An imaging device and a method to reconstruct an image

Abstract

An imaging device and a method to reconstruct an image are proposed. The device comprises an array of individually addressable light-emitting devices configured to generate light at one or more wavelengths that is projected onto a sample to be imaged; a single-pixel detector configured to detect the light generated by the array of light-emitting devices to reconstruct an image of said sample; and an electronic controller configured to control the array of light-emitting devices. An optical conditioning element is arranged at a given distance from the array to ensure that light from device illuminates a specific and distinct region of the sample. The light-emitting devices are configured to perform spatio-temporal encoding by each one illuminating its respective distinct region of the sample but all generating the light at the same time and being temporarily modulated at a different and unique frequency, using encoding signals generated by the electronic controller.

Description

[0001] AN IMAGING DEVICE AND A METHOD TO RECONSTRUCT AN IMAGE

[0002] Technical Field

[0003] The present invention generally relates to imaging systems and methods. More specifically, the invention relates to an ultrafast, or high-speed, single-pixel imaging device and to a method to reconstruct an image using such imaging device.

[0004] Background of the Invention

[0005] One of the key challenges in optical technologies is maximizing the amount of spatiotemporal information that can be retrieved from a sample. Efforts in this direction include state-of-the- art optical systems based on cameras that feature millions of pixels. While providing high- spatial resolution, cameras typically have a response time of milliseconds, strongly limiting imaging speed to some hundreds of frames-per-second (fps). This speed is often insufficient to extract sound conclusions from dynamic events as relevant as those that take place in living biological systems or in fast-moving industrial production lines.

[0006] Several techniques have been developed to address this issue. A widely extended solution is to use the so-called ’’ultrafast cameras”, that is, specially designed cameras capable of capturing events at speeds of tens of thousands of fps. However, they can be costly, bulky, and difficult to integrate in most common optical architectures. In addition, they are designed to operate in the visible spectrum of light. Note the growing interest for cameras with infrared (IR) capabilities, thanks to the unique advantages they offer, ranging from retrieving the chemical fingerprint of materials to night vision, or deep penetration in biological tissue. As of today, no ultrafast camera exists capable of operating with IR wavelengths.

[0007] Alternatively, it is possible to use cameras that feature a single pixel and exhibit a response time of nanoseconds or below, thus providing a high temporal resolution. A paradigmatic example is a simple photodiode. In this case, light can be detected at rates of millions of fps, but because the whole spatial information is merged into a single pixel, these cameras have effectively no spatial resolution. Importantly, by sending encoded-illumination patterns, it has been demonstrated that spatial information can be retrieved from a single-pixel camera [1- 3], In these instances, a spatial-light modulator device [2] or acousto-optic deflectors [1 ,3,4] are normally used to provide the necessary illumination encoding. Unfortunately, these devices cannot be easily integrated into existing optical systems, and require complex alignment and very high maintenance. Steps toward reducing the complexity of single-pixel imaging devices include exploiting computational ghost imaging schemes. They are based on sequentially illuminating a sample with different light patterns, and then collecting the emitted, transmitted, or reflected light with a single pixel detector. By using an LED array as an illumination module, significant improvement of the structured illumination rate in the computational ghost imaging system can be achieved compared to the traditional digital micromirror device [5], In particular, by properly designing an LED array driving circuit, the full use of the LED switching time can be obtained, providing display rates as high as 12.5 MHz [6], As such, images with 32 x 32 pixel resolution can be reconstructed at about 3,000 fps - corresponding to about 4096 sequences. Furthermore, these schemes are compatible with compressive algorithms that can further reduce the number of illumination sequences needed, albeit with the potential of resulting in image artifacts. Still, the sequential nature of the approach is not optimal for high-speed imaging applications. Indeed, imaging speed is the same as when sequentially switching the different LEDs of the array, one at a time. Also, the decoding process required to reconstruct the images can amplify noise, which results in no appreciable gain in signal-to-noise ratio compared to individual LED switching.

[0008] CN108366195-A discloses a high-speed spatial light modulation method based on an LED array and an imaging system of the method. The system comprises a computer, a control board card, an LED array, a light beam modulation lens, an object, a light beam collection lens, and a single-pixel photoelectric detector. The application method comprises the following steps: firstly, the LED array generates Hadamard patterns and stores the Hadamard patterns in the control panel card; secondly, the control board card controls the LED array to display the Hadamard patterns in sequence through scanning twice, and collects a voltage average value of a process of displaying the first scanning pattern and the second scanning pattern; and finally, the computer sums the products of the Hadamard patterns and the corresponding voltage average values to restore an image. For the system to be quick, the LEDs in the array need to be either completely off or completely on. More importantly, since the patterns must be scanned twice, the imaging speed — without the use of compressive algorithms — is reduced to half compared to sequentially switching the individual LEDs.

[0009] US2017059408-A1 discloses methods and devices for generating multispectral illuminating light having an addressable spectrum. The gist of the approach consists of converting a monochromatic camera - featuring multiple pixels - into a multi-color camera by encoding color information into temporal frequencies. The illuminating device, thus, comprises a multispectral light source and a modulator to temporally modulate each color (from a monochromatic or a quasi-monochromatic source) with a different and unique frequency. The whole sample of interest is then illuminated with all the colors simultaneously, and by capturing the information with the monochromatic camera, it is possible to decode color information and reconstruct a multispectral image. Note, though, that the speed of this approach is only a fraction of the camera frame rate. More importantly, the number of pixels of the final image is exclusively given by the number of pixels of the monochromatic camera. Therefore, it is fundamentally impossible to use this approach to reconstruct images with a single-pixel imaging system or detector - the reconstructions would contain only one pixel of spatial information, which does not constitute an image.

[0010] CN109770849-A discloses a single-pixel, multi-spectral SFDI system based on phase-locking photon counting and compressed sensing. The overall idea is to perform SFDI with spatial illumination encoding using a space light modulator (a digital micromirror device, for instance). To retrieve a multi-spectral SFDI image, the system uses the same principle as US2017059408-A1 , illuminating the whole sample with multiple encoded colors. Thus, the system comprises a multi-path square signal generator, a modulated LED laser device, a space light modulator, a single-pixel camera, and a computer. The modulated LED laser device comprises a plurality of LED laser device bodies with different wavelengths and a current driver of the laser device bodies. All paths of modulated lasers are collected into the space light modulator through coaxial cables, light beams output by coaxial optical fibers are expanded into wide-field light through a beam expander, a modulated light source at the specific space frequency generated by the space light modulator is projected to the tissue body surface through a lens, and the single-pixel camera comprises a compressed encoding module and a signal collecting module.

[0011] Other single-pixel imaging systems are known by CN115631111-A, CN109770849-A, W02020007078-A1 , and CN115187765-A1.

[0012] Most of these prior art solutions lack individually addressable light-emitting devices. Instead, these systems use light-emitting devices to project an image with content that is refreshed periodically, line by line. Likewise, the prior art solutions do not disclose an array of lightemitting devices capable of encoding spatial information, where each device is temporarily modulated at a different and unique frequency, with each light-emitting device illuminating a specific and distinct region of the sample.

[0013] New or alternative ultrafast single-pixel imaging devices and image reconstruction methods are therefore needed. References:

[0014] 1. Mikami, H. et al. Ultrafast confocal fluorescence microscopy beyond the fluorescence lifetime limit. Optica 5, 117 (2018).

[0015] 2. Edgar, M. P., Gibson, G. M. & Padgett, M. J. Principles and prospects for singlepixel imaging. Nat. Photonics 13, 13-20 (2019).

[0016] 3. Diebold, E. D., Buckley, B. W., Gossett, D. R. & Jalali, B. Digitally synthesized beat frequency multiplexing for sub-millisecond fluorescence microscopy. Nat. Photonics 7, 806-810 (2013).

[0017] 4. Tsyboulski, D., Orlova, N., Ledochowitsch, P. & Saggau, P. Two-photon frequency division multiplexing for functional in vivo imaging: a feasibility study. Opt. Express 27, 4488 (2019).

[0018] 5. Zi-Hao Xu et al. 1000 fps computational ghost imaging using LED-based structured illumination. Opt. Express 26 Issue 3, pp. 2427-2434 (2018).

[0019] 6. Hongxu Huang et al. 25,000 fps Computational Ghost Imaging with Ultrafast Structured Illumination. MDPI (2022). otion of the Invention

[0020] To that end, the object of the present invention is to provide a new ultrafast single-pixel imaging device, i.e. and ultrafast camera, and a method to provide the spatial illumination encoding necessary to reconstruct an image with said imaging device.

[0021] This object is fulfilled by a device with the characteristics of claim 1 and by a method with the features of claim 11 .

[0022] Embodiments of the present invention provide, according to one aspect, an imaging device comprising: an array of individually addressable light-emitting devices (e.g. light emitting diodes (LEDs) or lasers) to generate light at one or more wavelengths that is shed or projected onto a sample to be imaged; at least one optical conditioning element arranged at a given distance from the array of light-emitting devices to ensure that light from each light-emitting device illuminates a specific and distinct region of the sample, the specific and distinct region of the sample that each light-emitting device illuminates being known a priori by the imaging device; a single-pixel detector to detect the light generated by the array of light-emitting devices to reconstruct an image of said sample; and an electronic controller to control the array of light-emitting devices. In the proposed imaging device, the light-emitting devices are designed to perform spatiotemporal encoding, with each device illuminating its respective distinct region of the sample, and with all devices generating the light at the same time, but with each being temporarily modulated at a different and unique frequency, based on encoding signals generated by the electronic controller. Additionally, the number of encoding signals corresponds to both the number of light-emitting devices in the array and the number of pixels in the final reconstructed image. The intensity of the light generated by each light-emitting device depends on the amplitude of the encoding signals.

[0023] In some embodiments, the optical conditioning element comprises a lens, e.g. a focal lens.

[0024] In some embodiments, the device additionally comprises at least one optical collection element (e.g. one or multiple lenses) that is / are arranged between the optical conditioning element and the single-pixel detector to collect the light reflected from, emitted by, or transmitted through the sample and to focus the collected light into the single-pixel detector.

[0025] The one or more wavelengths can be comprised in the infrared, ultraviolet, or in the visible.

[0026] The encoding signals can comprise sine waves, square waves, and / or triangular waves.

[0027] In some embodiments, the electronic controller comprises a field-programmable-gated-array (FPGA) or the like.

[0028] In some embodiments, the electronic controller comprises an array of electronic oscillators, each oscillating at a given and unique frequency.

[0029] In some embodiments, the device further includes a battery-powered controller, wherein the single-pixel detector comprises portable elements.

[0030] In some embodiments, the imaging device is used in conjunction with other imaging devices located nearby to increase the field of view and / or to acquire color or spectral images.

[0031] In a particular embodiment, the proposed imaging device comprises an array of light-emitting- diodes (LED) or lasers that are all ON at the same time, but each temporally modulated at a different and unique frequency. A field-programmable-gated-array (FPGA) or similar electronic devices suffice to drive the array. All the light from the array illuminates the sample of interest, with each light-emitting diode illuminating one specific and different region of the sample. The light reflected from, transmitted through, or emitted by that sample is collected with a lens or more complex optical system that focuses the light into a single-pixel detector. The so-collected signal contains the merged information from all the light-emitting devices, and their corresponding spatial positions on the sample. However, after processing the signal - for instance, by calculating the power spectrum of the signal or using a virtual lock-in algorithm - the intensity of the light of each light-emitting device can be retrieved. By knowing a priori the sample’s region that each light-emitting device illuminates, an image can be generated with a number of pixels depending on the number of illuminating light-emitting devices. The maximum speed at which the proposed imaging device can operate depends on the bandwidth of the light-emitting devices used and their response time. Given that current LEDs can operate up to frequencies of 50 MHz or above, imaging at millions of frames per second is possible. Importantly, at the cost of loss of light intensity, it is possible to operate the light-emitting devices faster than their switching on / off time - they only need to be modulated.

[0032] In some embodiments, the imaging device also includes an element configured / designed to recognize the different regions of the sample illuminated by each light-emitting device. In some embodiments, this element comprises a decoding module. In other embodiments, it is implemented as a field-programmable gate array (FPGA) and integrated into the single-pixel detector.

[0033] Present invention also proposes, according to another aspect, a method to reconstruct an image. The method comprises receiving, by each light-emitting device of an array of individually addressable light-emitting devices of an imaging device, an encoding signal from an electronic controller of the imaging device; in response to the received encoding signals, the light-emitting devices performing spatio-temporal encoding, where each light-emitting device illuminates a specific and distinct region of a sample to be imaged using at least one optical conditioning element, the latter being arranged at a given distance from the array of light-emitting devices, all the light-emitting devices generate, at the same time, light at one or more wavelengths that is projected onto the sample to be imaged, with each device being temporarily modulated at a different and unique frequency, and an intensity of the light generated by each light-emitting device depending on an amplitude of the received encoding signal; and reconstructing a generated image of said sample by detecting the generated light using a single-pixel detector of the imaging device.

[0034] In some embodiments, the image is generating by retrieving the intensity of the light reflected from, emitted by, or transmitted through the sample upon illumination with each light-emitting device, and by processing the light detected by the single-pixel detector. In some embodiments, at least one optical collection element, which is arranged between the optical conditioning element and the single-pixel detector, is used to collect the light reflected from, emitted by, or transmitted through the sample and to focus the collected light into the single-pixel detector.

[0035] Present invention offers the following advantages:

[0036] - Speed: millions of frames per second with moderate or even high spatial resolution.

[0037] - Sensitivity: by using LEDs that emit light in the IR, ultraviolet or even in the visible and a single photodetector at such wavelengths, it is possible to reconstruct an image.

[0038] - High signal-to-noise ratio: the reconstruction algorithm allows removing all possible noise from light sources other than the LED array. Because all LEDs illuminate the sample of interest simultaneously, the retrieved signal is N times higher than using a single LED at a time, where N is the total number of LEDs of the array.

[0039] - Cost: the array is cost-effective, even in the IR wavelengths. Single-pixel cameras are also cheaper than traditional cameras - orders of magnitude cheaper in the case of IR cameras.

[0040] - Portability: the array can be as small as 1 mm (especially when considering microLEDs). A single-pixel camera can also have a smaller footprint than a traditional camera - the electronics associated are much simpler.

[0041] - Ease of integration: since the device can be portable, it will be easy to integrate into any optical architecture, either in transmission or reflection illumination.

[0042] Therefore, present invention provides a simple-to-use low-cost single-pixel imaging device that surpasses the speed of existing architectures by generating an image of a sample / object from a single acquisition at millions of fps. The key component of the proposed technology is an array of individually addressable light-emitting devices where each one is driven with periodic electronic signals, such as sinusoidal signals, each with a specific and unique frequency that illuminates a specific and distinct region of a sample of interest. As such, the light-emitting devices are all emitting light, but each is temporally modulated at that specific frequency that illuminates that specific region of the sample, performing the required spatial- temporal encoding needed to reconstruct an image with a single-pixel detector.

[0043] Brief Description of the Drawings The previous and other advantages and features will be more fully understood from the following detailed description of embodiments, with reference to the attached figures, which must be considered in an illustrative and non-limiting manner, in which:

[0044] Figs. 1A and 1 B schematically show the transmission and reflection configuration, respectively, of the proposed imaging device, according to an embodiment of the present invention.

[0045] Fig. 2 shows another embodiment of the proposed imaging device.

[0046] Fig. 3 shows another embodiment of the proposed imaging device, particularly, for imaging of fast dynamic events, a) Cartoon depicting a bullet fired by a soft-air gun and transiting through the image plane where it is illuminated by a line of individually addressable LEDs. Each LED emits n-IR light flickering at a unique frequency. The magnification of the 4-f system was reduced to 1 / 5 to adapt the field of view of the camera to the size of the bullets (6 mm in diameter), b) Space-time (x-t) images of two bullets transiting through the image plane (line) and reconstructed at the frame-rate characteristic of (from left to right) a human eye (50 FPS), fast single-pixel camera (1 kFPS) and pixelated CMOS camera (10 kFPS), and the proposed single-pixel technology (1 MFPS).

[0047] Fig. 4 is a flow chart illustrating a method to reconstruct an image, according to an embodiment of the present invention.

[0048] Detailed Description of the Invention and of Preferred Embodiments

[0049] The present invention discloses an imaging device, e.g. an ultrafast and single-pixel camera, and a method to recover an image using such device.

[0050] Figs. 1A and 1 B illustrate an embodiment of an imaging device 1 for reconstructing an image of a sample 10 of interest. Fig. 1A illustrates the transmission path, and Fig. 1 B the reflection path of the device.

[0051] Particularly, the imaging device 1 of Fig. 1A comprises an array of (individually addressable) light-emitting devices 100; an optical conditioning element 110; the sample 10 of interest to be imaged; an optical collection element 120; and a single-pixel detector 200 that can comprise portable elements. The reflection path comprises the same elements, and additionally a beam splitter or a dichroic mirror 130. The optical conditioning element 110 enables that each element of the array of light-emitting devices 100 illuminates one specific and distinct region of the sample 10.

[0052] The light-emitting devices may comprise infrared, ultraviolet or visible light-emitting-diodes (LED) or lasers.

[0053] The array of light-emitting devices 100 may be formed by a single line of LEDs or lasers.

[0054] The imaging device 1 also includes an electronic controller (not shown), for example a fieldprogram mable-gated-array (FPGA) or the like, to drive the light-emitting devices of the array 100. In some embodiments, the electronic controller comprises an array of electronic oscillators, each oscillating at a given and unique frequency. Eventually, electronic oscillators can be programmable.

[0055] Particularly, the electronic controller produces a (user-programmable) encoding signal for each light-emitting device, such that all the light-emitting devices of the array can be set, simultaneously, to an ON state in order to generate light, though, each light-emitting device is temporally modulated at a different and unique frequency. It should be understood that the intensity of the light generated by each light-emitting device depends on the amplitude of the encoding signals.

[0056] The modulating encoding signals for driving the light-emitting devices of the array 100 can be analogic or digital and can comprise sine, square or triangular waves.

[0057] All the light generated by the array 100 is then projected or collimated onto the sample 1 using the optical conditioning element 110, for example, a lens or a more complex optical system, with each element of the array illuminating one specific and distinct region of the sample 10. Likewise, all the light reflected from or transmitted through the sample 10 is collected with the optical collection element 120, for example, another lens or a more complex optical system, that focuses the light into the single-pixel detector 200.

[0058] In the reflection path, as said before, a beam splitter or dichroic mirror 130 is used to allow part of the light to reach the sample 1 and be reflected towards the single-pixel detector 200.

[0059] The signal containing all the merged information from all the light-emitting devices can be further processed, for instance, by calculating, either in real-time or offline, the power spectrum of the signal using a virtual lock-in algorithm or a lock-in amplifier / spectrum analyzer. Thus, the intensity of the light of each light-emitting device after passing through or being reflected from the sample can be retrieved. Because the sample region that each lightemitting device illuminated, and its corresponding frequency modulation, are known parameters, an image can be generated.

[0060] Various techniques can be employed to identify or know the region of the sample illuminated by each light-emitting device. For instance, in some embodiments, the imaging device, whether integrated or external to the single-pixel detector 200- can include a decodification module or similar component designed to recognize the specific and distinct region of the sample 10 illuminated by each light-emitting device. If integrated in the single-pixel detector, such module / component can be implemented as a chip field-programmable gate array (FPGA) or the like.

[0061] With reference now to Fig. 2, therein another embodiment of the proposed imaging device 1 is shown, where a couple of lenses forming a beam expander are used as the optical conditioning element (110 in Fig. 1A). An image of the 10 by 4 IR-LED array 100 used in this particular implementation is also shown, as well as the plot corresponding to the Fourier Transform of the signal collected by the detector 200 when simultaneously driving the 40 individual LEDs with a unique temporal frequency, ranging form 0.7 to 32 MHz. Importantly, the frequency and amplitude values of the plot peaks are used for image reconstruction: the frequency relates to the spatial position of the LEDs - which is known a priori - while the amplitude to the amount of light absorbed by the sample. All the transmitted light is recorded with a single-pixel detector, namely, an avalanche photodiode (not shown) and digitalized with a fast (sampling frequency >39 MHz) acquisition card (not shown). In this embodiment, given that 40 LEDs were used, the final image features 40 pixels. With such a system, rapidly evolving systems, such as the spinning wheel shown, can be characterized at speeds as high as 600,000 fps.

[0062] Note that in other simpler embodiments, not illustrated, the imaging device 1 is only formed by the cited array of light-emitting devices 100, an optical element, e.g. a lens or array of lens, the single-pixel detector 200, and the electronic controller.

[0063] Fig. 3 shows another embodiment of the imaging device 1. In this case, an individual line of 10 IR-LEDs 100 is used for illumination. Again, each LED is modulated at a unique temporal frequency, from 1 to 20 MHz. This implementation can be used to characterize moving objects, such as a bullet 51 fired from a compressed air gun 50. When one or two bullets 51 pass through the illuminated line, a change in the transmitted intensity can be captured, and by properly decoding the recorded signal, an image can be reconstructed. In experiments herein, an imaging rate as high as 1 ,000,000 fps was achieved - note how such high speed was necessary to properly detect the bullets 51 .

[0064] Fig. 4 illustrates an embodiment of a method for reconstructing an image. The proposed imaging device 1 is used for executing the method. According to this embodiment, at step 301 , each one of the (individually addressable) light-emitting devices of the array 100 receives an encoding signal that temporally modulates each light-emitting device at a different and unique frequency (step 302), and that causes the generation of light that is projected onto the sample 10, simultaneously by all the light-emitting devices, with each light-emitting device illuminating one specific and distinct region of the sample (step 303). At step 304, the singlepixel detector 200 can detect the generated light to reconstruct the image of the sample 10.

[0065] In some particular embodiments, the light detected by the single-pixel detector 200 is processed, using some algorithm, to retrieve the intensity of the light reflected from, emitted by, or transmitted through the sample 10 upon illumination. Therefore, such retrieved intensity can be used for the image reconstruction of the sample 10.

[0066] The proposed imaging device can be used alongside additional imaging devices positioned in close proximity to expand the field of view or capture color or spectral images.

[0067] The embodiments described above are to be understood as a few illustrative examples of the present invention. It will be understood by those skilled in the art that various modifications, combinations, and changes may be made to the embodiments without departing from the scope of the present invention. In particular, different part solutions in the different embodiments can be combined in other configurations, where technically possible.

Claims

CLAIMS1. An imaging device, comprising: an array of individually addressable light-emitting devices (100) configured to generate light at one or more wavelengths, said light being projected onto a sample (10) to be imaged; a detector (200) configured to detect the light generated by the array of light-emitting devices (100) to reconstruct an image of said sample (10); and an electronic controller configured to control the array of light-emitting devices (100); characterized in that: the imaging device further comprises at least one optical conditioning element (110) arranged at a given distance from the array of light-emitting devices to ensure that light from each light-emitting device illuminates a specific and distinct region of the sample (10), the specific and distinct region of the sample (10) that each light-emitting device illuminates being known a priori by the imaging device; the detector (200) is a single pixel detector,; and the light-emitting devices are configured to perform spatio-temporal encoding, wherein: each light-emitting device illuminates its respective distinct region of the sample (10), all the light-emitting devices are configured to generate the light at the same time, with each one being temporarily modulated at a different and unique frequency, based on encoding signals generated by the electronic controller, the number of encoding signals is equal to the number of light-emitting devices in the array, and an intensity of the light generated by each light-emitting device depends on an amplitude of the encoding signals.

2. The imaging device of claim 1 , wherein the at least one optical conditioning element (110) comprises a lens.

3. The imaging device of claim 1 or 2, further comprising at least one optical collection element (120) that is arranged between the optical conditioning element (110) and the single-pixel detector (200) to collect the light reflected from, emitted by, or transmitted through the sample (10) and to focus the collected light into the single-pixel detector (200).

4. The imaging device of any one of the previous claims, wherein the light-emitting devices comprise light-emitting diodes or lasers.

5. The imaging device of any one of the previous claims, wherein one or more wavelengths are comprised in the infrared, ultraviolet or visible.

6. The imaging device of any one of the previous claims, wherein the electronic controller comprises a field-programmable-gated-array, FPGA.

7. The imaging device of any one of the previous claims, wherein the electronic controller comprises an array of electronic oscillators, each oscillating at a given and unique frequency.

8. The imaging device of any one of the previous claims, wherein the encoding signals comprise sine waves, square waves, and / or triangular waves.

9. The imaging device of any one of the previous claims, further comprising a battery-powered controller, wherein the single-pixel detector (200) comprises portable elements.

10. The imaging device of claim 3, wherein at least one optical collection element (120) comprises a lens.11 . A method to reconstruct an image, comprising: receiving, by each light-emitting device of an array of individually addressable lightemitting devices of an imaging device, an encoding signal from an electronic controller of the imaging device; in response to the received encoding signals, the light-emitting devices performing spatio-temporal encoding, wherein: each light-emitting device illuminates a specific and distinct region of a sample to be imaged using at least one optical conditioning element, the latter being arranged at a given distance from the array of light-emitting devices, and the specific and distinct region of the sample that each light-emitting device illuminates being known a priori by the imaging device; all the light-emitting devices generate light at the same time at one or more wavelengths, the light being projected onto the sample to be imaged, with each light-emitting device being temporarily modulated at a different and unique frequency based on the received encoding signal, and an intensity of the light generated by each light-emitting device depends on an amplitude of the received encoding signal; and reconstructing a generated image of the sample by detecting the generated light using a single-pixel detector.

12. The method of claim 11 , wherein the image being generated by retrieving an intensity of the light reflected from, emitted by, or transmitted through the sample upon illumination with each light-emitting device and by processing the light detected by the single-pixel detector.

13. The method of claim 11 or 12, wherein at least one optical collection element, which is arranged between the optical conditioning element and the single-pixel detector, is used to collect the light reflected from, emitted by, or transmitted through the sample and to focus the collected light into the single-pixel detector.

Images