on the field, but also an analysis with an error lower than 2% with respect to the spectral reflectance values of each pixel of the acquired image. The invention therefore allows having a high number of spectral bands for each image, LEIBO. / 33e2025 allowing an in-depth analysis thereof, through a single picture that is then filtered through numerous filters that allow analyzing the electromagnetic waves not only in the visual fields by also in the ultraviolet (UV) and infrared (IR) fields.

[0017] The advantages offered by the present invention will be clearer in light of the following detailed descriptions.

[0018] Description of the invention

[0019] According to the present invention, a system is attained for the acquisition and spectral analysis of images which comprises hardware and software elements.

[0020] The hardware elements are:

[0021] - an optical sensor for acquiring images with an acquisition capacity for electromagnetic waves with wavelength at least comprised between 300 nm and 1000 nm;

[0022] - at least a multi-filter apparatus connected to the optical sensor and filtering the entering electromagnetic waves. The multi-filter apparatus comprises in turn: o at least a photographic lens; o at least Bayer type filters for acquiring the wavelengths of the colors in the visible spectrum by mosaicing the acquired image; o at least a first filter capable of passing electromagnetic waves with wavelength at least comprised between 300 nm and 600 nm with a transmittance greater than 50%. Said first filter provides three pieces of spectral reflectance information for each pixel of an image acquired after said mosaicization; o at least a second filter capable of passing electromagnetic waves with wavelength at least comprised between 700 nm and 1000 nm with a transmittance greater than 50%, waves with wavelength comprised between 300 nm and 500 nm with a transmittance greater than 50% and waves with length comprised between 350 nm and 450 nm with a transmittance greater than 80% and waves with length comprised between 500 nm and 700 nm with a transmittance greater than 15%. Said second filter provides three pieces of spectral reflectance information for each pixel of an image acquired after LEIBO. / 33e2025 said mosaicization; o at least a focusing system for each of said first and second filters.

[0023] Other hardware elements of the system are:

[0024] - at least a Xenon photographic flash with an average spectral radiance at least equal to 2 pW / (cm2sr nm) for wavelengths at least comprised between 300 nm and 1000 nm and with a spectral radiance greater than 1.3 pW / (cm2sr nm) for wavelengths at least comprised between 350 nm and 1000 nm. Said photographic flash has positive spectral radiance, different from zero, for wavelengths comprised between 300 nm and 400 nm;

[0025] - a calibration target which comprises colors and boxes whose reflectance, for each single color or white square, are known.

[0026] The system then comprises at least a software apparatus which executes spectral analyses of the images acquired by means of the optical sensor with the multi-filter apparatus and the photographic flash. The software apparatus comprises: o one or more databases suitable for containing data regarding the spectral reflectance of elements to be examined; o a transformation module suitable for automatically aligning said images and providing data relative to the automatic transformations performed. Said data relative to the automatic transformations performed can also be manually modified; o a calibration module in which the colors of the calibration target within said image are associated with the relative known reflectance. Said calibration module calibrates the reflectance values of each pixel of said images based on the associations between image of the calibration target and the relative known reflectance; o a division module for the spectrum of the images acquired into a plurality of subbands, all having the same form and the same percentage transmittance. For each subband, said division module provides a unique reflectance value calculated based on the input data. Said unique reflectance value is provided for each pixel for the wavelength value, centered in said sub-band; o a comparison module which operates to compare between individual pixels and / or LEIBO. / 33e2025 groups of pixels of an image, by providing the differences of the spectral reflectance for each of the sub-bands obtained from the division module. The comparison module provides a graphical interface to a user for the selection of individual pixels and / or groups of pixels to be compared and / or by carrying out said selection of pixels in an automatic manner by selecting pixels and / or groups of pixels present in two or more regions of the acquired image; o an analysis module which performs graphical, numerical and / or statistical analyses on the spectral reflectance data of the bands obtained with said division module.

[0027] The system of the present invention is then employing by taking a single photograph of a scene in which also the calibration target is framed, or by first taking a photo in the same same, framing the target, and then a photo of the scene without target.

[0028] Once an image of the target is obtained in the same scene illumination conditions, or an image is taken of the elements to be examined, it is possible to send the image to the software apparatus in order to calibrate the colors thereof, to divide the spectrum thereof into various bands and thus be able to carry out different types of analyses.

[0029] In a preferred embodiment of the present invention, the system provides for the use of a reflex camera which is modified by removing a UV-IR cut filter and on whose optical sensor the multi-filter apparatus is mounted. The acquisitions of the multi-filter apparatus, in the example in which it comprises two channels, one with the first filter and the other with the second filter, are each reported on half of the optical sensor (since the channels of the example are two). In the event in which the multi-filter apparatus comprises for example four acquisition channels, each of these will report the electromagnetic waves acquired on a fourth of the optical sensor of the reflex camera.

[0030] The advantages offered by the present invention are evident in light of the description made up to now and will be even clearer due to the enclosed figures and to the relative detailed description.

[0031] Description of the figures LEIBO. / 33e2025

[0032] The invention will be described in at least a preferred embodiment as a non-limiting example with the aid of the enclosed figures, in which:

[0033] FIGURE 1 shows a general view of a system 100 according to the present invention;

[0034] FIGURE 2 shows the scheme of a multi -filter apparatus 110 with a first filter 113 and a second filter 114, with a beam splitter 117 and with a tilted mirror 118, according to the present invention;

[0035] FIGURE 3 shows a schematic representation of a multi-filter apparatus 110 in which a beam splitter 117 divides the entering electromagnetic waves into three outputs that are respectively directed towards a first filter 113, a second filter 114 and an additional filter 115 by means of two tilted mirrors 118 according to the present invention;

[0036] FIGURE 4 shows an axonometric view of a multi-filter apparatus 110 with two focusing systems 116 according to the present invention;

[0037] FIGURE 5 shows a section A-A’ of the multi-filter apparatus 110 of FIG. 4;

[0038] FIGURE 6 shows a device with an optical sensor 101 and a multi -filter apparatus 110 mounted on a quadcopter drone, on which a photographic flash 120 is also mounted.

[0039] Detailed description of the invention

[0040] The present invention will now be illustrated merely as a non-limiting or non-binding example, with reference to the figures which illustrate several embodiments relative to the present inventive concept.

[0041] With reference to FIG. 1, a general view is shown of a system 100 according to the present invention. In FIG. 1 as in the following description, the embodiment of the present invention is illustrated which up to present is deemed to be the best.

[0042] As shown in FIG. 1, according to the present invention, a system 100 is attained for the acquisition and spectral analysis of images comprising:

[0043] - an optical sensor 101, such as by way of a non-limiting or non -binding example the optical sensor 101 of a camera regarding which a filter for cutting the ultra-violet and infrared wavelengths said filter UV-IR has been removed; said optical sensor 101 is capable of LEIBO. / 33e2025 performing at least an image acquisition and has an acquisition capacity for electromagnetic waves with wavelength at least comprised between 300 nm and 1000 nm and still more preferably comprised between 200 nm and 1200 nm;

[0044] - at least a multi-filter apparatus 110 (shown in detail and in several embodiments in FIGS. 2, 3 and 5) connected to said optical sensor 101 and adapted to filter the electromagnetic waves (i.e. light / the images) entering into said optical sensor 101. Said multi-filter apparatus 110 comprises, in turn: o at least a photographic lens 111. Said lens 111, in several embodiments of the present invention is a wide-angle lens; o at least Bayer type filters 112 for acquiring the wavelengths of the colors in the visible spectrum relative to Red, Green and Blue (RGB) by mosaicing the acquired image; o at least a first filter 113 capable of passing electromagnetic waves with wavelength at least comprised between 300 nm and 600 nm and more preferably at least comprised between 200 nm and 600 nm with a transmittance greater than 50%, preferably greater than 70% and still more preferably greater than 80%. Where the transmittance is the capacity of a material to be crossed by a part of the incident light. Since therefore the transmittance, the ratio between intensity of the transmitted radiant flow and intensity of the incident radiant flow, is a non-dimensional size, therefore its percentage expression is indicative of the value of the intensity of the incident radiant flow with respect to that transmitted. Said first filter 113, coupled to said Bayer type filters 112, thus provides three pieces of spectral reflectance information for each pixel of an image acquired after said mosaicization; o at least a second filter 114 capable of passing electromagnetic waves with wavelength at least comprised between 700 nm and 1000 nm with a transmittance greater than 50%, preferably greater than 70% and still more preferably greater than 90%, waves with wavelength comprised between 300 nm and 500 nm with a transmittance greater than 50% and more preferably greater than 70%, waves with wavelength comprised between 350 nm and 450 nm with a transmittance greater than 80% and preferably LEIBO. / 33e2025 greater than 90% and waves with wavelength comprised between 500 nm and 700 nm with a transmittance greater than 15%. Said second filter 114, coupled to said Bayer type filters 112 is adapted to provide three pieces of spectral reflectance information for each pixel of an image acquired after said mosaicization; o at least a focusing system 116 for each filter 113, 114;

[0045] - at least a Xenon photographic flash 120 with an average spectral radiance at least equal to 2 pW / (cm2sr nm) for wavelengths at least comprised between 300 nm and 1000 nm and with a spectral radiance greater than 1.3 pW / (cm2sr nm) for wavelengths at least comprised between 350 nm and 1000 nm. Said photographic flash having positive spectral radiance, different from zero, for wavelengths comprised between 300 nm and 400 nm.

[0046] By way of a non-limiting or non-binding example, said photographic flash 120 is a Nikon® SB910FR flash from which the front optical group constituted by two Fresnel lenses has been removed.

[0047] - a calibration target 130 comprising colors and boxes whose reflectance, for each single color or white square, are known, since they have been previously accurately measured, for example by means of a spectroradiometer;

[0048] - at least a software apparatus 200 suitable for performing a spectral analysis of the images acquired by means of the optical sensor 101 with the multi -filter apparatus 110 and the photographic flash 120; said software apparatus 200 comprising: o one or more databases 201 suitable for containing data regarding the spectral reflectance of elements to be examined. By said “elements to be examined” it is intended by way of a non-limiting or non -binding example, in the scope of the present invention, pigments (to which pigment spectral reflectance databases 201 correspond, as shown in FIG. 1), skin (moles, marks, reddening, inflammation, vitiligo, acne, black spots, red spots and / or other items), vegetation (leaves, aesthetic symptoms of plant diseases), metal surfaces (to which spectral reflectance databases 201 correspond which are linked to surface damage, presence of rust in various stages of deterioration, presence of animal guano, chemical deterioration and other items), LEIBO. / 33e2025 elements that could be found at a crime scene (the databases 201 in this case contain data relative to the spectral reflectance of drug substances, body fluids and / or other forensic data), various materials (the databases 201 in these cases contain spectral reflectance data relative to various materials such as transparent plastics); o a transformation module 210 suitable for automatically aligning said images. The original pictures, indeed, even if executed with extreme accuracy, do not coincide pixel per pixel, also because with the variation of the wavelength, the refraction indices change in the various optical windows / glasses / filters of the apparatus 110 and thus the focusing changes. Together with the focusing, also the enlargement ratio changes and consequently, in order to attain a perfect superimposition between the images at the various bands, operations of scale, rotation and translation are necessary through a similar transformation. The alignment is automated by employing image recognition algorithms. Such algorithms are well known up to now, one need only recall very well-known example like those provided by MATLab® and / or SimuLINK®. Such algorithms recognize elements present in said images and therefore provide for operating transformations in order to ensure a coincidence and superimposition of the subjects in said images. Said transformation module 210 therefore provides data related to the automatic transformations performed. Said data relative to the automatic transformations performed can also be manually modifiable. For example the transformation module 210 can show that in order to superimpose a first image on a second image, it had to operate an enlargement of the first image equal to 1.68% (percentage with respect to the original total size of the first image). In the event in which the user detected the need to modify said scale, he / she could, through the interface of the transformation module 210, manually modify said value equal to 1.68% (for example by clicking on it and typing a different value by means of keyboard) and / or he / she could operate other transformations such as for example the rotation of one of the images; o a calibration module 220 in which the colors of the calibration target 130 within said LEIBO. / 33e2025 image are associated with the relative known reflectance. Said calibration module 220 calibrates the reflectance values of each pixel of said images based on the associations between image of the calibration target 130 and the relative known reflectance. As seen, such reflectance is measured for each single white color square, accurately with a spectroradiometer. These measurements are of maximum importance since they represent the point of connection between the precision of reflectance measured by means of the optical sensor 101 and the actual reflectance. By way of preferred example, said spectroradiometer allows a traceable calibration, referring to the international metrological institutes. Said reflectance values of the pixels of the images are evaluated starting from functions that provide the most affinity (preferably evaluated in percentage) between said known reflectance and the colors of the calibration target 130 acquired with the images. From each image acquisition, “N” numerical reflectance values are in fact obtained, where “N” is given by three (values deriving from the mosaicization operated by the Bayer filters 112) multiplied by the number of filters 113, 114 but also other filters 115 when present (see hereinbelow). The calibration module 220 then operates a transformation from N numbers to 1 number relative to the reflectance value of each pixel. Said N numbers are transformed through N mathematical functions which are those adapted to provide the most percentage affinity between said known reflectance and the colors of the calibration target 130 acquired with the images; o a division module 230 for the spectrum of the images acquired into a plurality of subbands, all having the same form and the same percentage transmittance. For each subband, said division module 230 provides a unique reflectance value calculated based on the input data which is “N” equal to three (mosaicization operated by the Bayer filters 112) multiplied by the number of filters 113, 114 (but also other additional filters 115, as will be illustrated hereinbelow, relative to several embodiments of the invention) employed in the multi -filter apparatus 110. Said unique reflectance value is provided for each pixel for the wavelength value, centered by said sub-band. LEIBO. / 33e2025

[0049] Indeed, specifically the sub-bands preferably have triangular or Gaussian bell-shaped form and reach a transmittance of 100% or in any case the maximum transmittance value, at the central / average wavelength value. Other forms of sub-bands are possible. The transformation operated by the calibration module 220 from N to 1 numbers must be valid for each sub-band. Therefore, it is understood how a higher number N allows obtaining a greater number of sub-bands (hence greater capacity for spectrum analysis) without losses in terms of data reliability. By way of a nonlimiting or non-binding example, a reliability in the calibration equal to or greater than 98% is obtained by employing said filters 113 and 114 and dividing the spectrum from 300 nm to 1000 nm into 13 sub-bands. By way of a non-limiting or non-binding example, the same reliability equal to or greater than 98% is obtained by employing three filters (the filters 113 and 114 and an additional filter 115) and dividing into 20 sub-bands. By way of a non-limiting or non-binding example, the same reliability equal to or greater than 98% is obtained by employing five filters (the filters 113 and 114 and three additional filters 115) and dividing into 30 sub-bands. Further divisions into sub-bands, with the increase of the number of filters (and consequently with the increase of the resolution of the optical sensor 101 which is divided by the number of filters employed) are possible without exiting from the protective scope of the present invention; o a comparison module 240 capable of performing a comparison between individual pixels and / or groups of pixels of an image, providing the differences of the spectral reflectance for each of the sub-bands obtained from the division module 230. Said comparison module 240 providing graphical interfaces to a user for the selection of the individual pixels and / or of the groups of pixels to be compared and / or performing said selection of pixels in an automatic manner by selecting pixels and / or groups of pixels present in two or more regions of the acquired image;

[0050] By way of a non-limiting or non-binding example, said modules 210-240 acquire input formats of the type TIFF 16, TIFF8, JPG, RAW (Nikon*. NEF, Hasselblad LEIBO. / 33e2025

[0051] *.FFF, PhaseOne *.IQ4 and Canon*. CR2) as well as images from different types of optical sensors 101 (Raggi-X, XRF, THz and / or other items).

[0052] Said calibration module 220 obtains a radiometric precision greater than 98% with a colorimetric precision AE2000 < 2 by using artificial intelligence (such as color and image detection algorithms) for the execution of said calibration.

[0053] Said modules 210-240 providing by way of a non-limiting or non-binding example formed of outputs such as TIFF 16, TIFF8, JPEG, PNG. o an analysis module 250 capable of performing graphical, numerical and / or statistical analyses on the spectral reflectance data of the bands obtained with said division module 230. Said analysis module 250 is capable of performing a PCA (Principal Component Analysis) of said images, a colorimetric reading and data reading of the spectral reflectance, an analysis of clustering segmentation type based on each spectral data and image available, evaluation of a Normalized Difference Index (NDI) in order to discern surface characteristics and pigments, a rendering of the false infrared and ultraviolet colors as well as another analysis type.

[0054] In the scope of the example made relative to the use of the system 100 for detecting “elements to be examined” such as various materials, in several embodiments, the system 100 can be employed for averting possible falsifications of impressions / fmgerprints made with plastics or other transparent materials. The optical sensor 101 is, in these cases, the sensor of a fingerprint reader that automatically carries out a spectral analysis in order to determine if, in the acquired image, the spectral reflectance is indicative of skin or of plastic elements potentially linked to falsification attempts. In these cases, the entire multi-filter apparatus 110 is closed within a device that processes the acquired images of the fingerprints. Preferably said device that processes the acquired images of the fingerprints is compact, such that it can be set on a surface. Preferably said device comprises a surface on which the fingertips are set for the detection of the fingerprints. Below said surface, enclosed in a body of the device, said multifilter apparatus 110 is placed in front of said optical sensor 101. In several embodiments of the invention, the software apparatus 200 comprises artificial intelligence algorithms adapted LEIBO. / 33e2025 to automatically detect the presence of a fingertip on said surface (through the recognition of the images) and consequently carrying out an acquisition of the image of said fingertip for a spectral analysis. In several embodiments of the invention, the device comprises a door which covers the surface and which can be connected to the acquisition by the optical sensor 101 such that when the door is opened, the optical sensor 101 automatically triggers a series of acquisitions. Such acquisitions can be examined by artificial intelligence in order to eliminate those that are “empty” (before or after a user sets his / her fingertip on the surface) and those “moved” and usually keep those only with the best definition.

[0055] In several embodiments of the present invention, the system 100 comprises an optical sensor 101 in front which two independent photographic lenses 111 are placed in parallax, each with a focusing system 116 independent of each other and respectively sending the acquired electromagnetic waves towards a first filter 113 and a filter of Bayer 112 type and towards a second filter 114 and a filter of Bayer type 112. The two lenses 111 bringing the electromagnetic waves, each on half of said optical sensor 101.

[0056] In several embodiments of the present invention like that described in FIG. 3, the system 100 can comprise one or more additional filters 115. By way of a non-limiting or non-binding example, one or more additional filters 115 can be taken from a group comprising:

[0057] - a filter with transmittance greater than 90% for each wavelength comprised between 420 nm and 700 nm;

[0058] - a filter with transmittance greater than 95% for each wavelength comprised between 390 nm and 710 nm;

[0059] - a filter with transmittance greater than 90% for each wavelength comprised between 360 nm and 660 nm, with transmittance greater than 60% for wavelengths comprised between 300 nm and 360 nm and with transmittance greater than 35% for wavelengths comprised between 700 nm and 1000 nm;

[0060] - a filter with transmittance greater than 90% for each wavelength comprised between 800 nm and 1000 nm, with transmittance greater than 30% for wavelengths comprised between 420 nm and 700 nm and with average transmittance equal to about 15% for wavelengths LEIBO. / 33e2025 comprised between 300 nm and 420 nm;

[0061] - a filter with transmittance greater than 95% for each wavelength comprised between 630 nm and 880 nm;

[0062] - a filter with transmittance greater than 95% for each wavelength comprised between 800 nm and 1000 nm;

[0063] - a filter with average transmittance equal to about 25% for wavelengths comprised between 800 nm and 1000 nm;

[0064] - a filter with average transmittance equal to about 40% for wavelengths comprised between 300 nm and 400 nm and equal to about 5% for wavelengths comprised between 700 nm and 800 nm;

[0065] - a filter with average transmittance equal to about 50% for wavelengths comprised between 300 nm and 400 nm and transmittance greater than 25% for wavelengths comprised between 700 nm and 1000 nm.

[0066] In several embodiments of the present invention such as those shown in FIGS. 2-5, said multifilter apparatus 110 comprises at least a beam splitter 117 and at least a tilted mirror 118 adapted to separate the entering electromagnetic waves in a photographic lens 111 in two or more wave beams. Said wave beams each being directed towards one of said filters 113, 114 and towards said Bayer type filters 112. Specifically, FIG. 2 schematically shows the electromagnetic waves (three beams represented in light gray entering into the lens 111 for the three colors Red Green and Blue - RGB) and their path within the apparatus 110. The beam splitter 117, as is observed in FIG. 2, divides into two the light beams entering into the apparatus 110 and the tilted mirror 118 generates a second “channel” of rays parallel to the first (that in which the beam splitter 117) is placed. Each of both “channels” finish in one half of the optical sensor 101.

[0067] In some of the embodiments of the invention in which at least a beam splitter 117 and at least a tilted mirror 118 are present, like those shown in FIGS. 2-5, the focusing system 116 for said second filter 114, comprises mechanical components that produce the approach and / or moving away of a lens 10 to / from the tilted mirror 118, allowing an independent focusing relative to LEIBO. / 33e2025 each of the filters 113, 114. Specifically the focusing system 116 preferably has means (such as washers connected to telescopic tubes) which allow an overall focusing such to adjust the inlet focus through the photographic lens 111, and then said mechanical components for an independent focusing relative to each of the filters 113-115. Said “mechanical means” are deemed well known in the field of optics and photography and can be, by way of a non-limiting or non-binding example, analogous washers that make telescopic tubes slide between them, and which therefore produce an increase or a reduction of the distance between tilted mirror 118 and said lens 10.

[0068] In several embodiments of the present invention like that shown in FIG. 3, in which the system 100 comprises said additional filters 115 and said mechanical components of said focusing system in order to adjust the distance between a tilted mirror 118 and a lens 10. The beam splitters 117 and the tilted mirrors 118 are adapted to separate the electromagnetic waves into two or more wave beams. Said wave beams are directed by said tilted mirrors 118, each towards one of said filters 113-115 and towards said Bayer type filters 112. Said focusing system 116 for said filters 114, 115 comprises said mechanical components, allowing said independent focusing.

[0069] With specific reference to FIG. 3, a merely schematic representation is shown of a multi-filter apparatus 110 in which the entering light rays are divided, by two beam splitters 117, into three “channels” and directed (also by means of two tilted mirrors 118) respectively towards a first filter 113, a second filter 114 and an additional filter 115. FIG. 3 is to be intended a schematic figure since the position of the “channel” in which the additional filter 115 is placed is, in several preferred embodiments of the invention, orthogonal to the plane of the same FIG. 3. Divisions into a greater number of “channels” can be operated and are also preferably, one need only think of a division into four “channels” in which each channel occupies a quadrant of the optical sensor 101. FIG. 3 shows how each “channel” is in provided with a focusing system 116 thereof.

[0070] With reference to FIG. 4, an axonometric view is shown of a multi-filter apparatus 110, which at its interior integrates a beam splitter 117 and a tilted mirror 118, dividing the entering light LEIBO. / 33e2025 into two “channels” whose outputs are visible in the lower right part of FIG. 4 and which represents the windows that are then placed in contact with the optical sensor 101. FIG. 4 also shows a rigid base for the multi -filter apparatus 110. According to the present invention, indeed, in several embodiments, the system 100 employs said base in order to ensure stability in capturing images and / or videos and for also ensuring the capture of multiple images (one need only think of images with and without calibration target 130) in a manner so as to reduce as much as possible the differences between two acquired images of the same “analyzed to be analyzed”.

[0071] Specifically, in FIG. 4, a multi -filter apparatus 110 is shown which has an “overall” focusing system 116 placed immediately behind the lens 111, and a second focusing system 116 for the only channel in which a second filter 114 is placed, which can be accessed by means of a slit (shown in FIG. 4) present on the body of the apparatus 110.

[0072] In FIG. 4, a section plane is observed, marked with the letters A- A’ . The section attained by said section plane is shown in FIG. 5 in which it is possible to observe, in detail, the arrangement of the elements already described, within the multi -filter apparatus 110. FIG. 4 and FIG: 5 do not show said Bayer type filters 112 since, in these embodiments of the present invention, these are placed directly on the optical sensor 101.

[0073] With reference to FIGS. 4 and 5, knurled head screws are shown which, once unscrewed, allow extracting suitable supports for the first and second filters 113, 114. Said supports contain, in several embodiments of the present invention, also a photographic diaphragm placed in contact with said filters 113 and / or 114.

[0074] In several embodiments of the present invention such as that shown in FIGS. 1-5, the optical sensor 101 is that of a device (such as by way of a non-limiting or non-binding example: a reflex camera) from which a cut filter of the ultraviolet and infrared bands (UV-IR cut) is removed. In these embodiments as shown in FIGS. 2, 3 and 5, said multi-filter apparatus 110 further comprises an optical window 119 having a thickness suitable for restoring the optical path reduced by the removal of the cut filter of the UV-IR bands. Said optical window 119 therefore has a thickness suitable for restoring the optical path modified by the broadening of LEIBO. / 33e2025 the band of the optical sensor 101 and corresponding to the infinity focusing in the maximum wavelength that passes through said optical window 119. Said optical window 119 has a spectral transmittance greater than 80% for wavelengths at least comprised between 300 nm and 1000 nm and still more preferably the optical window 119 has a spectral transmittance greater than 85% for wavelengths comprised between 200 nm and 2500 nm. The optical window 119 has surfaces crossed by electromagnetic waves parallel to each other. Said parallel surfaces have smoothness such that the surface roughness have size (expressed in nanometers) less than a quarter of the smallest wavelength acquired. By way of example, in the event in which the optical sensor 101 acquires wavelengths comprised between 300 nm and 1000 nm, the size of the surface roughness on said surfaces of said optical window 119 must be less than 75 nm (300 nm / 4).

[0075] In several embodiments of the present invention like that shown in FIG. 5, said optical window 119 is made of quartz crystals.

[0076] In other embodiments of the present invention, said optical window 119 is made of calcium fluoride (CaFi).

[0077] In other embodiments of the present invention, said optical window 119 is made of molten silica. Other materials which are less suitable but which can be employed in particular embodiments of the present invention are germanium, zinc selenide (ZnSe), chalcogenide (AMTIR-1).

[0078] In several embodiments of the present invention, said calibration module 220 comprises secondary calibration algorithms adapted to allow the delimitation on the images of shaded areas. Said delimitation is operated by a user in a manual manner through a graphical interface which allows, for example, selecting zones with uniform color, or which in another example allows manually tracing a perimeter delimiting one or more shaded areas. Said secondary calibration algorithms requiring a user to report one or more points of a same body / object framed in the image, having the same color, inside and outside said shade zones. Said secondary calibration algorithms carrying out a secondary calibration, associating with the colors in said shade zones the relative colors outside of the shade zones and calibrated with LEIBO. / 33e2025 the calibration module 220. In this manner, it is also possible to have correct reflectance values per shade zone in the image which would otherwise require transformations and graphical processing for correcting the colors in only those zones.

[0079] In several embodiments of the present invention, said at least an optical sensor 101 is the sensor or a device suitable for recording photographs and videos. In this manner, the video recording allows the extrapolation of photograms which are then each analyzed, as described above.

[0080] In several embodiments of the present invention such as that shown in FIG: 6, said at least an optical sensor 101 belongs to a device that is mounted on a displacement apparatus; said device and said displacement apparatus being manually controllable on site and / or remotely. By “displacement apparatus” it is intended herein an apparatus adapted to ensure movement on the ground, in water, in the air, on ropes, belts and / or other means, of the device (which is, by way of a non-limiting or non-binding example, a camera / video camera).

[0081] An example of a displacement apparatus, as shown in FIG. 6 is a quadcopter drone on which the device with the optical sensor 101, the multi-filter apparatus 110 and the photographic flash 120 are mounted. Both the drone and the device are remotely switchable.

[0082] Another example of displacement apparatus is a frame on which four carriages slide, according to two main directions that are orthogonal to each other; such carriages are parallel two by two and are each connected with the device. Due to the simultaneous movement of a pair of carriages and to the independent movement of the pairs of carriages, the device can fully record a scene that takes place within the projection operated according to the direction perpendicular to the plane on which the device is moved. This type of displacement apparatus can be mounted at crime scenes in order to have the device perform an accurate scanning of a broad surface (e.g. a floor or a wall) on which forensic examinations are to be performed. In addition, this type of displacement apparatus can, by way of a non-limiting or non-binding example, be used for analyzing surface defects, for example on a planar metal plate of large size which is therefore accurately “scanned” (i.e. filmed or even recorded by means of a series of photographs) by the device. LEIBO. / 33e2025

[0083] Finally, it is clear that modifications, additions or variations that are obvious for a man skilled in the art can be made to the invention described up to now, without departing from the protective scope that is provided by the enclosed claims.

Claims

LEIBO. / 33e2025Claims1. System (100) for the acquisition and spectral analysis of images, characterized in that it comprises:- an optical sensor (101) capable of performing at least an image acquisition; said at least an optical sensor (101) having a capacity for acquiring electromagnetic waves with a wavelength at least between 300 nm and 1000 nm;- at least a multi-filter apparatus (110); said multi-filter apparatus (110) being connected to said optical sensor (101) and filtering the electromagnetic waves entering said optical sensor (101); said multi-filter apparatus (110) comprising, in turn: o at least a photographic lens (111); o at least Bayer type filters (112) for acquiring the wavelengths of the colors in the visible spectrum by mosaicing the acquired image; o at least a first filter (113) capable of passing electromagnetic waves with a wavelength at least between 300 nm and 600 nm with a transmittance greater than 50%; said first filter (113), coupled with said Bayer type filters (112) capable of providing three pieces of spectral reflectance information for each pixel of an image acquired after said mosaicization; o at least a second filter (114) capable of passing electromagnetic waves with a wavelength at least between 700 nm and 1000 nm with a transmittance greater than 50%, waves with a wavelength between 300 nm and 500 nm with a transmittance greater than 50% and waves with a length between 350 nm and 450 nm with a transmittance greater than 80% and waves with a length between 500 nm and 700 nm with a transmittance greater than 15%; said second filter (114), coupled to said Bayer type filters (112) capable of providing three spectral reflectance information for each pixel of an image acquired after said mosaicization; o at least a focusing system (116) for each filter (113, 114);LEIBO. / 33e2025- at least a Xenon photographic flash (120) with an average spectral radiance of at least 2 pW / (cm2 sr nm) for wavelengths at least between 300 nm and 1000 nm and with a spectral radiance greater than 1.3 pW / (cm2 sr nm) for wavelengths at least between 350 nm and 1000 nm; said photographic flash having a positive spectral radiance different from zero for wavelengths between 300 nm and 400 nm;- a calibration target (130) comprising colors and boxes whose reflectance, for each single color or white square, are known;- at least a software apparatus (200) suitable for performing a spectral analysis of the images acquired through the optical sensor (101) with the multi-filter apparatus (110) and the photographic flash (120); said software apparatus (200) comprising: o one or more databases (201) suitable for containing data regarding the spectral reflectance of elements to be examined; o a transformation module (210) suitable for automatically aligning said images; said transformation module (210) suitable for providing data relating to the automatic transformations performed; said data relating to the automatic transformations performed being manually modifiable; o a calibration module (220) in which the colors of the calibration target (130) within said image are associated with the relevant known reflectance; said calibration module (220) being suitable for calibrating the reflectance values of each pixel of said images on the basis of the associations between the image of the calibration target (130) and the relevant known reflectance; o a division module (230) for the spectrum of the acquired images into a plurality of sub-bands all having the same shape and the same percentage transmittance; for each sub-band, said division module (230) provides a unique reflectance value calculated on the basis of the input data; said unique reflectance value being provided for each pixel for the wavelength value, centered by said subband; o a comparison module (240) capable of performing a comparison betweenLEIBO. / 33e2025 individual pixels and / or groups of pixels of an image by providing the differences in the spectral reflectance for each of the sub-bands obtained by the division module (230); said comparison module (240) providing a graphical interface to a user for selecting individual pixels and / or groups of pixels to be compared and / or performing said pixel selection automatically by selecting pixels and / or groups of pixels present in two or more regions of the acquired image; o an analysis module (250) capable of performing graphical, numerical and / or statistical analyses on the spectral reflectance data of the bands obtained with said division module (230).

2. System (100), according to the previous claim 1, characterized in that it comprises an optical sensor (101) in front of which two independent photographic lenses (111) are placed, in parallax.

3. System (100), according to the previous claim 1 or 2, characterized in that it comprises one or more additional filters (115).

4. System (100), according to any of the previous claims, characterized in that said multifilter apparatus (110) comprises at least a beam splitter (117) and at least an tilted mirror (118) suitable for separating the electromagnetic waves entering a photographic lens (111) into two or more wave beams; said wave beams being each directed towards one of said filters (113, 114) and towards said Bayer type filters (112).

5. System (100), according to the previous claim 4, characterized in that said focusing system (116) for said second filter (114), comprising mechanical components that produce the approach and / or the separation of a lens (10) to / from the tilted mirror (118) allowing an independent focusing relative to each of the filters (113, 114).LEIBO. / 33e20256. System (100), according to the previous claims 3-5, characterized in that said beam splitter (117) and at least an tilted mirror (118) are suitable for separating the electromagnetic waves into two or more wave beams; said wave beams being each directed towards one of said filters (113, 114, 115) and towards said Bayer type filters (112); said focusing system (116) for said filters (114, 115) comprising said mechanical components allowing said independent focusing.

7. System (100), according to any of the preceding claims, characterized in that said optical sensor (101) is that of a device from which a cut filter of the ultraviolet and infrared bands (UV-IR cut) is removed; said multi -filter apparatus (110) further comprising an optical window (119) having a thickness suitable for restoring the optical path reduced by the removal of the cut filter of the UV-IR bands; said optical window (119) having a thickness suitable for restoring the optical path modified by the broadening of the band of the optical sensor (101) and corresponding to infinity focusing in the maximum wavelength that passes through said optical window (119); said optical window (119) having a spectral transmittance greater than 80% for wavelengths at least between 300 nm and 1000 nm; said optical window (119) having surfaces crossed by electromagnetic waves parallel to each other and having, said surfaces, surface roughness of size less than a quarter of the smallest wavelength acquired.

8. System (100), according to any of the preceding claims, characterized in that said calibration module (220) comprises secondary calibration algorithms suitable for allowing the delimitation of shaded areas on the images; said secondary calibration algorithms requiring a user to report one or more points of the same body / object framed in the image, having the same color, inside and outside said shaded areas; said secondary calibration algorithms performing a secondary calibration, associating the colors in said shaded areas with the relative colors outside the shaded areas and calibrated with theLEIBO. / 33e2025 calibration module (220).

9. System (100), according to any of the preceding claims, characterized in that said at least an optical sensor (101) is the sensor of a device suitable for recording photographs and videos.

10. System (100), according to any of the preceding claims, characterized in that said at least an optical sensor (101) belongs to a device that is mounted on a displacement apparatus; said device and said displacement apparatus being manually controllable on site and / or remotely.

Images