A method for generating a preview image of an ocular surface

Abstract

The present invention discloses a method for obtaining a preview image of ocular surface. The preview image is obtained by illuminating the ocular surface with a light source, such as a laser or a diode, configured to emit light at a narrow spectral band, having narrow spectral bandwidth. This allows to obtain a surface image. More precisely, the method according to the invention comprises the following steps: a) illuminating the ocular surface with a light source configured to provide light within a wavelength spectral band restricted during exposure to a given spectral bandwidth; capturing with a detector at least one frame within the predefined spectral bandwidth; c) processing the captured frames using Fourier transformation and spatial filtering to generate the preview image of the ocular surface. The predefined spectral bandwidth is up to 10 nm.

Description

[0001] A method for generating a preview image of an ocular surface

[0002] TECHNICAL FIELD

[0003] The present invention relates to a method for preview surface imaging in fullfield eye imaging system.

[0004] BACKGROUND OF THE DISCLOSURE

[0005] The field of ophthalmic imaging has witnessed significant advancements over the past few decades, driven by the need for precise, non-invasive diagnostic tools capable of capturing detailed images of the eye's internal structures. Among these, full-field (FF) imaging technique offers the ability to capture comprehensive volumetric data of the eye in a single acquisition. Unlike traditional scanning methods using point detectors, FF imaging utilizes a two-dimensional (2D) light detector to illuminate and capture the entire eye area simultaneously, and relies on capturing the backscattered signal from said entire volume at once. This approach not only increases image quality and enhances imaging speed, but also reduces motion artifacts and crosstalk noise, which are common challenges in pointscanning systems. To achieve said effects, typically a number of image volumes is being taken for later averaging or other numerical operations.

[0006] Despite the advantages mentioned above, FF imaging systems face certain limitations, particularly in providing real-time preview images for operators. In the FF imaging setup it is not practical to extract an imaged area preview picture from the standard data for the operator, which could be used in decision-making and choosing the region of interest. In known FF setups, generating a preview image requires processing the same volumetric data as used for the actual measurement, which is computationally intensive and time-consuming, and there is no faster path for preparing the preview image. As a result, the image processing speeds fall below the video rate speeds and thus make the preview generation process too slow and impractical for real-time decision-making during examinations. This limitation is further exacerbated in high-resolution microscope quality systems, where the imaged area is confined to a limited region of the retina, thereby failing to offer a comprehensive overview necessary for selecting regions of interest.

[0007] To address aforementioned challenges by providing lower-quality preview data, a solution has been proposed to integrate a parallel setup in a typical Optical Coherence Tomography (OCT) scanning mode into existing FF systems. While this approach offers a potential solution, it significantly increases the device's size, cost, and complexity, making it less feasible for widespread clinical use.

[0008] Another approach to solve the aforementioned problems involves incorporating a fundus camera into the FF system to obtain en-face images of the retina for preview purposes. Fundus cameras are designed to capture detailed images of the retina using specialized optical configurations, such as annular illumination mode, to minimize reflections from the eye's surfaces. However, integrating a fundus camera into an FF system introduces additional complexities, including the need for precise alignment and the challenges associated with annular illumination, which requires careful engineering to avoid central light reflections that can obscure retinal imaging. Moreover, traditional fundus cameras often necessitate pharmacological pupil dilation to evaluate the retinal periphery effectively.

[0009] Document US20130250240A1 discloses a digital holographic adaptive optics (DHAO) system to replace hardware components in a conventional AO system with numerical processing for wavefront measurement and compensation of aberration by the principles of digital holography. The method comprises the steps of: providing a narrow laser beam entering an eye through the cornea and lens, forming a focused spot on the retina; capturing and numerically storing the wavefront profile of the emergent beam by digital holography; providing a flood illumination to the retina; capturing and numerically storing the wavefront profile of the emergent light by digital holography; and numerically combining and processing the holograms to obtain an aberration-compensated image of the retina. This approach is faster than conventional AO because it does not involve feedback and iteration, and the dynamic range of deformation measurement is essentially unlimited. This system can be incorporated into a conventional fundus camera and it can generate profiles of the retinal vasculature and measure blood flow. Among non-patent literature, publication In-vivo retinal imaging with off-axis full-field time-domain optical coherence tomography (Optics Letters, Vol. 41 , Issue 21 , pp. 4987-4990, 2016) discloses a volumetric tissue imaging system mainly composed of a low coherence light source and a camera, full-field optical coherence tomography (FF-OCT). It introduces path-length differences between the reference and the sample light in neighbouring pixels using an off-axis reference beam. In this solution, the temporal carrier frequency in scanned time-domain OCT is replaced by a spatial carrier frequency.

[0010] While the methods described above allow high-resolution imaging of the retina, their configurations are complex and do not allow for obtaining rapid previews necessary for dynamic clinical environments.

[0011] In light of the considerations presented above, there is a pressing need for an innovative solution that enables fast preview imaging within FF systems without compromising their inherent advantages. Such a solution would not only enhance the efficiency of clinical workflows but also improve diagnostic accuracy by allowing for more precise targeting of regions of interest.

[0012] SUMMARY OF THE INVENTION

[0013] According to the invention, there is provided a method for generating a preview image of an ocular surface using a full-field ophthalmic imaging system, the method comprising the following steps: a) illuminating the ocular surface with a light source configured to provide light within a wavelength spectral band restricted during exposure to a given spectral bandwidth; b) capturing with a detector at least one frame within the predefined spectral bandwidth; c) processing the captured frames to generate the preview image of the ocular surface, wherein the predefined spectral bandwidth is up to 10 nm. Preferably, in step a, the light source is configured to sweep across the spectral band restricted to the predefined spectral bandwidth during exposure.

[0014] The light source configured to sweep across the spectral band may be a tunable laser.

[0015] Sweeping over the selected spectral band may be performed more than once. In such case, sweeping may be performed at least twice over the same spectral band or over different spectral bands.

[0016] More specifically, in one example, after completing one sweep over the selected spectral band, the subsequent sweep is repeated over the same spectral band.

[0017] In another example, after completing one sweep over a first spectral band, the subsequent sweep is performed over a second spectral band, different from the first spectral band.

[0018] Alternatively to the light source configured to perform sweeping, the light source may be configured to provide a continuous fixed spectral band during the exposure.

[0019] In such case, the light source continuously emits the light in the spectral band restricted to the predefined spectral bandwidth during exposure.

[0020] Preferably, such a light source used is a diode, such as a superluminescent diode or a light-emitting diode.

[0021] Alternatively to the light source continuously emitting the light in the selected spectral band, the continuous spectral band may be provided by the light source emitting a broadband wavelength range, the broadband light source being combined with a filter transmitting the spectral band restricted to the predefined spectral bandwidth.

[0022] Preferably, the filter used is a bandpass filter, an interference filter or monochromator.

[0023] Preferably, the spectral band falls within a spectral range of 780 nm to 900 nm. However, it may be broader, e.g. up to 1300 nm.

[0024] Preferably, the predefined spectral bandwidth is up to 2 nm. Preferably, the predefined spectral bandwidth is even up to 1 nm.

[0025] Preferably, the detector used is a CMOS camera or a CCD camera.

[0026] Preferably, in step b at least two frames are captured, and step c further includes temporal averaging of multiple captured frames.

[0027] BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The disclosure of the present invention will now be discussed, by way of example only, with reference to the accompanying drawing in which:

[0029] Figure 1 shows comparison of measured volume depth ranges between broad- and narrow-spectrum illumination conditions.

[0030] DETAILED DESCRIPTION OF THE DISCLOSURE

[0031] The present invention aims to provide a fast preview mode in known full-field (FF) eye imaging systems, and addresses this need by introducing a novel method for fast speckle-less preview imaging using the hardware in digital holographic fundus camera mode in such FF ophthalmic systems. It eliminates the reliance on additional imaging modalities or complex additional configurations, thereby preserving the simplicity and affordability of FF systems. Furthermore, the invention ensures that the preview images encompass a sufficiently wide field of view, providing operators with a comprehensive overview of the retina to inform their diagnostic decisions.

[0032] A digital fundus camera is an ophthalmic imaging device designed to capture detailed images of the retina, optic disc, macula, and posterior pole. Its structure facilitates non-invasive visualization of the eye's interior, aiding in the diagnosis and monitoring of various ocular conditions.

[0033] The core of the fundus camera's optical system is based on the principle of monocular indirect ophthalmoscopy. It employs a specialized lens assembly to focus light onto the retina and capture the reflected image. The illumination and observation paths are designed to be coaxial and separate, minimizing reflections and optimizing image clarity. Typically, an annular illumination system directs light through the periphery of the pupil, while the central axis is reserved for image capture, reducing corneal reflections. A controlled light source, such as a xenon flash or LED, provides the necessary illumination for imaging. In non-mydriatic cameras, infrared light is used for alignment and focusing, allowing for imaging without pharmacological pupil dilation. Once alignment is achieved, a visible light flash captures the image. The reflected light from the retina is captured by a high- resolution digital sensor, such as a CCD or CMOS sensor. The digital data is then processed to produce a detailed image of the fundus. Advanced software algorithms enhance image quality, adjust exposure, and facilitate analysis.

[0034] The integration of digital holography into fundus imaging enables the capture of three-dimensional information of retinal structures. A coherent light source is split into the object beam, which illuminates the retina, and the reference beam. The light reflected from the retina interferes with the reference beam, and the resulting hologram is captured by a digital sensor. The system thus records the interference pattern between a reference beam and the light reflected from the retina. This interference pattern, or hologram, contains both amplitude and phase information, allowing for numerical reconstruction of the retinal image in three dimensions using computational algorithms. Implementing holographic mode in fundus cameras involves incorporating coherent light sources and interferometric setups into the existing optical design. Advancements in digital sensors and computational power have made it feasible to integrate holographic imaging into compact, clinical-grade fundus cameras.

[0035] Fundus cameras typically use a range of visible light wavelengths, from approximately 380 to 700 nm. Some cameras also employ near-infrared wavelengths for specific imaging applications. In holography mode spatially coherent light is used, referring to light waves that maintain a constant phase relationship across different points in space that are perpendicular to the direction of propagation. A light source producing such spatially coherent light for the purpose of the invention is for example a laser. Thanks to the use of spatially coherent light, a controlled phase in the entire image is obtained. As mentioned above, a signal is transmitted simultaneously to both the eye and a reference mirror. The portion of the signal directed to the eye undergoes a phase change upon interaction with ocular structures and is subsequently reflected or returned. The portion directed to the reference mirror is reflected without undergoing any phase alteration. The interference between the phase-modified signal from the eye and the unmodified reference signal enables the determination of the phase change introduced by the eye. This phase information can be used to extract structural or functional data about the eye.

[0036] Typically, in standard FF measurement the light wavelength sweep range is few tens of nm, for example 25 nm or more, even up to 80 nm. This broad wavelength sweep range allows to obtain detailed volumetric information for producing 3D images, but is time-consuming due to large amount of data to be processed, and therefore cannot be used in fast preview imaging.

[0037] According to the invention, fast preview images are obtained by employing a narrow wavelength spectral band in FF measurement system, restricted to narrow spectral bandwidth. The spectral band is selected from a broader spectral range suitable for imaging of the eye.

[0038] Narrow wavelength spectral band, in one embodiment, may be a sweep range which is swept by the light source (such as a laser) during the measurement. Sweep range is the range of optical wavelengths over which a light source (e.g. a tunable laser) is swept during measurement. In another embodiment, the light source may provide the selected narrow spectral range continuously, without sweeping. This issue will be explained in detail later in the description.

[0039] The broader is the spectral band, the finer is the axial resolution (depth resolution). This is schematically shown in Figure 1 , depicting comparison of measured volume depth ranges between broad- and narrow-spectrum illumination conditions, causing the difference in data amount needed to be processed between these two cases. As can be seen, in narrow-spectrum illumination conditions the depth is reduced, leading to obtaining the image of the surface instead of the volumetric image. Axial resolution Az is inversely proportional to the bandwidth: where Aois central wavelength, and AX is spectral band (e.g. sweep range), defined by a parameter known as full width half maximum, FWHM. Generally, longer central wavelengths (e.g. 950 nm) are often used for deep tissue imaging, while shorter central wavelengths allow for less deep penetration.

[0040] The wavelength spectral band (sweep range) used corresponds to the bandwidth of spectral decorrelation of speckle, i.e. where:

[0041] AX - bandwidth required for speckle decorrelation

[0042] X - central wavelength n - refractive index (typically assumed 1 .3-1 .4)

[0043] L - optical path difference (caused by the system geometry or surface roughness).

[0044] Using the illumination light with the wavelength bandwidth equal to AX is sufficient to obtain a good quality preview image. For example, in retina imaging at 850 nm, L = 0.1 mm, n = 1.4, AX equals approximately 0.8 nm. In cornea imaging AX equals approximately 0.5 nm.

[0045] The wavelength spectral range used in the measurement system preferably falls within the range of 780 to 900 nm, but may be more broad, for example up to 1300 nm. Therefore, in selected exemplary implementations, the wavelength spectral band may be: from 780 nm to 790 nm, from 800 nm to 810 nm, from 850 nm to 855 nm, from 850 nm to 851 nm, etc.

[0046] Limiting the wavelength sweep range causes a drastic drop in the imaging depth, typical for OCT, and an image of the surface instead of the volumetric image is obtained. As a result, the amount of data for analysis decreases significantly, which allows to minimize the computation time needed to perform before presenting the finished image to the operator. While the data set obtained lacks depth information, the image quality on the retinal surface remains high, and with the narrowing of the spectrum, the depth of the 3D cube decreases, as shown in Fig. 1 . Therefore, while employing such narrow sweep range conditions, no volumetric (depth) information is obtained, but rather an en-face surface image can be quickly produced. Such a surface image is suitable to be presented as a preview information image.

[0047] Generally, the ocular surface is illuminated with a coherent light source, such as a laser, configured to emit light at a narrow spectral band. Therefore, during the exposure, the wavelength spectrum is restricted to a narrow spectral band, such that at any given moment of exposure. Narrowing the wavelength spectrum during exposure can be achieved in two ways, both aiming to ensure that, at any point in time during the exposure, only a narrow spectral band contributes effectively to the obtained image.

[0048] For the sake of clarity, it should be noted that in the context of the present invention, the term “spectral bandwidth” refers to the width of the selected spectral band (such as e.g. 3 nm); the term “spectral band” refers to the selected range of wavelengths over which the bandwidth extends (e.g. 850 nm to 852 nm), and the term “spectral range” is a broader wavelength interval from which the spectral band may be selected (e.g. 780 nm to 900 nm).

[0049] The specific illumination protocols employed during the preview mode may include, by way of example and without limitation, the following configurations.

[0050] In one embodiment, wavelength sweeping is employed. The source, typically a tunable laser or a broadband source with tunable filtering, sweeps and scans over a narrow range of wavelengths, i.e. over a narrow spectral band. The wavelength sweep range is a spectral band having a selected bandwidth. The spectral bandwidth used in the method according to the invention should not exceed 10 nm. However, preferably, the sweep range is less than 2 nm, or even less than 1 nm. The wavelength spectral band (sweep range) can be even as narrow as 0,1 nm, or even narrower, or may be of any value up to 10 nm. Preferably, it is for example from 0,5 up to 1 ,5 nm. Therefore, in one embodiment, the feature leading to switching to the fast preview mode occurs in the programmatically selected, narrow laser wavelength sweep range. The sweep range depends on the actual illumination conditions and the sample properties (layer thickness and refractive index).

[0051] In this embodiment, the system employs a tunable laser (already existing in the FF imaging system), and it is possible to limit the tuning range during each exposure cycle to a narrow spectral window. This results in the desired effect of a spectrally narrow illumination during each frame or exposure period.

[0052] For example, a single frame may be recorded while the laser is swept across a narrow wavelength spectral bandwidth, such as 0,8 nm. Once this sweep is completed and the frame is captured, the system can proceed to acquire the next frame. This can be done in one of two ways:

[0053] 1 ) Repetitive narrow sweep

[0054] After completing the sweep (e.g., 0,8 nm from Ai to A2), the laser is returned to the starting wavelength (Ai), and the same sweep is repeated to capture the next frame. In this case, the central wavelength is fixed.

[0055] 2) Progressive sweep over adjacent ranges:

[0056] The laser can continue sweeping over a different, adjacent narrow wavelength range (e.g., e.g., 0,8 nm from Ai to A2 and e.g., 0,8 nm from A2 to A3), thereby advancing continuously without the need to reverse tuning direction. This method speeds up acquisition since no time is lost returning the laser to the starting point. It also allows the system to sequentially cover a broader spectrum, frame by frame, while still maintaining narrowband conditions at each moment of exposure. In this case, the central wavelength is varied. For example, if selected wavelength spectral band is 10 nm, the first measurement may be performed at the spectral range of 800 to 810 nm, and the subsequent measurement may be performed at the spectral range of 810 to 820 nm, etc.

[0057] Therefore, in case of wavelength sweeping, the possible variants include repetitive tuning over the same narrow range during the exposure period, or switching I tuning between different narrow ranges or discrete wavelengths. In both cases, the key principle is that at any point in time during the laser operation and exposure, the instantaneous spectral content remains within a narrow range. Moreover, in both cases, the obtained frames may be displayed separately or may be averaged before displaying.

[0058] The central wavelength of the narrow sweep range, in a non-limiting example, may be for example 850 nm.

[0059] The exposure time for each individual frame captured during sweeping is even as short as ca. 1 ms. This leads to significant benefits, such as faster individual frame capture and less data per frame compared to standard solutions (even 10x less), what simplifies post-processing and storage.

[0060] In each sweeping range, at least 1 frame is captured. Preferably, if more than 1 frame is captured, for example 2 frames, the frames are averaged.

[0061] In another embodiment, which represents an alternative approach to achieving a narrow wavelength spectrum during exposure, maintaining exposition at a narrow spectral band may be realised by performing the exposure using light confined to a narrow, fixed spectral band throughout the exposure duration. This embodiment is distinct from the embodiment involving narrowband wavelength sweeping or tuning (involving temporal wavelength variation), since it utilizes the entire narrow spectral band simultaneously — i.e. , the full narrow band is present and used at once during the exposure.

[0062] In this embodiment, a light source configured to emit solely within a predetermined narrow spectral range may be utilized. Such a light source may be selected to have a bandwidth sufficiently narrow to meet the spectral requirements of the exposure application, for instance, 10 nm or less, e.g. 1 nm.

[0063] A light source that emits only a narrow fixed spectral band throughout the exposure duration may be used. The light source may be a narrowband-emitting semiconductor device such as a diode, e.g. a light-emitting diode (LED) or a superluminescent diode (SLD).

[0064] Alternatively, in such embodiment, a broadband light source may be used, together with a filter which transmits only a narrow fixed spectral band. A broadband light source may be a halogen lamp or xenon arc lamp, employed in conjunction with an optical filtering system. The filtering system may include, but is not limited to, bandpass filters, interference filters, or monochromators, which are configured to selectively transmit only a narrow portion of the source spectrum corresponding to the desired exposure wavelength. These filters may be fixed in position and spectral characteristics to ensure stability and consistency of the spectral band during the entire exposure interval. The optical filter may be integrated directly into the illumination path or in proximity to the exposure target (eye).

[0065] The light source emits a narrow spectral band all at once, without requiring active tuning or sweeping. Instead of using the tunable laser to scan through a narrow wavelength range, a fixed narrowband light source — such as said superluminescent diode (SLD) or light-emitting diode (LED) — can be added to the system. These sources can be chosen or filtered to emit only the desired narrow spectral range (e.g., up to 10 nm bandwidth), and they do so simultaneously across the full band, without temporal modulation. Thereby, there is provided continuous illumination in a selected range.

[0066] Because the diode or the broadband source with a filter does not require wavelength scanning, the exposure time is further reduced. This approach effectively simulates the laser’s narrow sweep.

[0067] In this embodiment, at least one image frame is captured, however capturing multiple frames, for example 10 frames, allows to increase the image quality. Thus, it is possible to use a number of frames and average them for increased image quality or using just one frame to maximize the refresh rate. Temporal averaging of consecutive frames enhances signal-to-noise ratio.

[0068] The narrow spectral band, in one of the non-limiting examples, may be 2 nm in a wavelength range from 849 to 850 nm.

[0069] When the preview mode according to the invention is finished and the actual measurement is started, the diode is turned off and the laser is turned on with full tuning width.

[0070] Because the narrow spectral band is emitted continuously, there is no delay from laser tuning. The camera can begin acquiring data immediately, which significantly reduces the total acquisition time. Thus, in this embodiment, the system switches from spectral scanning (as in the first embodiment) to simultaneous narrowband illumination — achieving the same effect of narrow spectral exposure, but through different means.

[0071] In all embodiments, the used wavelength band, either wavelength sweep range or fixed bandwidth, should generally be not broader than 10 nm. The wavelength range (sweep range or fixed band) is preferably within a range of 780 to 900 nm, and in such case it is preferable that the bandwidth does not exceed 2 nm. In some embodiments, such as using the wavelength regime of 1300 nm, the preferable bandwidth may exceed 2 nm. The bandwidth is selected and adjusted depending on the thickness of the layer being imaged and its refractive index.

[0072] The process of obtaining a preview image according to the invention, regardless of the embodiment, is represented by the following steps: a) illuminating the ocular surface with a light source configured to provide light within a given, narrow wavelength spectral band, at a given wavelength band; b) capturing with a detector at least one frame within the given narrow wavelength spectral band; c) processing the at least one captured frame to generate a surface preview image. In this step, processing is preferably done using Fourier transformation and spatial filtering, in a known manner.

[0073] In step a, the wavelength spectral band is narrow, which means that the wavelength spectral band is up to 10 nm. Preferably, it is even less than 1 nm, e.g. 0,8 nm.

[0074] Illumination by a narrow spectral band may be achieved by wavelength sweep through the given spectral band or by using a light source that provides light in this narrow spectral band, such as a LED or SLD diode, or a broadband source combined with a filter, as already explained above.

[0075] The wavelength band, either swept or a narrow spectrum, falls for example within a spectral range of 780 nm to 900 nm. However, the range may be different and may vary according to the needs - for example, the range may be up to 1300 nm. Step b performed at least once. In case of laser sweeping, once on sweep in step b is completed and the frame is captured, the system can proceed to acquire the next frame. This can be done in one of two ways, using repetitive narrow sweep or using progressive sweep over adjacent ranges. In repetitive sweep, the central wavelength of the given narrow wavelength spectral range remains the same during subsequent measurements. In progressive sweep, the central wavelength of the given narrow wavelength spectral range is changed in subsequent measurements.

[0076] The detector used in the present method, for either of the embodiments, is for example an ultrafast CMOS or CCD camera.

[0077] In holographic imaging arrangement, the strong back-reflection is still present in the raw data. These reflections can interfere with the desired signal by introducing artifacts and noise into the recorded holograms. This can degrade image quality and obscure vital retinal details. Due to the phase information available in the data, the excess reflection signal can be easily removed from the picture in fast numerical postprocessing by spatial filtering of the data after Fourier transformation, where the Fourier transformation converts the data into spatial frequency domains. Utilizing spatial filters helps to isolate the desired signal from unwanted reflections, by suppressing artifacts from strong back-reflections using bandpass filters. Transforming the hologram into the frequency domain allows for the separation of the desired signal from noise and artifacts. Therefore, no special illumination shape is needed and the FF system modules can be reused without any modifications.

[0078] Unlike conventional methods, the present invention incorporates dynamic laser tuning, wavelength-specific modulation of illumination, and advanced numerical post-processing techniques. Using the holographic fundus camera mode makes it possible to obtain a fast preview-class image in the FF eye imaging device without applying any additional hardware or mechanical modifications, by just modifying the illumination mode which is adjustable on the light source driving software level. This allows to obtain preview image and the actual measurement image using the same setup.

Claims

CLAIMS1. A method for generating a preview image of an ocular surface using a fullfield ophthalmic imaging system, the method comprising the following steps: d) illuminating the ocular surface with a light source configured to provide light within a wavelength spectral band restricted during exposure to a given spectral bandwidth; e) capturing with a detector at least one frame within the predefined spectral bandwidth; f) processing the captured frames to generate the preview image of the ocular surface, wherein the predefined spectral bandwidth is up to 10 nm.

2. The method according to claim 1 , wherein in step a, the light source is configured to sweep across the spectral band restricted to the predefined spectral bandwidth during exposure.

3. The method according to claim 2, wherein the light source is a tunable laser.

4. The method according to claim 2 or 3, wherein step b is performed at least twice, and wherein after completing one sweep over the spectral band, the subsequent sweep is repeated over the same spectral band.

5. The method according to claim 2 or 3, wherein step b is performed at least twice, and wherein after completing one sweep over a first spectral band, the subsequent sweep is performed over a second spectral band, different from the first spectral band.

6. The method according to claim 1 , wherein in step a, the light source is configured to provide a continuous fixed spectral band during the exposure.

7. The method according to claim 6, wherein the light source continuously emits the light in the spectral band restricted to the predefined spectral bandwidth during exposure.

8. The method according to claim 6 or 7, wherein the light source used is a diode, preferably a superluminescent diode or a light-emitting diode.

9. The method according to claim 6, wherein the light source emits a broadband wavelength range, and wherein the broadband light source is combined with a filter transmitting the spectral band restricted to the predefined spectral bandwidth.

10. The method according to claim 9, wherein the filter used is a bandpass filter, an interference filter or monochromator.

11. The method according to any of the preceding claims, wherein the spectral band falls within a spectral range of 780 nm to 900 nm.

12. The method according to any of the preceding claims, wherein the predefined spectral bandwidth is up to 2 nm.

13. The method according to any of the preceding claims, wherein the predefined spectral bandwidth is up to 1 nm.

14. The method according to any of the preceding claims, wherein the detector used is a CMOS camera or a CCD camera.

15. The method according to any of the preceding claims, wherein in step b at least two frames are captured, and wherein step c further includes temporal averaging of multiple captured frames.

Images