Non-destructive quantitative determination of at least one performance indicator in the rearing of live insect populations in complex media

Abstract

The present invention relates to a non-destructive method for quantitatively determining at least one performance index in rearing a population of live insects in a complex medium, the method comprising the steps of: (i) irradiating a complex medium containing a population of live insects with light comprising one or more wavelengths in the range of 420 nm to 2500 nm, (ii) collecting light reflected by the complex medium to obtain a reflectance spectrum of the medium, and (iii) correlating said spectrum to a quantitative value of at least one performance index. The present invention also relates to the use of this method and the use of a spectrophotometer to quantitatively determine at least one performance index of a population of live insects in a complex medium.

Description

[Technical field]

[0001] The present invention belongs to the field of analytics and more specifically relates to a non-destructive method that allows the determination of a quantitative value of at least one performance indicator.

[0002] Over the last few years, insect farming has increased significantly. It is interesting for many reasons, for the agricultural industry, since insects are a source of desired proteins, and for other industrial sectors, since insects are also a source of chitin, which can be converted by deacetylation into chitosan. Its applications are numerous, for example in cosmetics, the medical and pharmaceutical sector, the nutritional and food industry, water treatment, etc.

[0003] It is therefore necessary to introduce industrial models adapted to insects that allow the automatic monitoring of insect rearing in order to maintain optimal growth conditions.

[0004] This monitoring is possible through the quantification of performance indicators (Key Performance Indicators or KPIs).

[0005] "Performance indicator" (Key Performance Indicator or KPI) means information, preferably intended to be expressed numerically, that makes it possible to measure and evaluate the result of one or more actions, to monitor performance developments and / or to analyze a situation.

[0006] The methods currently used are manual and cumbersome (selection and counting of insect populations, visual inspection of rearing tanks) and are not suitable for industrial-scale rearing. Moreover, these methods lack precision and can damage the insects during development, especially during the latter operations.

[0007] Therefore, there is a need to automate the monitoring of live insect population growth, as well as the monitoring of equipment for sorting and feeding live insect populations.

[0008] The present invention overcomes these technical problems and relates to a non-destructive method for determining the quantitative value of at least one performance indicator. Moreover, the inventors have shown in a surprising and unexpected manner that the method according to the invention makes it possible to determine the quantitative value of at least one performance indicator with very low uncertainty in complex media. Summary of the Invention

[0009] The present invention relates to a non-destructive method for quantitatively determining at least one performance indicator of the rearing of a population of live insects in a complex medium, said method comprising: (i) irradiating a complex medium containing a population of live insects with light containing one or more wavelengths in the range of 420 nm to 2500 nm; (ii) collecting light reflected by the composite medium to obtain a reflectance spectrum of the medium; (iii) correlating the spectrum with a quantitative value of at least one performance indicator.

[0010] The present invention also relates to the use of a method as defined above for carrying out monitoring of the growth of a population of live insects, for determining the efficiency of a sorting machine and / or for determining the efficiency of a device for feeding a population of live insects.

[0011] The present invention also relates to the use of a spectrophotometer for quantitatively determining at least one performance index of the rearing of a population of live insects in a complex medium. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] In the context of this application, unless otherwise indicated, ranges of values ​​given are understood to include the endpoints.

[0013] The first object relates to a non-destructive method for quantitatively determining at least one performance indicator of the rearing of a population of live insects in a complex medium, said method comprising: (i) irradiating a complex medium containing a population of live insects with light containing one or more wavelengths in the range of 420 nm to 2500 nm; (ii) collecting light reflected by the composite medium to obtain a reflectance spectrum of the medium; (iii) correlating the spectrum with a quantitative value of at least one performance indicator.

[0014] By "non-destructive method" is preferably meant a series of non-invasive steps that make it possible to preserve the integrity of the live insect population.

[0015] By "insect" is preferably meant any stage of development, such as adult, larvae, nymphs and / or eggs, preferably insects in the adult, larvae and / or nymphal stages. Thus, the term encompasses any stage of development (also called growth stage or development stage) from eggs or egg capsules to adults, including nymphs such as larvae and pupae.

[0016] By "population of insects" is preferably meant a set of at least two insects of the same or different species and / or of the same or different developmental stages.

[0017] In particular, the insect population according to the invention relates to an individualized insect population. The insect population contemplated by the invention therefore does not correspond to insects that hatch eggs inside their abdomens (ovoviviparous insects).

[0018] Preferably, the population of insects, more particularly the population of live insects, is a set of at least two insects of the same species at the same developmental stage, examples of populations of insects include, but are not limited to, a set of larvae, a set of nymphs or a set of adults.

[0019] "Lymph" preferably means the larval stage of an insect, and thus the immature stage of an insect, including maggots of the Diptera order and caterpillars of the Lepidoptera order, as well as the apterous stages of the Orthoptera order. The term encompasses any developmental stage of the larva.

[0020] By "nymph" is preferably meant a stage of insect metamorphosis, in particular an intermediate stage between larva and adult, including in Diptera the pupa, in Coleoptera the nymph, in Lepidoptera the hard-shelled pupa, and, if necessary, for Diptera, an intermediate stage at which specific physiological (pre-pupal) or behavioral changes in the individual appear, such as a significant sclerotinization of the cuticle.

[0021] By "adult insect" (or adult) is preferably meant the mature stage of an insect, in particular the final stage of insect development that ensures the reproduction of the species.

[0022] Advantageously, the insect is selected from the group of Coleoptera, Diptera, Lepidoptera, Orthoptera, Neuroptera, Hymenoptera, especially the order Blattopteran including Isoptera and Mantopteran, Phasmatopteran, Hemiptera, Heteroptera, Ephemeroptera, Mecoptera and mixtures thereof, preferably from the group of Coleoptera, Diptera, Lepidoptera, Orthoptera and mixtures thereof, more preferably the insect belongs to the group of Coleoptera. Preferably, the Diptera belongs to the suborder Ditrysia. Preferably, the Lepidoptera belongs to the suborder Ditrysia, more preferably to the superfamily Pyralidae. Preferably, the Neuroptera belongs to the suborder Hemiptera. More preferably, the order Neuroptera belongs to the family Mantodea, the family Ephemeridae, the family Chrysopidae, the family Nerulidae, or a mixture thereof. Preferably, the beetle belongs to the suborder Pycnonotida, in particular the family Tenebrionidae, the family Coccinellidae, the family Cerambycidae, the family Curculionidae, or a mixture thereof. More preferably, the beetle is selected from the group Tenebrionidae, the Tenebrionidae, the giant Tenebrionidae, the Tenebrionidae, the Red Flour Beetle, the Palm Weevil, and a mixture thereof, and even more preferably, the group Tenebrionidae, the Tenebrionidae, and a mixture thereof. Thus, the insects targeted by the present invention preferably belong to the group of beetles, and more particularly to the family Tenebrionidae. Preferably, the insects targeted by the present invention belong to the species Tenebrionid molitor and / or Tenebrionid molitor.

[0023] A "complex medium" preferably comprises: - an assemblage comprising at least two, preferably at least three, different components other than a population of live insects, and / or at least one insect that is at at least one different developmental stage from at least one insect of the population of live insects, and / or - means at least one dead insect.

[0024] By "an assembly comprising at least two, preferably at least three, different components" is meant an assembly that preferably comprises at least two, preferably at least three, components that are at least partially composed of identical or similar molecules, but in different concentrations.

[0025] By "identical or similar molecules" is preferably meant chemical molecules that have the same or similar chemical structure.

[0026] For example, these are molecules with a common chemical structure that have the same vibrations of hydrogen atoms as those of the molecular family to be measured (proteins, lipids, sugars, chitin, water) during excitation of the latter by an infrared source. For example, if hydrogen is attached to nitrogen (through the N-H2-COOH group), the molecule belongs to the family of proteins.

[0027] Thus, the proteins have the same or similar chemical structure.

[0028] Similarly, monosaccharides, oligosaccharides and polysaccharides have the same or similar chemical structures, triglycerides and fatty acids have the same or similar chemical structures, and chitin and chitin derivatives have the same or similar chemical structures.

[0029] "Chitin derivative" preferably means a derivative of a polysaccharide containing N-acetylglucosamine units and D-glucosamine units, in particular a chitin-polypeptide copolymer (sometimes also called "chitin-polypeptide complex").

[0030] Indeed, in a complex medium according to the invention, each component may in particular comprise two or more compounds selected from the following: -protein, -mono-, oligo- and polysaccharides, triglycerides and fatty acids, and / or -Chitin and chitin derivatives.

[0031] More specifically, each of the above compounds is present in at least two selected from a component of a complex medium and a live insect population.

[0032] Preferably, the complex medium is an insect rearing medium.

[0033] Examples of components may be, but are not limited to, a substrate, and / or dead insects of the same or different species as the species of the live insect population, and / or live insects of a different species than at least one insect of the live insect population, preferably live insects of a different species than the species of the live insect population, and / or live insects at a different developmental stage than at least one insect of the live insect population, and / or feces (also called frass), and / or contaminants.

[0034] "Substrate" preferably means a nutrient substrate.

[0035] "Nutritional substrate" preferably means a nutritious substrate or a mixture of nutritious substrates intended to be consumed by insects. It may be a solid or liquid nutritious substrate. The solid nutritious substrate may be a mixture of solid products or by-products from the processing of cereals, oleaginous plants, oleaginous proteinaceous plants and / or proteinaceous plants. This substrate may be supplemented with vitamins or minerals. "By-products" preferably means unavoidable substances produced during the process of producing the desired product. The solid substrate may be a product or by-product from the processing of wheat (common wheat, durum wheat), corn, barley, rice, triticale, oats, sorghum, rye, spelt, millet, quinoa, buckwheat, rapeseed, sunflower, flax, soybean and / or pea. Preferably, the solid substrate is a product or by-product from the processing of wheat, more preferably the substrate is wheat bran and / or wheat middlings. Alternatively, dried distillers grains with solubles may be used.

[0036] By "feces" is preferably meant insect droppings (excreted feces) regardless of their development stage.

[0037] "Contaminants" preferably means undesirable organisms such as bacteria or fungi, or other insects such as moths, and / or undesirable elements produced by the latter, such as moth silk. Preferably, the contaminants are moths and / or moth silk. More preferably, the contaminants are moth silk.

[0038] Advantageously, a complex medium comprises at least two, preferably at least three components that are distinct from the live insect population.

[0039] Advantageously, the at least two, preferably at least three, different components are selected from substrate and faeces.

[0040] Advantageously, the complex medium comprises at least one insect which is at at least one different developmental stage from at least one insect of the live insect population.

[0041] Advantageously, the complex medium comprises at least one insect at at least one different developmental stage from the population of live insects, said population of live insects therefore being a population of insects at a given developmental stage, more particularly a population of insects of a given species at a given developmental stage, in which case it is possible in particular to distinguish a population of larvae from a population of nymphs or a population of nymphs from a population of adults.

[0042] Advantageously, the complex medium contains at least one dead insect.

[0043] "Quantitative determination" preferably means the determination of a quantitative value.

[0044] Preferably, a "performance indicator" or "KPI" ("key performance indicator") of live insect population rearing means an indicator for a live insect population, or a component of a complex medium, or a characteristic of a live insect population, or a characteristic of a component of a complex medium, at a given time.

[0045] By "quantitative value of a performance index" is preferably meant a quantitative value related to the population of live insects (e.g. weight percentage), or a component of a complex medium (e.g. weight percentage), or a characteristic of the population of live insects (e.g. average individual mass), or a characteristic of a component of a complex medium (e.g. weight percentage of moth silk aggregates) at a given time. The quantitative value of a component of a complex medium can be, for example, the amount of said component contained in the complex medium, preferably a mass, such as the weight percentage of said component relative to the total weight represented by the complex medium and the population of live insects.

[0046] Examples of quantitative values ​​of performance indicators include, but are not limited to, the quantitative value of a given insect population in the composite medium, expressed as a percentage by mass or weight of the total weight represented by the composite medium and the population of live insects, the quantitative value of a substrate in the composite medium, expressed as a percentage by mass or weight of the total weight represented by the composite medium and the population of live insects, the quantitative value of feces in the composite medium, expressed as a percentage by mass or weight of the total weight represented by the composite medium and the population of live insects, the quantitative value of a contaminant in the composite medium, expressed as a percentage by mass or weight of the total weight represented by the composite medium and the population of live insects, the individual average mass (IMM) of insects in the composite medium, expressed as a percentage by mass.

[0047] Thus, in a surprising and unexpected way, the inventors have developed a method that allows the quantification of specific populations of insects in a complex medium and other components of the medium. For example, the method according to the invention allows the quantification of different developmental stages of insects such as larvae, nymphs and / or adults, as well as dead insects, in a complex medium.

[0048] "Individual mean mass" (or "IMM") preferably means the individual mean mass of the insects.

[0049] Advantageously, the performance indicators are selected from insect population quantification, substrate quantification, faeces quantification, pollutant quantification and / or insect average individual mass quantification, preferably insect population quantification, insect average individual mass quantification and / or faeces quantification, even more preferably insect population quantification and / or faeces quantification, such as dead insects.

[0050] Advantageously, the performance indicator is fecal quantification.

[0051] Advantageously, the performance indicator is a quantification of the insect population.

[0052] Advantageously, the insect population consists of dead insects.

[0053] Advantageously, the population of insects consists of insects of a given developmental stage, which insects are advantageously alive.

[0054] By "illuminating" is preferably meant irradiating the medium with light comprising one or more wavelengths.

[0055] It is noted that according to the invention, the irradiation step is carried out on a complex medium comprising a population of live insects, i.e. on a complex medium in the presence of a population of live insects.

[0056] Advantageously, the irradiation step (i) is carried out with light comprising one or more wavelengths comprised in the range of 800 nm to 2200 nm, preferably 1000 to 1950 nm, and even more preferably 1100 to 1650 nm.

[0057] By "collection step" is preferably meant obtaining by a detector the energy intensity reflected by the medium as a function of the wavelength of the light emitted in step (i). The light in step (i) is advantageously decomposed light, preferably monochromatically decomposed light.

[0058] By "monochromatic light" is preferably meant light that contains one or more wavelengths of the same color.

[0059] Advantageously, the monochromatic light comprises a set of wavelengths, the shortest and longest of which are separated by at most 50 nm, preferably at most 25 nm, more preferably at most 15 nm, even more preferably at most 10 nm, for example about 8.75 nm.

[0060] Advantageously, the complex medium is irradiated with one or more monochromatic lights comprised in the range 420 nm to 2500 nm, preferably 800 to 2200 nm, preferably 1000 to 1950 nm, even more preferably 1100 to 1650 nm.

[0061] In the case of several monochromatic lights, steps (i) and (ii) are performed several times, each monochromatic light being emitted in succession.

[0062] According to a preferred embodiment, several irradiation steps (i) are performed successively to illuminate the sample with light having a wavelength in the range of 420 nm to 2500 nm, preferably 800 to 2200 nm, preferably 1000 to 1950 nm, even more preferably 1100 to 1650 nm, and a collection step (ii) is performed after each irradiation step.

[0063] By "range of light" it is preferably meant the sum of light covering all of the wavelengths included between the two end points of the range.

[0064] For example, to obtain a light range of X to Y nm, several successive irradiation steps (i) can be carried out via successive monochromatic light emission, where a first monochromatic light starts at X nm, the next monochromatic light is emitted at defined steps, for example 0.5 nm, up to Y (in this case the second monochromatic light is emitted at X+0.5 nm, the third at X+1 nm, etc.). Preferably, the steps are between 0.2 nm and 10 nm, more preferably between 0.5 nm and 5 nm, even more preferably about 0.5 nm.

[0065] By "reflectance spectrum" is preferably meant a spectrum representing the intensity of energy reflected by a culture medium as a function of the wavelength at which said culture medium is irradiated.

[0066] Advantageously, steps (i) and (ii) of the method according to the invention are repeated several times, preferably at least once, to improve the quality of the reflectance spectrum of the culture medium.

[0067] Preferably, after collection step (ii) of the method according to the invention, a diffuse reflectance spectrum of the culture medium is obtained.

[0068] "Diffuse reflectance spectrum" preferably means a reflectance spectrum representing the energy intensity reflected by a culture medium having the particularity of reflecting and dispersing the energy intensity in all directions of space according to the wavelength at which the culture medium is irradiated.

[0069] Advantageously, steps (i) and (ii) of the method according to the invention are carried out using a spectrophotometer, in particular a near infrared spectrophotometer (also called NIRS spectrometer for "near infrared spectroscopy"), such as a NIRS XDS Rapid Content Analyzer made by Metrohm, a NIRS Analyzer Processor made by Metrohm or a NIRS XDS Process Analyzer made by Metrohm.

[0070] The operation of a spectrophotometer is known to those skilled in the art. The light, generally polychromatic, coming for example from a tungsten or halogen lamp, is sent through an input slit to generate spatially coherent light. This light then reaches a diffraction grating, thus resulting in a spatial separation of the wavelengths that form it. Through diffuse reflection, a part of the signal is sent back to the detector. The output slit makes it possible to limit the passage of the wavelengths that reach the sample for the illumination step. This output slit therefore defines a bandwidth of a given size. As a result, depending on the size of the bandwidth, several wavelengths can reach the sample simultaneously. The wider the output slit, the greater the number of wavelengths that reach the sample and therefore the greater the variation in the resolution of the reflected spectrum. However, the smaller the width of the output slit, the finer the resolution of the reflected spectrum, but the lower the signal-to-noise ratio. To allow the successive passage of spatially separated wavelengths at the output slit, the diffraction grating may be mounted on an encoding motor that makes it possible to vary its angular position and thus define the acquisition steps, or the device comprises a diode array.

[0071] Preferably, the acquisition step of the spectrophotometer is between 0.2 nm and 8 nm, more preferably between 0.5 nm and 4 nm, and even more preferably about 0.5 nm.

[0072] Preferably, the bandwidth of the spectrophotometer is from 1 nm to 50 nm, more preferably from 5 to 20 nm, even more preferably from 8 nm to 10 nm, such as about 8.75 nm.

[0073] A spectrophotometer can include one or more detectors, preferably at least two detectors (also called sensors).

[0074] Preferably, the spectrophotometer comprises a silicon detector for detecting the intensity of energy reflected by the sample for wavelengths between 400 nm and 1100 nm, and a lead sulfate detector for detecting the intensity of energy reflected by the sample for wavelengths between 1100 nm and 2500 nm. Typically, when it is desired to obtain a diffuse reflectance spectrum of the sample, the spectrophotometer comprises a hemispherical reflecting dome positioned above the sample to direct all reflected light to the detector.

[0075] According to a preferred embodiment, the illumination step (i) and collection step (ii) are performed using a spectrophotometer and are advantageously repeated several times while changing the position of the sample, the succession of steps (i) and (ii) usually being called a scan.

[0076] Preferably, steps (i) and (ii) are repeated at least 12 times, more preferably at least 16 times, even more preferably at least 24 times, such as 32 times (or 32 scans). Similarly, the position of the sample is preferably changed at least 2 times, more preferably at least 4 times, such as 8 times. Even more preferably, steps (i) and (ii) are repeated 32 times per sample position and the sample position is changed between 4 and 8 times.

[0077] Steps (i) and (ii) of the method according to the invention can be carried out when the complex medium is stationary or in motion.

[0078] When the complex medium is moving, this is called dynamic mode. Conversely, when the complex medium is stationary, this is called static mode.

[0079] Regardless of whether steps (i) and (ii) are performed according to a static or dynamic mode, the spectrophotometer can be integrated into an optical device, including a motorized mobile support to which the spectrophotometer is coupled. This makes it possible to limit the number of spectrophotometers required to acquire data over a free surface of the complex medium sufficient to represent the entire free surface of the complex medium. The motorized mobile support, if necessary accompanied by a transport device in which a breeding tank containing the complex medium is placed, can generate a relative movement between the spectrophotometer and said complex medium. The acquisition area of ​​the spectrophotometer while performing steps (i) and (ii) thus sweeps a portion of the total free surface of the complex medium, which is much larger than the surface for which data can be acquired with a fixed sensor.

[0080] In the dynamic mode, as a simple example, the movable support can include a rotating arm parallel to the upper surface of a transport device (e.g., a belt conveyor) on which the tank containing the complex medium is placed. The arm is therefore also substantially parallel to the free surface of the complex medium. The optical device is configured such that the rotating arm performs one or more rotations (e.g., three rotations) while the breeding tank passes under the sensor at a constant speed imposed by the transport device on which it is placed.

[0081] In stationary mode, as a simple example, the mobile support includes means for translating the spectrophotometer in two orthogonal directions, called longitudinal and transverse, in a plane parallel to the free surface of the complex medium. A tank containing the complex medium is placed under the spectrophotometer and fixed. The optical device is configured such that the zone scanned by the acquisition area of ​​the spectrophotometer has a "racetrack" shape at the surface of the complex medium. The racetrack shape is an oval shape comprising two parallel longitudinal straight tracks connected by a curved and / or transverse part. To maximize the surface scanned by the acquisition area of ​​the spectrophotometer, the scan zone can have the shape of two racetracks, i.e. a scanned racetrack-shaped zone is included in another scanned racetrack-shaped zone with a larger dimension.

[0082] "Correlation step" means a step that allows the determination of a quantitative value of a given performance index according to the reflectance spectrum of the medium. Thus, the quantitative value of a given performance index can be correlated with said spectral data via the reflectance of the medium measured at one or more specific wavelengths.

[0083] Thus, the correlation step (iii) is carried out using the entire or part(s) of the reflectance spectrum. Preferably, the correlation step (iii) is carried out using the entire reflectance spectrum.

[0084] Advantageously, step (iii) of correlation is carried out using a previously established predictive model for this component.

[0085] Preferably, the model was previously established by regression analysis.

[0086] Regression analysis is a statistical method that allows to model the relationship between various variables, more specifically between a dependent variable and one or more independent variables, the value of which affects the value of the dependent variable. At the end of the analysis, a predictive model is thus established and represented by an equation. In the context of the present invention, the dependent variable is the quantitative value of a performance index, and the independent variable is the amount of light reflected by an insect or a component of a complex medium of a living insect population at one wavelength or over a given wavelength range.

[0087] Preferably, the regression analysis is a simple linear regression, a multiple linear regression or a partial least squares regression (called PLS). More preferably, the regression analysis is a PLS regression. These regression analyses are well known to those skilled in the art and can be carried out, for example, using commercially available software, preferably the software Vision sold by Metrohm.

[0088] To carry out the regression analysis, several samples of the medium must be constructed, preferably at least 50 samples, more preferably at least 100 samples, even more preferably at least 150 samples, such as 180 samples. Each sample is then irradiated with light comprising one or more wavelengths comprised in one or more ranges from 420 nm to 2500 nm, preferably from 800 to 2200 nm, preferably from 1000 to 1950 nm, even more preferably from 1100 nm to 1650 nm, and then the light reflected by each sample is collected to obtain the reflectance spectrum of each sample. Each sample is also analyzed by a reference method to determine a reference quantitative value in the population of live insects or in a given component of the complex medium. Preferably, this reference method comprises manually separating the various components of the complex medium from the population of live insects and then weighing them. Alternatively and even more preferably, in the case of an artificial complex medium, the mass of each component present in the medium is known, and therefore there is no need to manually separate the various components of the medium. The same applies to live insect populations. Exemplary methods are described in the Examples below.

[0089] The PLS regression can be applied to a few wavelengths, or to a broad spectral range, or to the entire reflectance spectrum.

[0090] Furthermore, PLS regression allows the determination of independent variables that cannot be observed (e.g. have little or no physical meaning), which are called latent variables or factors. PLS regression therefore allows the creation of new variables.

[0091] According to a preferred embodiment, the regression analysis is a PLS regression, which includes the selection of several factors to be included in the equation of the predictive model. This selection of the number of factors is important, because if there are not enough factors, this leads to under-modeling and a decrease in the accuracy of the model, and conversely, if there are too many factors, it leads to over-modeling and a decrease in the robustness of the model. Advantageously, the optimal number of factors to be used in the PLS regression is determined by calculating the prediction residual sum of squares (PRESS) for each number of factors, and thus the PRESS value represents the prediction error and needs to be minimized. Preferably, the PRESS value for each number of factors is calculated via an external validation set or cross-validation, more preferably via cross-validation.

[0092] Therefore, the number of factors with the minimum PRESS value is selected to establish the equation of the predictive model. Because the PRESS value depends on the components of the medium, the minimum value selected is not the same depending on the model established for a given component of the medium.

[0093] Preferably, during cross-validation, all data of the samples involved in establishing the predictive model are grouped together, then data for each sample in turn is removed and the model is created using the spectra and values ​​of the remaining samples determined during the relevant reference analysis. Cross-validation is described in more detail in the Examples below.

[0094] Advantageously, the method further comprises, after the collecting step (ii) and before the correlating step (iii), a step of mathematical processing of the reflectance spectra, including first derivatives, second derivatives, Savinsky type processing, and / or SNV type normalization ("standard normal variate").

[0095] Preferably, the mathematical processing step is an SNV type normalization followed by first or second derivative, or a Savinsky type processing followed by first or second derivative.

[0096] Advantageously, the mass percentage of insects in the larval stage in the complex medium of the method according to the invention is between 0% and 90%, preferably between 20% and 85%, more preferably between 25% and 80% of the total weight of the complex medium and the sample of live insects.

[0097] Advantageously, the mass percentage of adult insects in the complex medium of the method according to the invention is between 0% and 100%, preferably between 10% and 90%, more preferably between 20% and 80% of the total weight of the complex medium and the sample of live insects.

[0098] Advantageously, the mass percentage of nymphal stage insects in the complex medium of the method according to the invention is between 0% and 50%, preferably between 0.1% and 20%, of the total weight of the complex medium and the sample of live insects.

[0099] Advantageously, the mass percentage of dead insects in the complex medium of the method according to the invention is between 0% and 50%, preferably between 0.1% and 40%, of the total weight of the complex medium and live insect sample.

[0100] Advantageously, the mass percentage of the substrate in the complex medium of the method according to the invention is between 0% and 60%, preferably between 5% and 50%, of the total weight of the complex medium and the sample of live insects.

[0101] Advantageously, the mass percentage of faeces in the complex medium of the method according to the invention is between 0% and 50%, preferably between 5% and 40%, of the total weight of the complex medium and the live insect sample.

[0102] Advantageously, the mass percentage of contaminants such as moth silk aggregates in the complex medium of the method according to the invention is between 0% and 90%, preferably between 0% and 50%, more preferably between 0% and 40% of the total weight of the complex medium and the live insect sample.

[0103] Advantageously, the individual average mass of insects in the complex medium of the method according to the invention is calibrated according to two subgroups: - a group having an estimated mass of 0.5 mg to 50 mg, preferably 0.7 mg to 48 mg, more preferably 1 mg to 45 mg; - Another subgroup of 35 to 200 mg, preferably 38 to 185 mg, more preferably 40 mg to 175 mg.

[0104] The present invention also relates to the use of the method according to the invention, preferably using a spectrophotometer as defined above, for carrying out monitoring of the growth (or development or growth) of a population of live insects, preferably for controlling the quality of insect rearing, for determining the efficiency of a sorting, preferably insect sorting, machine and / or for determining the efficiency of a device for feeding a population of live insects.

[0105] The present invention also relates to the use of a spectrophotometer for the quantitative measurement of at least one performance index in a complex medium.

[0106] Advantageously, the spectrophotometer is as defined above.

[0107] Advantageously, the performance indicator is as defined above.

[0108] Advantageously, the spectrophotometer is used to carry out monitoring of the growth (or development or growth) of a population of live insects, preferably for controlling the quality of rearing of a population of live insects, to determine the efficiency of a machine for sorting, preferably for sorting a population of live insects, and / or to determine the efficiency of a device for feeding a population of live insects.

[0109] The invention is illustrated in a non-limiting manner by the examples and figures. [Brief description of the drawings]

[0110] [Figure 1] FIG. 1 includes two figures. The figure on the left (A) shows the mass of live insects in the larval stage in sampling A calculated from the training set according to the mass of live insects in the larval stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the larval stage in sampling A calculated from the external validation set according to the mass of live insects in the larval stage determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0111] [Diagram 2] FIG. 2 includes two figures. The figure on the left (A) shows the mass of live insects in the larval stage in sampling B calculated from the training set according to the mass of live insects in the larval stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the larval stage in sampling B calculated from the external validation set according to the mass of live insects in the larval stage determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0112] [Diagram 3]FIG. 3 includes two figures. The figure on the left (A) shows the mass of live insects in the nymph stage in sampling B calculated from the training set according to the mass of live insects in the nymph stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the nymph stage in sampling B calculated from the external validation set according to the mass of live insects in the nymph stage determined by the reference analysis, the masses being expressed as a weight percentage of the total weight of the sample.

[0113] [Figure 4] FIG. 4 includes two figures. The figure on the left (A) shows the mass of live insects in the adult stage in sampling C calculated from the training set according to the mass of live insects in the adult stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the adult stage in sampling C calculated from the external validation set according to the mass of live insects in the adult stage determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0114] [Diagram 5] FIG. 5 includes two figures. The figure on the left (A) shows the mass of live insects in the nymph stage in sampling C calculated from the training set according to the mass of live insects in the nymph stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the nymph stage in sampling C calculated from the external validation set according to the mass of live insects in the nymph stage determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0115] [Figure 6]FIG. 6 includes two figures. The figure on the left (A) shows the mass of live insects in the adult stage in sampling D calculated from the training set according to the mass of live insects in the adult stage determined by the reference analysis, and the figure on the right (B) shows the mass of live insects in the adult stage in sampling D calculated from the external validation set according to the mass of live insects in the adult stage determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0116] [Figure 7] FIG. 7 includes two figures: the figure on the left (A) shows the mass of feces in sampling A calculated from the training set according to the mass of feces determined by the reference analysis, and the figure on the right (B) shows the mass of feces in sampling A calculated from the external validation set according to the mass of feces determined by the reference analysis, the masses being expressed as a weight percentage of the total weight of the sample.

[0117] [Figure 8] FIG. 8 includes two figures: the figure on the left (A) shows the mass of the feces in sampling B calculated from the training set according to the mass of the feces determined by the reference analysis, and the figure on the right (B) shows the mass of the feces in sampling B calculated from the external validation set according to the mass of the feces determined by the reference analysis, the masses being expressed as a weight percentage of the total weight of the sample.

[0118] [Figure 9] FIG. 9 includes two figures: the figure on the left (A) shows the mass of the nutrient substrate in sampling A calculated from the training set according to the mass of the nutrient substrate determined by the reference analysis, and the figure on the right (B) shows the mass of the nutrient substrate in sampling A calculated from the external validation set according to the mass of the nutrient substrate determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0119] [Figure 10]FIG. 10 includes two figures: the left figure (A) shows the mass of the nutrient substrate in sampling B calculated from the training set according to the mass of the nutrient substrate determined by the reference analysis, and the right figure (B) shows the mass of the nutrient substrate in sampling B calculated from the external validation set according to the mass of the nutrient substrate determined by the reference analysis, the masses being expressed as weight percentages.

[0120] [Figure 11] Figure 11 includes two figures: the left (A) shows the mass of the nutrient substrate in sampling C calculated from the training set according to the mass of the nutrient substrate determined by the reference analysis, and the right (B) shows the mass of the nutrient substrate in sampling C calculated from the external validation set according to the mass of the nutrient substrate determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0121] [Figure 12] FIG. 12 includes two figures: the figure on the left (A) shows the mass of dead insects in sampling A calculated from the training set according to the mass of dead insects determined by the reference analysis, and the figure on the right (B) shows the mass of dead insects in sampling A calculated from the external validation set according to the mass of dead insects determined by the reference analysis, the mass expressed as a weight percentage of the total weight of the sample.

[0122] [Figure 13] FIG. 13 includes two figures: the figure on the left (A) shows the mass of dead insects in sampling B calculated from the training set according to the mass of dead insects determined by the reference analysis, and the figure on the right (B) shows the mass of dead insects in sampling B calculated from the external validation set according to the mass of dead insects determined by the reference analysis, the mass being expressed as a weight percentage of the total weight of the sample.

[0123] [Figure 14]FIG. 14 includes two figures: the figure on the left (A) shows the mass of dead insects in sampling C calculated from the training set according to the mass of dead insects determined by the reference analysis, and the figure on the right (B) shows the mass of dead insects in sampling C calculated from the external validation set according to the mass of dead insects determined by the reference analysis, the mass expressed as a weight percentage of the total weight of the sample.

[0124] [Figure 15] FIG. 15 includes two figures: the figure on the left (A) shows the mass of dead insects in sampling D calculated from the training set according to the mass of dead insects determined by the reference analysis, and the figure on the right (B) shows the mass of dead insects in sampling D calculated from the external validation set according to the mass of dead insects determined by the reference analysis, the mass expressed as a weight percentage of the total weight of the sample.

[0125] [Figure 16] FIG. 16 includes two figures: the left side (A) shows the mass of the contaminant (moth silk aggregates) in sampling E calculated from the training set according to the mass of the contaminant (moth silk aggregates) determined by the reference analysis, and the right side (B) shows the mass of the contaminant (moth silk aggregates) in sampling E calculated from the external validation set according to the mass of the contaminant (moth silk aggregates) determined by the reference analysis, the mass being expressed as a weight percentage relative to the total weight of the sample.

[0126] [Figure 17] Figure 17 includes two figures. The left side (A) shows the individual mean mass (IMM) of insects in sampling E calculated from the training set according to the individual mean mass determined by the reference analysis ranging from 1.03 mg to 42.86 mg, and the right side (B) shows the individual mean mass (IMM) of insects in sampling E calculated from the external validation set according to the individual mean mass determined by the reference analysis ranging from 1.03 mg to 42.86 mg.

[0127] [Figure 18] Figure 18 includes two figures. The left side (A) shows the individual mean mass (IMM) of insects in sampling E calculated from the training set according to the individual mean mass determined by the reference analysis ranging from 40.85 mg to 173.46 mg, and the right side (B) shows the individual mean mass (IMM) of insects in sampling E calculated from the external validation set according to the individual mean mass determined by the reference analysis ranging from 40.85 mg to 173.46 mg. EXAMPLES

[0128] Example 1: Creation of a model implemented in the method according to the invention I- Sampling, Spectral Acquisition, and Reference Analysis --sampling - "Industrial tank" sampling Four samples (A, B, C and D) were prepared from samples of Tenebrio molitor at various developmental stages (larvae, adults and / or nymphs). 2 Rearing tanks were artificially filled by mixing live insects with substrate, dead insects and / or feces.

[0129] The substrate is a nutrient substrate called "INSECTUS" commercially available from MIJTEN.

[0130] The characteristics of each sample are shown in Table 1.

[0131] [Table 1] * The samples were carried out on a special tank with a metal grid at the bottom.

[0132] The mass percentages are given relative to the total weight of the components introduced into the tank.

[0133] The term "grate" refers to a metal grid whose mesh allows laid eggs to pass through but retains adult insects on its surface.

[0134] - "Test Tank" Sampling Sampling E was also prepared using a sample of the larval stage of Tenebrio molitor. 2 In a tank having the dimensions of 100 x 100 mm, the larval stage of the mealworm was reared for 4 weeks in a complex medium containing a nutrient substrate called "INSECTUS" marketed by the company MIJTEN.

[0135] During these four weeks of rearing, the insects consumed the nutrient substrate and produced feces, and also allowed contaminants such as moths to infest the insect rearing tanks.

[0136] The characteristics of sample E are shown in Table 2.

[0137] [Table 2] "Moth silk aggregate" means moth silk aggregated with complex medium and / or parts of a moth.

[0138] Sampling E was used to quantify two performance indicators, namely the individual mean mass (IMM) of insects expressed in mg and the mass percentage of contaminants, in particular the mass percentage of moth silk aggregates. Furthermore, 234 samples of this type of sampling E were used to develop a model for the "contaminants" KPI, while 386 samples of this type of sampling E ranged from 1.03 mg to 42.86 mg and 245 samples were used to develop a model for the "IMM" KPI, ranging from 40.85 mg to 173.46 mg.

[0139] --Spectral acquisition - Acquisition of spectra in "industrial tank" sampling Spectra were acquired in diffuse reflectance mode using a NIRS (standing for near infrared spectroscopy) spectrometer, and the acquisition conditions are summarized in Table 3.

[0140] [Table 3] For each tank, four different positions are analyzed by the NIRS Analyzer Pro spectrometer, so that each of the four spectra obtained at the end of the acquisition corresponds to an average of 32 scans (the averaging is performed automatically by the device).

[0141] The four spectra obtained over the four different locations are then averaged to obtain only a single average spectrum representative of the scanned industrial tank.

[0142] - Acquisition of spectra in "test tank" sampling A NIRS (standing for Near Infrared Spectroscopy) spectrometer was used to obtain spectra in diffuse reflectance mode under the conditions shown in Table 3 above.

[0143] For each tank, eight different positions are analyzed by the NIRS Analyzer Pro spectrometer. Each test tank is moved manually across these eight different positions. Thus, each of the eight spectra obtained at the end of the acquisition corresponds to an average of 32 scans.

[0144] The eight spectra obtained over the eight different locations are then averaged to obtain only a single average spectrum representative of the scanned test tank.

[0145] --Reference analysis - Reference analysis in "industrial tank" sampling Considering that the rearing tanks were prepared manually by introducing the components (feces, live insects, dead insects, nutrient substrate) into the tanks before the scan with the NIRS spectrometer, the amount of each component is known. Indeed, before acquiring the spectrum, the mass of each component is measured and expressed as a percentage of the total weight of the components introduced into the tank. These values ​​constitute the reference values ​​for each industrial tank.

[0146] - Reference analysis in "test tank" sampling After obtaining the spectra, a portion (representing 1 / 8 of each rearing tank) is manually sorted using screens of different mesh sizes according to the different components (faeces, live insects, dead insects and contaminants). Once each component is isolated, the mass of the latter is measured and expressed as a mass percentage relative to the total weight of the component introduced into the tank, except for the "IMM" KPI, which is expressed in milligrams.

[0147] II-Determination of quantitative model The software Vision used for the acquisition of the spectra mentioned in section I is also used to process the data and develop the quantitative models.

[0148] --Mathematical pre-processing of the spectra Various mathematical pretreatments, such as first derivative, second derivative, SNV (Standard Normal Variate) type normalization, or Savinsky type treatment, were determined to apply to the spectra obtained in I.

[0149] --Developing quantitative models: Training set After their mathematical pre-processing, the spectra are correlated via partial least squares regression or PLS with the values ​​determined during the reference analysis in section I. PLS involves the selection of a certain number of factors (VL). If there are not enough factors, this leads to under-modeling and reduced accuracy, conversely, too many factors leads to over-modeling and reduced robustness of the model.

[0150] For each performance indicator to be determined (faeces, populations of a given insect, IMM, nutrient substrate and pollutants), the database (spectra and values ​​determined during the relevant reference analysis) was divided into two sets: a learning set consisting of spectra and values ​​determined during the relevant reference analysis for 75% of the samples, and For the remaining samples (25% of the samples), an external validation set consisting of spectra and values ​​determined during the relevant reference analysis.

[0151] A model is established using all of the data in the training set by setting the maximum number of factors for PLS in the software to 16, and then establishing the above model by varying the number of factors from 1 to 16. The optimal number of factors to be used for PLS is then determined through cross-validation, the latter being performed on the entire training set.

[0152] During cross-validation, the data for each sample (spectra and values ​​determined during the associated reference analysis) are sequentially removed from the training set and a model is established using the spectra and values ​​determined during the associated reference analysis of the remaining samples. First, the data of the sample, i.e. its spectrum and its associated values ​​determined during the reference analysis, is removed from the learning set, and then the model is established using the spectra and associated values ​​determined during the reference analysis of the remaining samples.

[0153] The resulting model is then used to analyze the spectrum of the removed sample, i.e. to predict the quantitative value.

[0154] - Finally, the value predicted by the model (for the spectrum of the removed sample) is subtracted from the value determined during the reference analysis associated with this spectrum, then these differences are squared and summed. This allows to calculate the prediction residual sum of squares (PRESS) for each factor of the PLS. The PRESS value thus represents the prediction error.

[0155] The data of the sample that was removed from the training set is then reintroduced into it, and the above steps are then repeated by removing the data of another sample.

[0156] The number of factors with the minimum PRESS value is thus determined, which therefore makes it possible to obtain, for each performance index to be determined, the optimal number of factors to be used for PLS.

[0157] The best performing established models (with parameters described in detail below) are retained and their robustness is assessed on an external validation set.

[0158] The various parameters used to select the best model during training set creation are: -R 2 Multiple correlation coefficient: It corresponds to the agreement between the data predicted by the model and the values ​​determined during the reference analysis. It must tend to approach 1.

[0159] - SEC Standard Error of Calibration: This value is equivalent to the standard deviation and is a statistical parameter that reflects the upper limit of the accuracy of future predictions. It should be as low as possible.

[0160] -Standard error of SECV cross-validation: this value corresponds to the SEC during cross-validation carried out on the learning set during PLS. Preferably, SECV should have a value as close as possible to SEC. We have determined that in order for the model to have acceptable uncertainty, SECV should be no more than two times the standard deviation relative to the value determined during the reference analysis.

[0161] --Quantitative model development: external validation The data of the samples of the external validation set represent the range of values ​​determined during the reference analysis. This data makes it possible to evaluate the robustness of the model developed with the best performing training set by studying the following parameters: -R 2 Pearson correlation coefficient: This corresponds to the agreement between the data predicted by the model for the external validation set and the values determined during the reference analysis. It should tend towards 1.

[0162] -SEP Standard error of prediction: This value is a statistical parameter that reflects the residuals (or prediction errors) obtained from the predictions for the external validation set. Ideally, SEP ≤ SECV. If not (SEP > SECV), it must be ensured that SEP is statistically below SECV (or SECV < SEP ≤ 1.3 × SECV).

[0163] -Bias: Bias is the average value of the residuals (or prediction errors) calculated based on the reference values of the components of the validation set. Bias should be close to 0, which indicates that the deviations are randomly distributed. However, a significant bias value indicates a systematic error, for example, a change in the device or reference analysis. This bias can then be corrected by software.

[0164] -RPD: RPD stands for "Relative Percent Difference" and is a statistical indicator that compares the variability of the reference data used for the calibration accuracy. Therefore, this value can be used to evaluate the quality of the predictions made by the calibration. An RPD greater than 1.4, preferably greater than 2, is considered to be associated with good predictions and thus the model is considered sufficient. Conversely, an RPD value less than 1.4 indicates that the model is not appropriate.

[0165] Select the model with the best performance and thus determine the modeling parameters for the final calibration.

[0166] Example 2: Results of the calibration test - Living insects -- Development of a model for the quantification of living insects Following the method described in Example 1, the development of a model for determining the amount of live insects at a particular development stage in a given complex medium (various samplings) was established.

[0167] Therefore, appropriate mathematical pre-processing was applied to each spectrum obtained.

[0168] The mathematical pretreatment suitable for the performance indicators for the larval stage insects in sampling A was determined to be the normalization of the SNV types, followed by the first derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 13 factors (Figure 1, A and B, respectively).

[0169] The mathematical pretreatment suitable for the performance indicators for the larval stage insects in sampling B was determined to be the first derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 12 factors (Figure 2, A and B, respectively).

[0170] Furthermore, for this same sampling B, it was determined that a suitable mathematical pretreatment for the performance indicators for the nymphal stage insects was the normalization of the SNV types, followed by the second derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 9 factors (Figure 3, A and B, respectively).

[0171] The mathematical pretreatment suitable for the performance indicators for adult stage insects in sampling C was determined to be the first derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 11 factors (Figure 4, A and B, respectively).

[0172] In this same sampling C, it was determined that the mathematical pretreatment suitable for the performance index on nymphal stage insects was the normalization of the SNV types, followed by the second derivative over the spectral range from 1100 nm to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 8 factors (Figure 5, A and B, respectively).

[0173] Finally, it was determined that the mathematical pretreatment suitable for the performance indicators for the adult stage insects in sampling D was the first derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 11 factors (Figure 6, A and B, respectively).

[0174] A summary of the results is shown in Table 4 below.

[0175] [Table 4] It should be noted that in case B above, if the larvae constitute a population of live insects in the sense of the present invention, the nymphs are a component of the complex medium as insects of at least one different developmental stage from at least one insect of the population of live insects, or vice versa. Similarly, in case C above, if the nymphs constitute a population of live insects in the sense of the present invention, the adults are a component of the complex medium as insects of at least one different developmental stage from at least one insect of the population of live insects, or vice versa. However, these two populations consist of live insects.

[0176] In conclusion, for each model obtained, the R 2 The correlation coefficient is very high.

[0177] According to each of the models developed, the correlation between the calculated values ​​and those determined by the reference analysis is very high.

[0178] For each model for a given population of living insects, the RPD values ​​are always greater than 2 and very high, thus indicating a good quality of prediction of the models across the developed range.

[0179] The measurement uncertainty represents the maximum error associated with predicting the mass of a given population of living insects. The values ​​obtained are relatively low.

[0180] These results confirm that analysis by NIRS spectroscopy is a suitable method to determine the quantitative values ​​of live insect populations, especially in complex media containing feces, substrates and / or dead insects.

[0181] Furthermore, the results show that the method according to the invention makes it possible to distinguish different populations of living insects, even if the latter are mobile in complex media, without prior manipulation of the sample.

[0182] The results of sampling D demonstrate that the method according to the invention can be used when insects are reared on a metal grid and therefore the presence of the metal grid does not affect the results.

[0183] Example 3: Calibration Test Results - Feces --Development of a model for quantification of feces The development of a model for determining fecal volume was carried out according to the method described in Example 1.

[0184] An appropriate mathematical pre-processing is applied to each spectrum obtained.

[0185] In particular, it was determined that a suitable mathematical pretreatment for the performance indicators for feces in sampling A was the normalization of the SNV types, followed by the second derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 9 factors (Figure 7, A and B, respectively).

[0186] The mathematical pretreatment suitable for the performance indicators for the feces in sampling B was determined to be the normalization of the SNV types followed by the first derivative over the spectral range from 1100 nm to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 14 factors (Figure 8, A and B respectively).

[0187] A summary of the results is shown in Table 5 below.

[0188] [Table 5] In conclusion, for each model obtained, the R 2 The correlation coefficient is very high.

[0189] According to each of the models developed, the correlation between the calculated values ​​and those determined by the reference analysis is very high.

[0190] The RPD values ​​are always very high, greater than 2, thus indicating a good quality of prediction of the model over the developed range.

[0191] The measurement uncertainty represents the maximum error associated with the prediction of the fecal mass. The obtained values ​​are relatively low.

[0192] These results confirm that analysis by NIRS spectroscopy is a suitable method to determine the quantitative values ​​of insect feces, especially in complex media containing live insects of different development stages, nutrient substrates and / or dead insects.

[0193] Example 4: Calibration Test Results - Nutrient Substrates --Development of a model for quantifying nutrient substrates The development of a model for determining the amount of nutrient substrate in a complex medium was carried out according to the method described in Example 1.

[0194] An appropriate mathematical pre-processing is applied to each spectrum obtained.

[0195] The mathematical pretreatment suitable for the performance indicators on nutrient substrates in sampling A was determined to be the normalization of the SNV types, followed by the second derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 10 factors (Figure 9, A and B, respectively).

[0196] The mathematical pretreatment suitable for the performance indicators on nutrient substrates in sampling B was determined to be the normalization of the SNV types, followed by the first derivative over the spectral range of 1100-1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 10 factors (Figure 10, A and B, respectively).

[0197] For sampling C, after investigation it was determined that the appropriate mathematical pretreatment was the first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 10 factors (Figure 11, A and B, respectively). A summary of the results is shown in Table 6 below.

[0198] [Table 6] For each model obtained, the R 2 The correlation coefficient is very high.

[0199] For each model for nutrient substrates, the RPD values ​​were always very high, greater than 2, thus indicating a good quality of prediction of the models across the developed range.

[0200] The measurement uncertainty represents the maximum error associated with the prediction of the mass of the nutrient substrate. The obtained values ​​are relatively low.

[0201] These results confirm that analysis by NIRS spectroscopy is a suitable method to reliably determine quantitative values ​​of nutrient substrates in various complex media containing live insects of various developmental stages, feces and / or dead insects.

[0202] Example 5: Calibration Test Results - Dead Insects --Development of a model for quantifying dead insects The development of a model for determining the amount of dead insects in complex media was carried out according to the method described in Example 1.

[0203] An appropriate mathematical pre-processing is applied to each spectrum obtained.

[0204] The mathematical pretreatment suitable for the spectra originating from sampling A was determined to be the normalization of the SNV types, followed by the first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 11 factors (Figure 12, A and B, respectively).

[0205] The mathematical pretreatment suitable for the spectra originating from sampling B was determined to be the normalization of the SNV types followed by the second derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 13 factors (Figure 13, A and B, respectively).

[0206] The appropriate mathematical pretreatment of the spectra acquired at sampling C was determined to be the second derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with six factors (Figure 14, A and B, respectively).

[0207] Finally, it was determined that the appropriate mathematical pretreatment of the spectra acquired at sampling D was the normalization of the SNV types, followed by the first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 10 factors (Figure 15, A and B, respectively).

[0208] A summary of the results is shown in Table 7 below.

[0209] [Table 7] In conclusion, for each model obtained, the R 2 The correlation coefficient is very high.

[0210] For each model for dead insects, the RPD values ​​were always very high, greater than 2, thus indicating a good quality of prediction of the models across the developed range.

[0211] The measurement uncertainty represents the maximum error associated with the prediction of the amount of dead insects. The obtained values ​​are relatively low.

[0212] These results confirm that analysis by NIRS spectroscopy is a suitable method to reliably determine the quantitative values ​​of dead insects, especially in complex media containing live insects of various development stages, nutrient substrates and / or feces.

[0213] Furthermore, these results demonstrate that NIRS spectroscopy analysis is a suitable method to distinguish dead from live insects in complex media.

[0214] Example 6: Calibration Test Results - Contaminants -- Development of models for quantification of pollutants The development of a model for determining the amount of contaminants in complex media was carried out according to the method described in Example 1.

[0215] It was determined that the appropriate mathematical pretreatment of the spectra obtained from sampling E was a Savinsky type treatment followed by the first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 12 factors (Figure 16, A and B, respectively).

[0216] A summary of the results is shown in Table 8 below.

[0217] [Table 8] In conclusion, the R 2 The correlation coefficient is very high.

[0218] The correlation between the calculated values ​​and those determined by the reference analysis is very strong.

[0219] For the models for the pollutants, the RPD values ​​were greater than 2, indicating good quality of prediction of the models over the developed range.

[0220] These results confirm that analysis by NIRS spectroscopy is a suitable method to reliably determine quantitative values ​​of contaminants, especially in complex media containing live insects, nutrient substrates and feces.

[0221] Example 7: Calibration Test Results - IMM --Development of a model to determine the average mass of each individual in the range of 1.03 mg to 42.86 mg. The development of a model for determining the average individual mass of insects in complex media was carried out according to the method described in Example 1.

[0222] In particular, it was determined that the mathematical pretreatment suitable for the IMM index at sampling E in the range of 1.03 mg to 42.86 mg was the first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 14 factors (Figure 17, A and B, respectively).

[0223] A summary of the results is shown in Table 9 below.

[0224] [Table 9] In conclusion, the R 2 The correlation coefficient is very high.

[0225] In fact, the correlation between the calculated values ​​and those determined by the reference analysis is very strong.

[0226] For the model for the determination of the individual mean mass of insects, the RPD values ​​were greater than 2, indicating a good quality of prediction of the model over the range from 1.03 mg to 42.86 mg.

[0227] --Development of a model to determine the average mass of each individual in the range of 40.85 mg to 173.46 mg. The development of a model for determining the average individual mass of insects in complex media was carried out according to the method described in Example 1.

[0228] The mathematical pretreatment suitable for the IMM index at sampling E for the development of a model in the range of 40.85 mg to 173.46 mg was determined to be the normalization of the SNV types, followed by the first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set make it possible to approve a model with 15 factors (Figure 18, A and B, respectively).

[0229] A summary of the results is shown in Table 10 below.

[0230] [Table 10] In conclusion, the R 2 The correlation coefficient is very high.

[0231] In fact, the correlation between the calculated values ​​and those determined by the reference analysis is very strong.

[0232] For the model for the determination of the individual mean mass of insects, the RPD values ​​were greater than 2, indicating a good quality of prediction of the model over the range from 40.85 mg to 173.46 mg.

[0233] These results confirm that analysis by NIRS spectroscopy is a suitable method to non-invasively determine the average individual mass of insects, especially in complex media containing live insects of various developmental stages, nutrient substrates and / or dead insects.

Claims

1. 1. A non-destructive method for quantitatively determining at least one performance indicator of the rearing of a population of live insects in a complex medium, said method comprising: (i) irradiating the complex medium containing the live insect population with light comprising one or more wavelengths in the range of 420 nm to 2500 nm; (ii) collecting light reflected by the complex medium to obtain a reflectance spectrum of the medium; (iii) correlating the spectrum with a quantitative value of at least one performance indicator.

2. 10. The method of claim 1, wherein the complex medium comprises at least two distinct components other than the live insect population.

3. The method of claim 2 , wherein the at least two different components are selected from substrate and feces.

4. 4. The method according to any one of claims 1 to 3, wherein the complex medium comprises at least one insect at at least one different developmental stage from at least one insect of the live insect population.

5. The method according to any one of claims 1 to 4, wherein the complex medium comprises at least one dead insect.

6. 6. The method according to any one of claims 1 to 5, wherein the performance indicator is selected from insect population quantification, substrate quantification, feces quantification, pollutant quantification and / or average individual mass quantification of the insects.

7. The method of claim 6 , wherein the performance indicator is a quantification of the feces.

8. The method of claim 6 , wherein the performance indicator is a quantification of an insect population.

9. 9. The method of claim 8, wherein the insect population consists of the dead insects.

10. 9. The method of claim 8, wherein the population of insects consists of live insects of a given developmental stage.

11. 11. The method according to any one of claims 1 to 10, further comprising a step of mathematical processing of the reflectance spectra after the collecting step (ii) and before the correlating step (iii), comprising first derivative, second derivative, Savinsky type processing and / or SNV type normalization.

12. Use of the method according to any one of claims 1 to 11 for carrying out monitoring of the growth of a population of live insects, for determining the efficiency of a sorter and / or for determining the efficiency of a device for feeding the population of live insects.

13. Use of a spectrophotometer for the quantitative determination of at least one performance index of a population of live insects in a complex medium.

14. 14. The use according to claim 13, wherein the performance indicator is selected from quantification of the insect population, quantification of substrate, quantification of feces, quantification of pollutants and / or quantification of the average individual mass of the insects.

Images