SPECTROSCOPY APPARATUS

Abstract

There is provided a spectroscopy apparatus for measuring fluorescence signals from a photosynthetic object. The spectroscopy apparatus comprises: one or more light excitation sources (26,28) operable to carry out time-varying excitation of the fluorescence from the photosynthetic object; and one or more fluorescence-sensitive detection channels (36,38,44,46) configured to simultaneously record the fluorescence as a function of time with a microsecond to millisecond time resolution and as a function of wavelength with a wavelength resolution of 10 nm or better, responsive to the excitation of the fluorescence from the photosynthetic object by the or each light excitation source (26,28).

Claims

1. A spectroscopy apparatus for measuring fluorescence signals from a photosynthetic object, the spectroscopy apparatus comprising: one or more light excitation sources operable to carry out time-varying excitation of the fluorescence from the photosynthetic object; and one or more fluorescence-sensitive detection channels configured to simultaneously record the fluorescence as a function of time with a microsecond to millisecond time resolution and as a function of wavelength with a wavelength resolution of 10 nm or better, responsive to the excitation of the fluorescence from the photosynthetic object by the or each light excitation source.

2. A spectroscopy apparatus according to claim 1 wherein the one or more fluorescence-sensitive detection channels includes one or more fluorescence-sensitive detection units or devices.

3. A spectroscopy apparatus according to claim 1 wherein the wavelength resolution of the recorded fluorescence information is achieved continuously across the entire recorded fluorescence spectrum.

4. A spectroscopy apparatus according to claim 1 wherein the wavelength resolution of the recorded fluorescence information is achieved using three or more distinct narrow wavelength bands.

5. A spectroscopy apparatus according to claim 1 wherein the recorded fluorescence information includes fluorescence induction information.

6. A spectroscopy apparatus according to claim 1 wherein the recorded fluorescence information includes non-photochemical quenching information.

7. A spectroscopy apparatus according to claim 1 wherein the time-varying excitation is in the form of a repeating pulsed excitation that has a microsecond to millisecond pulse duration.

8-10. (canceled)

11. A spectroscopy apparatus according to claim 1 wherein the time-varying excitation is in the form of a periodically modulated excitation.

12. (canceled)

13. A spectroscopy apparatus according to claim 1 wherein the time resolution is in the range of 0.5 microseconds to 10 milliseconds.

14. A spectroscopy apparatus according to claim 1 wherein the wavelength resolution is in the range of 1 nm to 10 nm.

15. A spectroscopy apparatus according to claim 1 wherein the electronic circuit includes a processor and memory including computer program code, the memory and computer program code configured to, with the processor, enable the electronic circuit at least to analyse the recorded fluorescence information from the photosynthetic object so as to identify or characterise a condition of the photosynthetic object.

16. A spectroscopy apparatus according to claim 15 wherein the memory and computer program code are configured to, with the processor, enable the electronic circuit at least to analyse modified derivative functions of the recorded fluorescence information from the photosynthetic object so as to identify or characterise a condition of the photosynthetic object.

17. (canceled)

18. A spectroscopy apparatus according to claim 15 wherein the memory and computer program code are configured to, with the processor, enable the electronic circuit at least to analyse the recorded fluorescence information from the photosynthetic object to identify or characterise the condition of the photosynthetic object by providing the recorded fluorescence information as input to a machine learning algorithm or model and identify or characterise the condition of the photosynthetic object based on an output of the machine learning algorithm or model.

19. A spectroscopy apparatus according to claim 18 wherein the machine learning algorithm or model includes a long short-term memory algorithm or a neural network.

20. A spectroscopy apparatus according to claim 15 wherein the condition of the photosynthetic object includes at least one of: a physiological condition of the photosynthetic object; a health condition of the photosynthetic object; and a stress condition of the photosynthetic object.

21. (canceled)

22. A method of measuring fluorescence signals from a photosynthetic object using a spectroscopy apparatus according to claim 1, the method comprising the steps of: by the or each light excitation source, carrying out time-varying excitation of the fluorescence from the photosynthetic object; and by the or each fluorescence-sensitive detection channel, simultaneously recording the fluorescence as a function of time with a microsecond to millisecond time resolution and as a function of wavelength with a wavelength resolution of 10 nm or better, responsive to the excitation of the fluorescence from the photosynthetic object by the or each light excitation source.

23. (canceled)

24. (canceled)

25. A computer-implemented method of identifying or characterising a condition of a photosynthetic object, the method comprising the steps of: recording fluorescence information from the photosynthetic object by carrying out the method according to claim 22; and analysing the recorded fluorescence information from the photosynthetic object so as to identify or characterise a condition of the photosynthetic object.

26. A computer-implemented method of identifying or characterising a condition of a photosynthetic object, the method comprising the steps of: collecting a set of data by carrying out the method according to claim 22, wherein the collected set of data includes the recorded fluorescence information from the photosynthetic object; creating a training set including the collected set of data; training a machine learning algorithm or model using the training set; and identifying or characterising the condition of the photosynthetic object based on an output of the machine learning algorithm or model.

27. (canceled)

28. A computer-implemented method according to claim 26 wherein the step of identifying or characterising the condition of the photosynthetic object based on an output of the machine learning algorithm or model includes analysis of stress phenomena associated with a stress condition of the photosynthetic object.

79. A computer-implemented method according to claim 26 wherein the step of identifying or characterising the condition of the photosynthetic object based on an output of the machine learning algorithm or model includes plant, phenotyping or genotyping.

30-34. (canceled)

Description

[0090] Preferred embodiments of the invention will now be described, by way of non-limiting examples, with reference to the accompanying drawings in which:

[0091] FIG. 1 shows a spectroscopy apparatus according to an embodiment of the invention;

[0092] FIG. 2 illustrates a repeating burst pulse excitation mode employed for simultaneous time-resolved and wavelength-resolved measurement of fluorescence signals from photosynthetic objects; and

[0093] FIGS. 3 to 10 show data obtained from simultaneous time-resolved and wavelength-resolved measurements of fluorescence from plants.

[0094] The figures are not necessarily to scale, and certain features and certain views of the figures may be shown exaggerated in scale or in schematic form in the interests of clarity and conciseness.

[0095] The following embodiments of the invention are described with reference to the configuration of the invention as a plant health and stress monitoring system, particularly for early biotic and abiotic stress detection and analysis and for optimising plant growth conditions. It will be appreciated that the following embodiments of the invention apply mutatis mutandis to other applications relating to photosynthetic objects and other types of fluorescence emitted by photosynthetic objects, non-limiting examples of which are described throughout the specification.

[0096] Chlorophyll fluorescence is a sensitive indicator of many different kinds of plant stresses, since the photosynthetic machinery responds to all kinds of environmental parameters and their changes (such as light, water, nutrition, temperature and others) in all green plants in a sensitive manner. Chlorophyll fluorescence measurements permit extraction of information about the status of the plant, and are affected by water stress, nitrogen, temperature, chlorophyll and flavonoid contents, pH level, photosynthetic status and other physiological parameters, independently from soil interference, leaf area or biomass status. Chlorophyll fluorescence tends to change long before damage resulting from stress becomes visible or detectable. Hence, chlorophyll fluorescence can be used as early warnings of stress to detect adverse conditions of plants.

[0097] The inventors have found that chlorophyll fluorescence, when measured as a combination of simultaneously recorded time-resolved and wavelength-resolved information, and measured in the modes described above and below according to the invention, is particularly suitable for providing early warnings of stress and/or to detect any other adverse conditions of plants and other photosynthetic tissue.

[0098] A spectroscopy apparatus according to an embodiment of the invention is shown in FIG. 1. The spectroscopy apparatus comprises an enclosure 20, an electronic circuit 22, a sample holder 24, a light excitation source and a fluorescence-sensitive detection channel. The enclosure 20 is configured to block external light from entering its interior in order to create a dark enclosure 20, preferably a light-tight enclosure.

[0099] In use, the sample holder 24 is housed inside the dark enclosure 20 (preferably a light-tight enclosure) and holds a photosynthetic object. In the embodiment shown, the photosynthetic object is a plant leaf. Other photosynthetic objects may be used in other embodiments.

[0100] The light excitation source includes at least a pair of light emitting diodes (LED) 26,28 that are powered by respective power supplies 30. More than two LEDs may be used. A first of the LEDs is a red LED 26 with a centre wavelength of 620 or 650 nm, while a second of the LEDs is a blue LED 28 with a centre wavelength of 420 or 475 nm. Each of the LEDs 26,28 can provide any one of: a low intensity actinic background illumination or a high intensity (≥50,000 microEinsteins per second per square metre) single pulse or burst pulse, or a repetitively modulated excitation, or a combination thereof. Thus each of the LEDs 26,28 be used as an actinic source, a high power excitation source, or a combination thereof. Single colour LEDs with other centre wavelengths are possible. Both LEDs 26,28 are fully programmable to provide a continuous background light excitation or a repeating pulsed light excitation, which may be in the form of a single pulse excitation or a burst pulse excitation. This is achieved by controlling each power supply 30 to supply a continuous power (for actinic background illumination) or pulsed power to the corresponding LED 26,28.

[0101] The repeating pulsed excitation preferably has a microsecond pulse duration (preferably 10 microsecond to a millisecond) and an excitation frequency of 0.1 kHz to 100 kHz (depending what is achievable for the particular pulse duration, allowing for suitable intervals of 0.1 to 10 milliseconds lengths). It will be appreciated that the choice of wavelength or colour of each of the LEDs 26,28 may vary depending on the required properties of the light excitation for a particular photosynthetic object.

[0102] It is envisaged that, in other embodiments of the invention, a periodically modulated light excitation may be used in place of the repeating pulsed light excitation (single pulse or burst pulse). The periodically modulated light excitation is preferably based on a sinusoidal or any non-sinusoidal irregular but periodically repeating waveform—with or without background bias level—with an excitation frequency in the range of 1 mHz to 100 kHz.

[0103] It is also envisaged that, in still other embodiments of the invention, a different type of light excitation source may be used and/or a different number of light emitting diodes (e.g. one, three or more) may be used.

[0104] In use, the LEDs 26,28 are positioned inside the dark enclosure 20 (preferably a light-tight enclosure) to direct their emitted light towards the photosynthetic sample or object (e.g. a leaf) on the sample holder 24. Optical lenses 32 are positioned in front of the LEDs 26,28 to focus the light emitted by the LEDs 26,28 onto the photosynthetic sample (e.g. a leaf). Short-pass filters 34 are positioned between the optical lenses 32 and the sample holder 24 to remove long wavelength emission.

[0105] The fluorescence-sensitive detection channel includes an optical fibre 36 and a spectrograph 38. The spectrograph 38 is configured to receive a signal from the optical fibre 26.

[0106] In use, the optical fibre 36 and associated input optics are positioned inside the dark enclosure 20 (preferably a light-tight enclosure) to receive the fluorescence emission from the photosynthetic sample or object (e.g. a leaf). An optical lens 40 is positioned between the optical fibre 36 and the sample holder 24 to focus the fluorescence emission onto the optical fibre 36. A long-pass filter 42 is positioned between the optical fibre 36 and the optical lens 40 to filter scattered excitation light.

[0107] The electronic circuit 22 includes a processor and memory including computer program code. The memory and computer program code are configured to, with the processor, enable the electronic circuit 22 to carry out various processing functions. In the embodiment shown, the electronic circuit 22 forms part of a dedicated microprocessor that is connected to the spectrograph 38, so that the electronic circuit 22 may be used to record the received fluorescence spectrum as a function of time, and analyse and display the received fluorescence spectrum. The microprocessor is controlled in turn by a desktop or notebook computer used for input/output operations and display. In other embodiments, the electronic circuit may be, may include or may form part of one or more of an electronic device, a portable electronic device, a portable telecommunications device, a mobile phone, a personal digital assistant, a tablet, a phablet, a laptop computer, a server, a cloud computing network, a smartphone, a smartwatch, smart eyewear, and a module for one or more of the same. It will be appreciated that references to a memory or a processor may encompass a plurality of memories or processors.

[0108] Use of the spectroscopy apparatus of the invention to measure various fluorescence signals from the leaf is described as follows, with reference to FIGS. 2 to 10.

[0109] The below simultaneous time-resolved and wavelength-resolved measurements are described with reference to a dark-adapted sample. In other embodiments of the invention, the below simultaneous time-resolved and wavelength-resolved measurements may apply mutatis mutandis to a light-adapted sample. Measurements on light-adapted samples would be particularly relevant under conditions where dark adaptation would be technically difficult or impossible, e.g. measurements from a distance.

[0110] Initially the leaf is left in the dark enclosure 20 for a period of time sufficient to enable the leaf to enter a desired dark-adapted or relaxed state. At the same time zero background signals are recorded by the spectrograph 38 for referencing all signals to zero.

[0111] Then, one of the LEDs 26,28 may be controlled to provide a continuous weak actinic light of desirable intensity (which may alternatively be of zero intensity) to initiate the photosynthetic process(es) that may be, but not limited to, FI and NPQ. Either the same or the other LED 26,28 is controlled to provide a repeating pulsed excitation (single pulse or burst pulse) at an excitation frequency of 0.1 kHz to 100 kHz to cause emission of fluorescence from the leaf. Each pulse of the repeating pulsed excitation has a pulse duration in the order of 10 microseconds to several hundreds of microseconds. The optical fibre 36 captures the fluorescence emission from the leaf. In this way, a full fluorescence spectrum is recorded by the spectrograph 38 in very short times (typically for 10 microseconds to several tens of microseconds intervals) upon each pulse of the high intensity LED 26,28. The trigger signal for the pulses of the repeating pulsed excitation may be supplied by the spectrograph 38 so that the pulse generation can be synchronised with the fluorescence detection, or may be supplied by another controller. The measuring time window of the spectrograph 38 is positioned either at the beginning (for Fo′ measurements) or at the end (for Fm′) of an individual excitation pulse, or at any other time delay after the excitation pulse.

[0112] In this way the spectroscopy apparatus of the invention is able to obtain simultaneous time-resolved and wavelength-resolved measurements of the fluorescence of the leaf at high resolution (typically at 10 microsecond intervals).

[0113] A reference photodiode 45 may be used to monitor scattered high excitation light from the LEDs 26,28 with high time resolution (typically 0.3 to 0.5 microseconds resolution) in order to allow correction of the measured fluorescence signals for small intensity variations of the measuring light pulses across the pulse length and between consecutive pulses.

[0114] Optionally, or in parallel in the same apparatus, depending on the desired application, the fluorescence-sensitive detection channel may include a plurality of photodiodes 44 (preferably three or more) equipped with different narrow-band optical filters. Each photodiode 44 may be connected to several respective measurement channels 46 providing sub-microsecond resolution (typically 0.3 to 0.5 microseconds) so that each photodiode 44 is configured to produce a time-signal upon pulse excitation of the fluorescence from a narrow wavelength window of the total fluorescence emission from the leaf.

[0115] FIG. 2 shows a schematic graph illustrating the repeating burst pulse excitation mode. An N number of pulses (typically up to several hundred pulses per burst pulse) at high intensity (≥50,000 microEinsteins per second per square metre) are applied by one of the LEDs 26,28 to the photosynthetic object in the repeating burst pulse excitation mode. Each individual pulse has a duration of 10 to 100 microseconds, with the burst pulses being separated by 0.1-10 millisecond intervals. In the particular case where the number of high intensity excitation pulses N per burst equals to 1, the repeating pulse excitation defaults to the single pulse excitation mode. During each individual pulse of the burst pulse, the fluorescence signals recorded by the spectrograph 38 and/or the wavelength-specific photodiodes 44 are measured in a time-resolved and wavelength-resolved manner simultaneously with a time resolution in the range of 0.5 to 10 microseconds. Such burst pulses may be repeated automatically using any interval and any total number of bursts, depending on the purpose of the measurement (including but not limited to FI, NPQ, or other measuring modes). The entire pulse sequence, measuring procedure and data recording may be under the control of a programmable dedicated microprocessor as described with reference to FIG. 1. Weak actinic background irradiation, if desired, can be applied by the LEDs 26,28 or by a separate LED as actinic source under dedicated microprocessor control.

[0116] FIG. 3A shows typical 3D time-resolved and wavelength-resolved chlorophyll fluorescence emission spectra from a plant leaf using low intensity (e.g. 600 microEinsteins per second per square metre) actinic background illumination. FIG. 3B shows the normalised time-resolved decay-associated chlorophyll fluorescence emission spectra resulting from a global kinetic analysis. FIG. 3C shows normalised decay-associated spectra of various intermediate fluorescing species induced by the actinic and LED pulsed excitation as resulting from a global kinetic/spectral target model analysis. FIG. 3D shows the corresponding time-dependent concentrations of these intermediately developing and decaying fluorescing species resulting from a global kinetic/spectral target model analysis for the case of a plant that has undergone mild heat treatment (mild heat stress). FIG. 3E shows the same data as FIG. 3D except that the data in FIG. 3E corresponds to a normally grown control plant without any applied stress. Both FI and NPQ information can be derived from these data. It can be seen from FIGS. 3A to 3C that the high resolution of the simultaneous time-resolved and wavelength-resolved measurements of the chlorophyll fluorescence of the leaf by the spectroscopy apparatus of the invention enables spectrally and kinetically resolved separation and identification of components that contribute to the FI, NPQ and other processes initiated by the actinic light in the photosynthetic object. It can be seen from FIGS. 3D and 3E that there are substantial differences in the kinetics of the intermediate fluorescing species from the heat-stressed leaf and control plant and from FIG. 3C that there are significant spectral differences between these species. These differences can be further analysed qualitatively and quantitatively to identify inter alia, but not limited to, the type and level of stress in order to determine the stress condition of the photosynthetic object, the general physiological status, its general health status, or to perform characterisation of the photosynthetic object according to phenotyping and/or genotyping criteria.

[0117] FIG. 4 shows exemplary time-resolved measured fluorescence signals from a plant leaf for three selected wavelength ranges (685 nm, 700 nm and 730 nm, each having a 5 to 10 nm spectral width) for a single high intensity pulse (pulse duration of 100 microseconds) taken from a burst pulse.

[0118] FIG. 5 shows exemplary time-resolved measured fluorescence signals from a plant leaf for one selected wavelength range (685 nm, with a 10 nm spectral width) for a selection of high intensity pulses (pulse duration of 130 microseconds) taken from the same burst pulse. During each pulse, time-resolved and wavelength-resolved fluorescence signals are measured with a 0.5 to 10 microseconds time resolution.

[0119] FIG. 6 shows measured FI signals on a semilogarithmic time axis for Fo′ functions (open symbols, representing a fluorescence signal at the beginning of each individual pulse) and Fm′ functions (closed symbols, representing a fluorescence signal at the end of each individual pulse) resulting from a burst pulse. The sample was a mildly dry-stressed tobacco leaf. The signal is double-normalised to the Fm′ signal and is shown for clarity for one particular wavelength range (685 nm, with a 10 nm spectral width) out of the entire recorded spectrum. The same symbol type is used to display the corresponding derivative functions of the FI signals in logarithmic time space. These signals are then classified in a global fashion (three or more/all wavelength ranges combined) by machine-learning algorithms.

[0120] FIG. 7 shows FI signals from a mildly nitrogen-depleted (nutrient-stressed) maize plant. The wavelength range of 685 nm with a 10 nm spectral width is shown on a logarithmic time axis.

[0121] FIG. 8 compares typical effects of different stresses (“nutrient-stressed” meaning mild nitrogen depletion; “drought-stressed” meaning early mild dry-stress on the plant; “normally grown” meaning no stress on the plant) on the FI functions of maize leaves (top row) and their logarithmic derivatives (bottom row). The signals are subsequently subjected to classification by machine-learning (AI) algorithms. The shown fluorescence signals are for the wavelength range of 700 nm with a 10 nm spectral width.

[0122] FIG. 9 shows a spectrally and temporally resolved NPQ experiment on a leaf of a normally grown barley plant for four different wavelength ranges (top to bottom: 686 nm, 700 nm, 720 nm and 738 nm, each having a 5 to 10 nm spectral width, selected from the entire spectral range extending from 650 nm to 850 nm). The spectrally dependent effects of an induction phase (600 microEinsteins per second per square metre), a relaxation phase and a second induction phase are shown in FIG. 9.

[0123] FIG. 10 shows normalised FI measurements on a barley leaf for three selected wavelength ranges (685 nm, 700 nm and 730 nm) for the Fm′ functions (closed symbols) and the Fo′ functions (open symbols) displayed on a semilogarithmic time axis. The same symbol type is used to display the corresponding derivative functions of the FI signals in logarithmic time space.

[0124] The inventors have also found that analysing derivatives of the FI (FIGS. 6, 7, 8, and 10) and the NPQ (FIG. 9) on a logarithmic time scale results in significant improvement in providing additional details that enables accurate discrimination of the condition of the leaf. Due to the derivative of a function measuring the sensitivity to a change of the function's output with respect to a change in the function's input, the analysis technique permits accurate tracking of changes in the FI and NPQ that describe the photosynthetic status of the leaf that responds differently to different stresses. Applying such kinds of mathematically derived functions (e.g. derivative functions, logarithmic derivative functions or other mathematical functions based on the originally measured signals) has been found to be particularly useful for the application of machine-learning algorithms (AI) in the classification and differentiation of stress phenomena and their application to phenotyping and genotyping applications.

[0125] The recorded FI and NPQ information and their logarithmic time scaled derivatives or other derived mathematical functions of the underlying original functions may be used to create a training set. The training set is then used to train a machine learning algorithm or model over time. A condition of the leaf may be identified and classified based on the output of the machine learning algorithm or model. The inventors have identified a long short-term memory (LSTM) algorithm as being particularly reliable in identifying different types and levels of stress based on chlorophyll fluorescence measurements. Specifically, using the recorded FI and NPQ information and their logarithmic time scaled derivatives in combination with the LSTM machine learning algorithm results in a classification accuracy of around 70% with small amounts of data. Further improvements in classification accuracy (90% and above accuracy) have been found possible using larger training sets. This is in part due to the robustness of the LSTM machine learning algorithm in recognizing patterns in time-series data, such as FI and NPQ data. Such an approach is also useful to characterise photosynthetic objects according to phenotyping and genotyping criteria.

[0126] The spectroscopy apparatus of the invention provides at least the following benefits: [0127] i) The spectrally resolved measurement by the invention provides a much broader database for analysis and characterisation of stress phenomena that allows for extraction of more reliable and precise information regarding the health and stress conditions of the plant or photosynthetic object. This is relevant inter alia, but not limited to, the efficient automatic monitoring of ecosystems, such as forests, coral reefs etc., and their responses to climate change and other environmental changes and factors. [0128] ii) The spectrally resolved measurement by the invention provides a much broader database for analysis and characterisation of internal processes triggered by all external factors within the photosynthetic object, not limited to, but in particular light-induced photosynthetic processes, thus enabling a more detailed and more accurate characterisation and separation of photosynthetic reactions, photosynthetic regulation mechanism and responses of photosynthetic objects. This is of high relevance both for fundamental photosynthesis and plant physiology research, as well as for plant breeding aiming at higher crop yields, higher resistance to climate changes, selection of more efficient photosynthetically driven energy producing plants or energy-producing microorganisms for solar conversion to biofuels. [0129] iii) For field applications focused on plant monitoring, growth optimisation and early stress detection, the invention provides a field-ready and user-friendly instrument that is capable of carrying out the requisite fluorescence measurement and analysis. [0130] iv) The incorporation of a machine learning model in the invention to perform automatic analysis and provide expert recommendations makes it easier for non-specialist users to obtain reliable information and take appropriate countermeasures against stress conditions. [0131] v) The machine learning model can be trained by end users to improve and optimise the information about the stress conditions of the plant, thus tailoring the spectroscopy apparatus specifically for certain plants and applications.

[0132] It will be appreciated that the above numerical values are merely intended to help illustrate the working of the invention and may vary depending on the requirements of the spectroscopy apparatus and the photosynthetic object.

[0133] The listing or discussion of an apparently prior-published document or apparently prior-published information in this specification should not necessarily be taken as an acknowledgement that the document or information is part of the state of the art or is common general knowledge.

[0134] Preferences and options for a given aspect, feature or parameter of the invention should, unless the context indicates otherwise, be regarded as having been disclosed in combination with any and all preferences and options for all other aspects, features and parameters of the invention.