Non-destructive quantitative determination of at least one performance indicator in rearing a population of live insects in a complex medium
Patent Information
- Application Number
- EP2022847617
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-12-28
- Filing Date
- 2022-12-27
- Publication Date
- 2024-11-06
AI Technical Summary
Current methods for monitoring insect populations in insect farming are manual, tedious, and lack precision, making them incompatible with industrial-scale farming and potentially damaging to the insects.
A non-destructive method using light irradiation within specific wavelengths (420 nm to 2500 nm) to obtain a reflection spectrum, correlated with quantitative performance indicators, allowing for precise monitoring of insect populations and growth conditions without damaging the insects.
Enables accurate, automated, and non-invasive determination of performance indicators in complex environments, improving monitoring efficiency and reducing damage to insects, while providing low uncertainty measurements.
Smart Images

Figure 00000033_0000 
Figure 00000033_0001 
Figure 00000033_0002
Abstract
Description
[0001] NON-DESTRUCTIVE QUANTITATIVE DETERMINATION OF AT LEAST ONE PERFORMANCE INDICATOR OF A BREEDING OF A POPULATION OF INSECTS LIVING IN A COMPLEX ENVIRONMENT
[0002] STATE OF THE ART
[0003] The present invention belongs to the field of analysis, and more particularly, it relates to a non-destructive method for determining a quantitative value of at least one performance indicator.
[0004] Insect farming has experienced significant growth in recent years. Insect farming has many advantages, whether for the agro-industry, where insects are an interesting source of protein, or for other industrial sectors, where insects are also a source of chitin, which can be transformed into chitosan by deacetylation, and which has numerous applications, such as in cosmetics, medicine and pharmaceuticals, dietetics and food, or even for water treatment.
[0005] It is therefore necessary to set up industrial models adapted to insects, allowing their breeding to be monitored automatically, in order to maintain optimal growth conditions.
[0006] This monitoring is possible thanks to the quantification of performance indicators (or, in English, “Key Performance indicators” or “KPis”).
[0007] By "performance indicator" (or, in English, "Key Performance indicator" or "KPi") we mean, preferably, information intended to be quantified which makes it possible to measure and evaluate the results of one or more actions, to monitor the evolution of a performance and / or to analyze a situation.
[0008] The methods currently used are manual and tedious (sorting and counting insect populations, visual inspection of breeding tanks) and incompatible with industrial-scale breeding. In addition, these methods lack precision and can sometimes damage growing insects, particularly when handling them.
[0009] There is therefore a need to automate the monitoring of the growth of a population of live insects and the monitoring of the equipment for sorting and feeding a population of live insects.
[0010] The present invention overcomes these technical problems and relates to a non-destructive method for determining a quantitative value of at least one performance indicator. Furthermore, the present inventors have shown, surprisingly and unexpectedly, that the method according to the invention makes it possible to determine a quantitative value of at least one performance indicator with very low uncertainties in a complex medium. SUMMARY OF THE INVENTION
[0011] The present invention relates to a non-destructive method for quantitatively determining at least one performance indicator of a breeding of a population of living insects in a complex environment, said method comprising the following steps:
[0012] (i) a step of irradiating the complex medium comprising the population of living insects with light comprising one or more wavelengths in a range from 420 nm to 2500 nm,
[0013] (ii) a step of collecting the light reflected by the complex medium in order to obtain a reflection spectrum of the medium, and
[0014] (iii) a step of correlating said spectrum to a quantitative value of at least one performance indicator.
[0015] The present invention also relates to the use of the method as defined above for monitoring the growth of the population of live insects, for determining the efficiency of a sorting machine and / or for determining the efficiency of the equipment for feeding the population of live insects.
[0016] The present invention also relates to the use of a spectrophotometer for the quantitative determination of at least one performance indicator of a breeding of a population of living insects in a complex environment.
[0017] DETAILED DESCRIPTION
[0018] In the context of this application, unless otherwise stipulated, the ranges of values indicated are understood to include limits.
[0019] A first subject relates to a non-destructive method for quantitatively determining at least one performance indicator of breeding a population of living insects in a complex environment, said method comprising the following steps:
[0020] (i) a step of irradiating the complex medium comprising the population of living insects with light comprising one or more wavelengths in a range from 420 nm to 2500 nm,
[0021] (ii) a step of collecting the light reflected by the complex medium in order to obtain a reflection spectrum of the medium, and
[0022] (iii) a step of correlating said spectrum to a quantitative value of at least one performance indicator.
[0023] By "non-destructive process" we preferably mean a set of non-invasive steps that allow the integrity of the living insect population to be preserved.
[0024] By "insect" is preferably meant an insect at any stage of development, such as an adult, larval, nymph and / or egg stage, preferably adult, larval and / or nymph stage. This term therefore covers any stage of evolution (also called growth or development stage) from the egg or ootheca to the adult insect, including the larva and the nymph such as the pupa. By "insect population" is preferably meant a group of at least two insects of the same or different species and / or at the same or different stage of development.
[0025] In particular, an insect population within the meaning of the present invention relates to a population of individualized insects. Therefore, the insect population as referred to by the invention does not correspond to an insect which incubates eggs inside its abdomen (ovoviviparous insect).
[0026] Preferably, an insect population, and more particularly the live insect population, is a set of at least two insects of the same species at an identical stage of development. Examples of insect populations include, but are not limited to, a set of larvae, a set of nymphs, or a set of adult insects.
[0027] The term "larva" preferably means the larval stage of insects and therefore the juvenile stage of the insect, which includes the maggot for Diptera and the caterpillar for Lepidoptera, as well as the wingless stages in Orthoptera. This term covers any stage of development of the larva.
[0028] By "nymph" is preferably meant the stage of metamorphosis of the insect and in particular the intermediate stages between the larva and the imago, which includes the pupa for Diptera, the nymph for Coleoptera, the chrysalis for Lepidoptera and, where appropriate, an intermediate stage during which certain physiological (prepupa) or behavioral modifications of the individuals appear, such as significant sclerotization of the cuticle for Diptera.
[0029] By "adult insect" (or imago) we preferably mean the mature stage of the insect, and in particular the last stage of development of the insect, which ensures the reproduction of the species.
[0030] Advantageously, the insects are chosen from the group of Coleoptera, Diptera, Lepidoptera, Orthoptera, Neuroptera, Hymenoptera, Dictyoptera including in particular Blattoptera, including Isoptera, and Mantoptera, Phasmoptera, 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, Diptera belong to the suborder Brachycera. Preferably, Lepidoptera belong to the suborder Ditrysia, more preferably to the superfamily Pyraloidea. Preferably, Neuroptera belong to the suborder Hemerobiiformia. More preferably, Neuroptera belong to the families Mantispidae, Ithonidae, Chrysopidae, Hemerobiidae, or their mixtures.Preferably, the beetles belong to the infraorder Cucujiformia, in particular to the families Tenebrionidae, Coccinellidae, Cerambycidae, Dryophthoridae, or mixtures thereof. More preferably, the beetles are chosen from Tenebrio molitor, Alphitobius diaperinus, Zophobas morio, Tenebrio obscurus, Tribolium castaneum, Rhynchophorus ferruginous, and mixtures thereof, even more preferably Tenebrio molitor, Alphitobius diaperinus, and mixtures thereof. The insects targeted by the invention therefore preferably belong to the group of beetles and more particularly to the family Tenebrionidae. Preferably, the insects targeted by the invention belong to the species Tenebrio molitor and / or Alphitobius diaperinus.
[0031] By “complex environment” we preferably mean:
[0032] - a set comprising at least two, preferably at least three, different constituents other than the population of living insects, and / or
[0033] - at least one insect at at least one stage of development different from at least one insect in the living insect population, and / or
[0034] - at least one dead insect.
[0035] By "set comprising at least two, preferably at least three, different constituents" is meant, preferably, a set comprising at least two, preferably at least three, constituents composed at least in part of identical or similar molecules but at different concentrations.
[0036] By "identical or similar molecules" is preferably meant chemical molecules having identical or similar chemical structures.
[0037] For example, these are molecules that have common chemical structures exhibiting the same vibrations, when excited by the infrared source, of the hydrogen atom as those of the family of molecules being measured (proteins, lipids, saccharides, chitins, water). For example, if hydrogen is bonded with nitrogen (via N-H2-COOH group) then the molecules will belong to the protein family.
[0038] Thus, proteins have identical or similar chemical structures.
[0039] Similarly, monosaccharides, oligosaccharides, and polysaccharides have identical or similar chemical structures; triglycerides and fatty acids have identical or similar chemical structures; chitin and chitin derivatives have identical or similar chemical structures.
[0040] By "chitin derivatives" is preferably meant polysaccharide derivatives comprising N-acetyl-glucosamine units and D-glucosamine units, in particular chitin-polypeptide copolymers (sometimes referred to as "chitin-polypeptide composites").
[0041] Indeed, in the complex medium according to the invention, each constituent may in particular comprise two or more of the compounds chosen from:
[0042] - proteins,
[0043] - monosaccharides, oligo and polysaccharides,
[0044] - triglycerides and fatty acids, and / or
[0045] - chitin and chitin derivatives.
[0046] More particularly, each compound above is present in at least two chosen from the constituents of the complex medium and the population of living insects.
[0047] Preferably, the complex medium is an insect rearing medium. Examples of constituents may be, but are not limited to, a substrate, and / or dead insects of the same or different species than the live insect population, and / or live insects of a different species than at least one insect in the live insect population, preferably than the live insect population, and / or live insects at a different developmental stage than at least one insect in the live insect population, and / or droppings (also called frass), and / or contaminants.
[0048] By "substrate" we preferably mean a nutrient substrate.
[0049] By "nutrient substrate" is preferably meant a nutrient substrate or a mixture of nutrient substrates, intended to be consumed by insects. It may be a solid or liquid nutrient substrate. The solid nutrient substrate may be a mixture of solid products or co-products resulting from the processing of cereals, oilseeds, oilseeds and / or protein crops. This substrate may be supplemented with vitamins or minerals. By "co-product" is preferably meant an unavoidable material created during a manufacturing process of a product of interest. The solid substrate may be a product or co-product resulting from the processing of wheat (soft wheat, durum wheat), corn, barley, rice, triticale, oats, sorghum, rye, spelt, millet, quinoa, buckwheat, rapeseed, sunflower, flax, soybeans and / or peas.Preferably, the solid substrate is a product or co-product from wheat processing, more preferably, the substrate is wheat bran and / or wheat middlings. Alternatively, dried distillers grains with solubles may be used.
[0050] By "droppings" we preferably mean droppings (feces) of insects whatever their stage of development.
[0051] By "contaminant" is preferably meant an undesirable living organism such as a bacterium or a fungus, or another insect such as a moth, and / or an undesirable element produced by the latter such as moth silk. Preferably, the contaminant is a moth and / or moth silk. More preferably, the contaminant is moth silk.
[0052] Advantageously, the complex medium comprises at least two, preferably at least three, different constituents other than the population of living insects.
[0053] Advantageously, the at least two, preferably at least three, different constituents are chosen from a substrate and excrement.
[0054] Advantageously, the complex medium comprises at least one insect at at least one stage of development different from at least one insect in the population of living insects.
[0055] Advantageously, the complex medium comprises at least one insect at at least one stage of development different from the population of living insects, said population of living insects then being a population of insects at a given stage of development, more particularly, said population of living insects being a population of insects of a given species at a given stage of development. Such a case makes it possible in particular to distinguish a population of larvae from a population of nymphs or a population of nymphs from a population of adults. Advantageously, the complex medium comprises at least one dead insect.
[0056] By "quantitative determination" is preferably meant the determination of a quantitative value.
[0057] Preferably, by "performance indicator" or "KPi" ("Key Performance Indicator -" in English) of a breeding of a population of live insects, we mean an indicator relating to the population of live insects or to a constituent of the complex environment or to a characteristic of the population of live insects or to a characteristic of a constituent of the complex environment at a given time.
[0058] By "quantitative value of a performance indicator" is meant, preferably, a quantitative value associated with the population of live insects (for example, a percentage by weight) or with a constituent of the complex medium (for example, a percentage by weight) or with a characteristic of the population of live insects (for example, an average individual mass) or a characteristic of a constituent of the complex medium (for example, a percentage by weight of moth silk aggregate) at a given time. The quantitative value of a constituent of the complex medium may for example be the quantity of said constituent included in the complex medium, and preferably the mass quantity such as a percentage by weight of said constituent relative to the total weight represented by the complex medium and the population of live insects.
[0059] Examples of a quantitative value of a performance indicator include, but are not limited to, the quantitative value of a given insect population in the complex medium expressed as a mass or as a percentage by weight relative to the total weight represented by the complex medium and the population of live insects, the quantitative value of a substrate in the complex medium expressed as a mass or as a percentage by weight relative to the total weight represented by the complex medium and the population of live insects, the quantitative value of droppings in the complex medium expressed as a mass or as a percentage by weight relative to the total weight represented by the complex medium and the population of live insects, the quantitative value of contaminants in the complex medium expressed as a mass or as a percentage by weight relative to the total weight represented by the complex medium and the population of live insects,the average individual mass (“AIM”) of insects in the complex environment expressed as mass.,
[0060] Thus, surprisingly and unexpectedly, the present inventors have developed a method for quantifying specific insect populations and other environmental constituents in a complex medium. For example, the method according to the invention makes it possible to quantify insects at different stages of development such as larvae, nymphs and / or adult insects, and dead insects in a complex medium.
[0061] By "average individual mass" (or "AIM") is preferably meant the average individual mass of the insects.
[0062] Advantageously, the performance indicator is chosen from a quantification of an insect population, a substrate, droppings, a contaminant and / or an individual average mass of the insects, preferably a quantification of an insect population, an individual average mass of the insects and / or droppings, and even more preferably a quantification of an insect population such as dead insects, and / or droppings.
[0063] Advantageously, the performance indicator is a quantification of the droppings.
[0064] Advantageously, the performance indicator is a quantification of an insect population.
[0065] Advantageously, the insect population is made up of dead insects.
[0066] Advantageously, the insect population consists of insects at a given stage of development, the insects being advantageously alive.
[0067] By "irradiation step" is preferably meant the emission of light comprising one or more wavelengths onto the medium.
[0068] It will be noted that according to the invention, the irradiation step is carried out on the complex medium comprising the population of living insects, that is to say on the complex medium in the presence of the population of living insects.
[0069] Advantageously, the irradiation step (i) is carried out with light comprising one or more wavelengths in the range from 800 to 2200 nm, preferably from 1000 to 1950 nm, and even more preferably from 1100 to 1650 nm.
[0070] By "collection step" is preferably meant the acquisition via a detector of the energy intensity reflected by the medium as a function of the wavelengths of the light emitted in step (i). The light from step (i) is advantageously a decomposed light, preferably a light decomposed into monochromatic lights.
[0071] By "monochromatic light" is preferably meant light comprising one or more wavelengths of the same color.
[0072] Advantageously, a monochromatic light comprises a set of wavelengths of which the shortest wavelength and the longest wavelength are spaced apart by a maximum of 50 nm, preferably by a maximum of 25 nm, more preferably by a maximum of 15 nm, even more preferably by a maximum of 10 nm, such as of the order of 8.75 nm.
[0073] Advantageously, the complex medium is irradiated with one or more monochromatic lights in a range from 420 nm to 2500 nm, preferably from 800 to 2200 nm, preferentially from 1000 to 1950 nm, and even more preferentially from 1100 to 1650 nm.
[0074] In the case of several monochromatic lights, steps (i) and (ii) are carried out several times with each monochromatic light emitted successively. According to a preferred embodiment, several irradiation steps (i) are carried out successively so as to irradiate the sample with a range of light whose wavelengths are between 420 nm and 2500 nm, preferably from 800 to 2200 nm, preferentially from 1000 to 1950 nm, and even more preferentially from 1100 to 1650 nm, a collection step (ii) being carried out for each irradiation step.
[0075] By "range of light" is preferably meant a sum of lights covering all wavelengths between the two limits of the range.
[0076] For example, in order to obtain a light range from X to Y nm, several sequential irradiation steps (i) can be carried out by emitting sequential monochromatic lights, the first monochromatic light starting at X nm, and the following monochromatic lights being emitted with a defined step, for example 0.5 nm (in this case, the second monochromatic light is emitted at X+0.5 nm, the third at X+1 nm, ...), up to Y. Preferably, this step is between 0.2 nm and 10 nm, more preferably, between 0.5 nm and 5 nm, even more preferably, of the order of 0.5 nm.
[0077] By "reflection spectrum" is preferably meant a spectrum representing the energy intensity reflected by the medium as a function of the wavelengths with which said medium was irradiated.
[0078] Advantageously, steps (i) and (ii) of the method according to the invention are repeated several times, and preferably at least once, in order to improve the quality of the reflection spectrum of the medium.
[0079] Preferably, at the end of step (ii) of collection of the method according to the invention, a diffuse reflection spectrum of the medium is obtained.
[0080] By "diffuse reflection spectrum" is preferably meant a reflection spectrum representing the energy intensity reflected by the medium having the particularity of reflecting and dispersing the energy intensity in all directions of space, as a function of the wavelengths with which said medium was irradiated.
[0081] Advantageously, steps (i) and (ii) of the method according to the invention are carried out using a spectrophotometer, in particular using a near-infrared spectrophotometer (also called a NIRS spectrometer for “Near-infrared spectroscopy”), such as for example the NIRS XDS Rapid Content Analyzer from Metrohm, the NIRS Analyzer Pro from Metrohm or the NIRS XDS Process Analyzer from Metrohm.
[0082] The operation of a spectrophotometer is known to those skilled in the art: light, generally polychromatic, coming for example from a tungsten or halogen lamp, is sent through an entrance slit in order to generate spatially coherent light. This light then arrives on a diffraction grating, leading to the spatial separation of the wavelengths constituting it. By diffuse reflection, part of the signal is then returned to the detectors. An exit slit limits the passage of wavelengths arriving at the sample for the irradiation step; this exit slit then defines a bandwidth of a given size. Consequently, several wavelengths can arrive simultaneously at the sample, depending on the size of the bandwidth. The wider the exit slit, the greater the number of wavelengths arriving at the sample, and therefore the more the resolution of the reflection spectrum is impaired.On the other hand, the narrower the width of the output slit, the finer the resolution of the reflection spectrum, but the more the signal-to-noise ratio deteriorates. In order to allow the sequential passage of the spatially separated wavelengths in the output slit, the diffraction grating can be mounted on an encoder motor which allows its angular position to be varied, and thus an acquisition step to be defined, or the device is equipped with a diode strip.
[0083] Preferably, the acquisition step of the spectrophotometer is between 0.2 nm and 8 nm, more preferably between 0.5 nm and 4 nm, even more preferably of the order of 0.5 nm.
[0084] Preferably, the bandwidth of the spectrophotometer is between 1 nm and 50 nm, more preferably between 5 and 20 nm, even more preferably between 8 nm and 10 nm, such as for example of the order of 8.75 nm.
[0085] The spectrophotometer may comprise one or more detectors, preferably at least two detectors (also called sensors).
[0086] Preferably, the spectrophotometer comprises a silicon detector for detecting the energy intensity reflected by the sample for wavelengths between 400 nm and 1100 nm, and a lead sulfide detector for detecting the energy intensity reflected by the sample for wavelengths between 1100 nm and 2500 nm. Usually, in the case where it is desired to obtain a diffuse reflection spectrum of the sample, the spectrophotometer comprises a hemispherical reflector dome placed above the sample in order to direct all the reflected light towards the detector.
[0087] According to a preferred embodiment, the steps (i) of irradiation and (ii) of collection are carried out using a spectrophotometer and repeated several times, advantageously by modifying the position of the sample. The sequence of steps (i) and (ii) is commonly called a scan.
[0088] 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 for example 32 times (i.e. 32 scans). Similarly, the position of the sample is preferably modified at least 2 times, more preferably at least 4 times, such as for example 8 times. Even more preferably, steps (i) and (ii) are repeated 32 times per position of the sample and the position of the sample is modified between 4 and 8 times.
[0089] Steps (i) and (ii) of the method according to the invention can be carried out when the complex medium is stationary or in motion.
[0090] When the complex medium is in motion, we speak of dynamic mode. Conversely, when the complex medium is stationary, we speak of static mode.
[0091] Whether steps (i) and (ii) are carried out in a static mode or in a dynamic mode, the spectrophotometer can be integrated into an optical device comprising a motorized mobile support to which the spectrophotometer is linked. This makes it possible to limit the number of spectrophotometers necessary to acquire data on a free surface of the complex medium sufficient to be representative of the entire free surface of the complex medium. The motorized mobile support, in association where appropriate with a conveyor device on which the breeding tank containing the complex medium is placed, makes it possible to create a relative movement between the spectrophotometer and said complex medium. Thus, the acquisition field of the spectrophotometer scans during steps (i) and (ii) a portion of the total free surface of the complex medium that is much larger than the surface on which data could be acquired with a fixed sensor.
[0092] In dynamic mode, and simply by way of example, the mobile support may comprise a rotating arm parallel to the upper surface of the conveying device (for example the 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 so that the rotating arm performs one or more rotations (for example three rotations) while a breeding tank passes under the sensor at a constant speed imposed by the conveying device on which it is placed.
[0093] In static mode, and simply as an example, the mobile support comprises 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. The tank containing the complex medium is positioned and immobilized under the spectrophotometer. The optical device is configured so that the area scanned by the acquisition field of the spectrophotometer has a “racetrack” shape, on the surface of the complex medium. A racetrack shape is an oblong shape comprising two parallel longitudinal straight paths connected by turns and / or transverse portions. To maximize the area scanned by the acquisition field of the spectrophotometer, the scanned area may have the shape of two racetracks, namely a racetrack-scanned area included in another larger racetrack-scanned area.
[0094] By "correlation step" is meant a step for determining a quantitative value of a given performance indicator based on the reflection spectrum of the medium. The quantitative value of a given performance indicator can therefore be correlated to the data of said spectrum via the reflectance of the medium measured at one or more specific wavelengths.
[0095] Thus, the correlation step (iii) is carried out by exploiting the whole, or part, or parts of the reflection spectrum. Preferably, the correlation step (iii) is carried out by exploiting the whole of the reflection spectrum.
[0096] Advantageously, correlation step (iii) is carried out using a predictive model previously established for said constituent.
[0097] Preferably, said model has been previously established by regression analysis.
[0098] A regression analysis is a statistical method that allows modeling relationships between different variables, more particularly between a dependent variable and (one) of the independent variable(s), the value of the independent variables affecting the value of the dependent variable. At the end of the analysis, a predictive model is therefore established and is represented by an equation. In the context of the present invention, a dependent variable is a quantitative value of a performance indicator and an independent variable is the quantity of light reflected by the insect of the population of living insects or a constituent of the complex environment for a given wavelength or over a range of wavelengths.
[0099] Preferably, the regression analysis is a simple linear regression, a multilinear regression, or a partial least squares regression (PLS). More preferably, the regression analysis is a PLS regression. These regression analyses are well known to those skilled in the art and can, for example, be carried out using commercial software, preferably using the Vision software marketed by Metrohm.
[0100] In order to carry out a regression analysis, several medium samples must be constituted, preferably at least 50 samples, more preferably at least 100 samples, even more preferably at least 150 samples such as for example 180. Then, each sample is irradiated with light comprising one or more wavelengths included in a range(s) from 420 nm to 2500 nm, preferably from 800 to 2200 nm, preferably from 1000 to 1950 nm and even more preferably from 1100 nm to 1650 nm, then the light reflected by each sample is collected in order to obtain a reflection spectrum for each sample. Each sample is also analyzed by a reference method in order to determine a quantitative reference value in a population of living insects or in a given constituent of the complex medium.Preferably, this reference method consists of manually separating the different constituents of the complex medium and the population of living insects and then weighing them. Alternatively and even more preferably, for artificial complex media, the mass of each constituent present in the medium is known, so it is not necessary to manually separate the different constituents of the medium. The same applies to the population of living insects. Examples of the method are described in the examples below.
[0101] PLS regression can be applied to multiple wavelengths, even over a wide spectral range or even across the entire reflection spectrum.
[0102] In addition, PLS regression allows the determination of independent variables that are not observable (for example, that have little or no physical meaning), they are called latent variables or factors. PLS regression therefore allows the creation of new variables
[0103] According to a preferred embodiment, the regression analysis is a PLS regression which involves the choice of a number of factors to be included in the equation of the predictive model. This choice of the number of factors is important since, if there are not enough factors, this leads to under-modeling and a degradation of the accuracy of the model, and conversely, too many factors induce over-modeling and a degradation of the robustness of the model. Advantageously, the optimal number of factors to be used for the PLS regression is determined by calculating the predictive residual sum of squares (PRESS or Prediction Residual error - Sum of Squares) for each number of factors, the PRESS value therefore representing the prediction error and having to be minimal. Preferably, the PRESS value of each number of factors is calculated using an external validation set or cross-validation, more preferably, using cross-validation.Thus, the number of factors having a minimum PRESS value will be retained to establish the equation of the predictive model. The PRESS value depends on the constituent of the environment, the minimum value to be retained is not the same depending on the model to be established for a given constituent of the environment.
[0104] Preferably, in cross-validation, all data from the samples that participated in the establishment of a predictive model are pooled, then the data for each sample are removed in turn and a model is built with the spectra and values, determined during the associated reference analyses, of the remaining samples. Cross-validation is further described in the examples below.
[0105] Advantageously, the method further comprises, after the collection step (ii) and before the correlation step (iii), a step of mathematical processing of the reflection spectrum comprising a first derivative, a second derivative, a Savinsky type processing and / or a SNV (“Standard Normal Variate”) type normalization.
[0106] Preferably, the mathematical processing step is an SNV-type normalization followed by a first or second derivative or a Savinsky-type processing followed by a first or second derivative.
[0107] Advantageously, the mass percentage of insects at the larva stage in a complex medium of the method according to the invention is between 0% and 90%, preferably between 20% and 85%, and more preferably between 25% and 80% of the total weight of the sample of the complex medium and the living insects.
[0108] Advantageously, the mass percentage of insects at the adult stage in a complex medium of the method according to the invention is between 0% and 100%, preferably between 10% and 90%, and more preferably between 20% and 80% of the total weight of the sample of the complex medium and the living insects.
[0109] Advantageously, the mass percentage of insects at the nymph stage in a 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 sample of the complex medium and the living insects.
[0110] Advantageously, the mass percentage of dead insects in a 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 sample of the complex medium and the living insects.
[0111] Advantageously, the mass percentage of substrate in a 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 sample of the complex medium and the living insects.
[0112] Advantageously, the mass percentage of excrement in a 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 sample of the complex medium and the living insects.
[0113] Advantageously, the mass percentage of contaminants such as moth silk aggregates in a complex medium of the method according to the invention is between 0% and 90%, preferably between 0% and 50%, and more preferably between 0% and 40% of the total weight of the sample of the complex medium and the living insects.
[0114] Advantageously, the average individual mass of insects in a complex medium of the method according to the invention is calibrated according to two subgroups:
[0115] - a group in which the mass is estimated between 0.5 mg and 50 mg, preferably between 0.7 and 48 mg, more preferably between 1 mg and 45 mg,
[0116] - another subgroup between 35 and 200 mg, preferably between 38 and 185 mg, more preferably between 40 mg and 175 mg.
[0117] The present invention also relates to the use of the method according to the invention for monitoring the growth (or evolution or development) of the population of live insects and preferably monitoring the growth (or evolution or development) of the population of live insects for quality control of insect breeding, for determining the efficiency of a sorting machine, and preferably for sorting insects, and / or for determining the efficiency of the equipment for feeding the population of live insects, preferably using a spectrophotometer as defined above.
[0118] The present invention also relates to the use of a spectrophotometer, for the quantitative determination of at least one performance indicator in a complex medium.
[0119] Advantageously, the spectrophotometer is as defined previously.
[0120] Advantageously, the performance indicator is as defined previously.
[0121] Advantageously, the spectrophotometer is used to monitor the growth (or evolution or development) of the population of live insects, and preferably the monitoring of the growth (or evolution or development) of the population of live insects for quality control of breeding of the population of live insects, to determine the efficiency of a sorting machine, and preferably of sorting of the population of live insects, and / or to determine the efficiency of the equipment for feeding the population of live insects.
[0122] The present invention is illustrated, in a non-limiting manner, by the examples and the figures.
[0123] DESCRIPTION OF FIGURES
[0124] [Fig. 1] Figure 1 includes two diagrams: the left one (A) represents the mass quantity of live insects at the larval stage in sample A calculated from the training set as a function of the mass quantity of live insects at the larval stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the larval stage in sample A calculated from the external validation set as a function of the mass quantity of live insects at the larval stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight on the total weight of the sample. [Fig.2] Figure 2 includes two diagrams: the left one (A) represents the mass quantity of live insects at the larval stage in sample B calculated from the training set as a function of the mass quantity of live insects at the larval stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the larval stage in sample B calculated from the external validation set as a function of the mass quantity of live insects at the larval stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0125] [Fig. 3] Figure 3 includes two diagrams: the left one (A) represents the mass quantity of live insects at the nymph stage in sample B calculated from the training set as a function of the mass quantity of live insects at the nymph stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the nymph stage in sample B calculated from the external validation set as a function of the mass quantity of live insects at the nymph stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0126] [Fig. 4] Figure 4 includes two diagrams: the left one (A) represents the mass quantity of live insects at the adult stage in sampling C calculated from the training set as a function of the mass quantity of live insects at the adult stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the adult stage in sampling C calculated from the external validation set as a function of the mass quantity of live insects at the adult stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0127] [Fig. 5] Figure 5 includes two diagrams: the left one (A) represents the mass quantity of live insects at the nymph stage in the C sample calculated from the training set as a function of the mass quantity of live insects at the nymph stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the nymph stage in the C sample calculated from the external validation set as a function of the mass quantity of live insects at the nymph stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight on the total weight of the sample.
[0128] [Fig. 6] Figure 6 includes two diagrams: the left one (A) represents the mass quantity of live insects at the adult stage in sampling D calculated from the training set as a function of the mass quantity of live insects at the adult stage determined by the reference analyses, and the right one (B) represents the mass quantity of live insects at the adult stage in sampling D calculated from the external validation set as a function of the mass quantity of live insects at the adult stage determined by the reference analyses, the mass quantities being expressed as a percentage by weight on the total weight of the sample. [Fig.7] Figure 7 includes two diagrams: the left one (A) represents the mass quantity of droppings in sample A calculated from the training set as a function of the mass quantity of droppings determined by the reference analyses, and the right one (B) represents the mass quantity of droppings in sample A calculated from the external validation set as a function of the mass quantity of droppings determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0129] [Fig. 8] Figure 8 includes two diagrams: the left one (A) represents the mass quantity of droppings in sample B calculated from the training set as a function of the mass quantity of droppings determined by the reference analyses, and the right one (B) represents the mass quantity of droppings in sample B calculated from the external validation set as a function of the mass quantity of droppings determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0130] [Fig. 9] Figure 9 includes two diagrams: the left one (A) represents the mass quantity of a nutrient substrate in sampling A calculated from the training set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, and the right one (B) represents the mass quantity of a nutrient substrate in sampling A calculated from the external validation set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0131] [Fig. 10] Figure 10 includes two diagrams: the left one (A) represents the mass quantity of a nutrient substrate in sampling B calculated from the training set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, and the right one (B) represents the mass quantity of a nutrient substrate in sampling B calculated from the external validation set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0132] [Fig. 11] Figure 11 includes two diagrams: the left one (A) represents the mass quantity of a nutrient substrate in the C sample calculated from the training set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, and the right one (B) represents the mass quantity of a nutrient substrate in the C sample calculated from the external validation set as a function of the mass quantity of a nutrient substrate determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0133] [Fig. 12] Figure 12 includes two diagrams: the left one (A) represents the mass quantity of dead insects in sampling A calculated from the training set as a function of the mass quantity of dead insects determined by the reference analyses, and the right one (B) represents the mass quantity of dead insects in sampling A calculated from the external validation set as a function of the mass quantity of dead insects determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0134] [Fig. 13] Figure 13 includes two diagrams: the left one (A) represents the mass quantity of dead insects in sample B calculated from the training set as a function of the mass quantity of dead insects determined by the reference analyses, and the right one (B) represents the mass quantity of dead insects in sample B calculated from the external validation set as a function of the mass quantity of dead insects determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0135] [Fig. 14] Figure 14 includes two diagrams: the left one (A) represents the mass quantity of dead insects in sampling C calculated from the training set as a function of the mass quantity of dead insects determined by the reference analyses, and the right one (B) represents the mass quantity of dead insects in sampling C calculated from the external validation set as a function of the mass quantity of dead insects determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0136] [Fig. 15] Figure 15 includes two diagrams: the left one (A) represents the mass quantity of dead insects in sampling D calculated from the training set as a function of the mass quantity of dead insects determined by the reference analyses, and the right one (B) represents the mass quantity of dead insects in sampling D calculated from the external validation set as a function of the mass quantity of dead insects determined by the reference analyses, the mass quantities being expressed as a percentage by weight of the total weight of the sample.
[0137] [Fig. 16] Figure 16 includes two diagrams: the left one (A) represents the mass quantity of contaminants (moth silk aggregates) in the E sample calculated from the training set as a function of the mass quantity of contaminants (moth silk aggregates) determined by the reference analyses, and the right one (B) represents the mass quantity of contaminants (moth silk aggregates) in the E sample calculated from the external validation set as a function of the mass quantity of contaminants (moth silk aggregates) determined by the reference analyses, the mass quantities being expressed as a percentage by weight on the total weight of the sample.
[0138] [Fig- 17] Figure 17 includes two diagrams: the left one (A) represents the individual mean mass (“IMM”) of insects calculated from the training set in sampling E as a function of the individual mean mass determined by the reference analyses in the range from 1.03 mg to 42.86 mg, and the right one (B) represents the individual mean mass (“IMM”) of insects in sampling E calculated from the external validation set as a function of the individual mean mass determined by the reference analyses in the range from 1.03 mg to 42.86 mg,
[0139] [Fig. 18] Figure 18 includes two plots: the left plot (A) shows the individual mean mass (“IMM”) of insects calculated from the training set in sampling E as a function of the individual mean mass determined by the reference analyses in the range from 40.85 mg to 173.46 mg, and the right plot (B) shows the individual mean mass (“IMM”) of insects in sampling E calculated from the external validation set as a function of the individual mean mass determined by the reference analyses in the range from 40.85 mg to 173.46 mg.
[0140] EXAMPLES
[0141] Example 1: Development of the model implemented in the method according to the invention
[0142] I- Sampling, acquisition of spectra and reference analyses
[0143] • Sampling
[0144] - “Industrial bin” sampling
[0145] Four samples (A, B, C and D) were prepared from samples of Tenebrio Molitor at different stages of development (larva, adult and / or nymph). Rearing tanks of 2.8 m 2 were artificially filled by mixing live insects with substrate, dead insects and / or droppings.
[0146] The substrate is a nutrient substrate, called “INSECTUS” marketed by the company MIJTEN.
[0147] The characteristics of each sample are presented in Table 1.
[0148] [Table 1] metallic.
[0149] The mass percentages are given in relation to the total weight of the constituents introduced into the tank.
[0150] The term "grid" refers to a metal grid whose mesh allows laying eggs to pass through while retaining adults on its surface.
[0151] - “Test tank” sampling
[0152] Sample E was also prepared from samples of Tenebrio Molitor at the larval stage. In tanks with a dimension of 0.24 m 2 , Tenebrio Molitor insects at the larva stage were raised for 4 weeks in a complex medium comprising a nutrient substrate called “INSECTUS” marketed by the company MIJTEN.
[0153] During these 4 weeks of rearing, the insects consumed nutrient substrate and produced droppings. In addition, contaminants such as mites were allowed to infest the insect rearing tanks. The characteristics of sampling E are presented in Table 2.
[0154] Table 2]
[0155] By "moth silk aggregates" is meant moth silks agglomerated with part of the complex medium and / or moths.
[0156] E-sampling was used to quantify two performance indicators, namely the individual average mass (“IMM”), expressed in mg, of insects and the mass percentage of contaminants and in particular the mass percentage of moth silk aggregates. In addition, 234 samples of this type of E-sampling were used to develop a model for the “contaminants” KPi while 386 samples of this type of E-sampling were used to develop a model for the “IMM” KPi over the range 1.03 mg to 42.86 mg and 245 samples over the range 40.85 mg to 173.46 mg.
[0157] • Acquisition of spectra
[0158] - Acquisition of spectra on “industrial tank” samples
[0159] A diffuse reflection spectrum was acquired using a NIRS (Near Infrared Spectroscopy) spectrometer, with acquisition conditions summarized in Table 3.
[0160] Table 3]
[0161] On each tank, 4 different positions are analyzed using the NIRS Analyzer Pro spectrometer. Thus, each of the 4 spectra obtained at the end of the acquisition corresponds to an average of 32 scans (average automatically performed by the equipment).
[0162] The 4 spectra obtained at the 4 different positions are then averaged to give only one average spectrum representative of the scanned industrial tank.
[0163] - Acquisition of spectra on “test tank” samples
[0164] A diffuse reflection spectrum was acquired using a NIRS (Near Infrared Spectroscopy) spectrometer under the conditions shown in Table 3, above.
[0165] On each tank, 8 different positions are analyzed using the NIRS Analyzer Pro spectrometer. Each test tank is manually moved to these 8 different positions. Thus, each of the 8 spectra obtained at the end of the acquisition corresponds to an average of 32 scans.
[0166] The 8 spectra obtained at the 8 different positions are then averaged to give only one average spectrum representative of the scanned test tank. Reference analyses
[0167] - Reference analyses on “industrial tank” samples
[0168] Since the breeding tanks were prepared manually by introducing constituents (droppings, live insects, dead insects, nutrient substrate) into the tanks before scanning by the NIRS spectrometer, the quantities of each constituent are known. Indeed, before acquiring the spectra, the mass of each constituent is measured and expressed as a percentage of the total weight of the constituents introduced into the tank. These values constitute the reference values for each industrial tank.
[0169] - Reference analyses on “test tank” samples
[0170] After the acquisition of the spectra, a part (representing 1 Z8 èmeof each breeding tank) is sorted manually with sieves of different mesh sizes according to the different constituents (droppings, live insects, dead insects and contaminants). Once each constituent is isolated, the mass of the latter is measured and is expressed as a mass percentage relative to the total weight of the constituents introduced into the tank, except for the KPi “IMM” which is expressed in milligrams.
[0171] II- Determination of quantitative models
[0172] The Vision software used for the acquisition of the spectra mentioned in point I, is also used to process the data and develop quantitative models.
[0173] • Mathematical preprocessing of spectra
[0174] Different mathematical pretreatments were determined to be applied to the spectra obtained in I, such as: first derivative, second derivative, SNV (Standard Normal Variate) type normalization or Savinsky type treatment.
[0175] • Development of quantitative models: learning game
[0176] After their mathematical preprocessing, the spectra are correlated with the values determined during the reference analyses at point I, using a partial least squares regression (PLS). PLS involves the choice of a certain number of factors (VL): if there are not enough factors, this leads to under-modeling and a degradation of the accuracy, and conversely, too many factors induce over-modeling and a degradation of the robustness of the model.
[0177] For each performance indicator to be determined (droppings, given insect population, IMM, nutrient substrate and contaminants) the database (spectra and values determined during the associated reference analyses) was divided into two sets:
[0178] - a learning set consisting of the spectra and values determined during the associated reference analyses, for 75% of the samples, and
[0179] - an external validation set consisting of the spectra and values determined during the associated reference analyses, for the remaining samples (25% of the samples). Models are established with all the data in the training set, by setting the maximum number of PLS factors to 16 in the software which then establishes said models by varying the number of factors from 1 to 16. Then, the optimal number of factors to use for the PLS is determined using cross-validation. This is carried out with the entire training set.
[0180] In cross-validation, the data for each sample (spectrum and values determined during the associated reference analyses) are removed in turn from the training set and a model is developed with the spectra and values, determined during the associated reference analyses, of the remaining samples:
[0181] - First, the data of a sample, i.e. its spectrum and associated values determined during the reference analyses, are removed from the training set and then a model is established with the spectra and values determined during the associated reference analyses of the remaining samples.
[0182] - Then the resulting model is used to analyze the spectrum of the sample that had been removed, i.e. to predict quantitative values.
[0183] - Finally, the values predicted by the model (for the spectrum of the removed sample) are subtracted from the values determined during the reference analyses associated with this spectrum, then their differences are squared and summed; this allows the calculation of the predictive residual sum of squares (PRESS or Prediction Residual error - Sum of Squares) for each factor of the PLS. The PRESS value therefore represents the prediction error.
[0184] - The data for the sample that had been removed from the training set are then reintroduced into it and then the above steps are repeated removing the data for another sample.
[0185] The number of factors with a minimum PRESS value is then determined. This therefore makes it possible to obtain an optimal number of factors to use for the PLS, for each performance indicator to be determined.
[0186] The established models that present the best performances (parameters detailed below) are retained and their robustness is evaluated on the external validation set.
[0187] The different parameters that are used to choose the best models when developing with the training set are as follows:
[0188] - Multiple correlation coefficient R 2: it corresponds to the adequacy between the data predicted by the model and the values determined during the reference analyses. It must tend towards 1.
[0189] - SEC calibration standard error: this value is a statistical parameter, equivalent to one standard deviation and reflecting the upper limit of accuracy for future predictions. It should be as low as possible.
[0190] - Cross-validation standard error SECV: this value corresponds to the SEC during the cross-validation performed in the training set, during the PLS. Preferably, the SECV should have the closest possible value to the SEC. For the model to have an acceptable uncertainty, it was determined that the SECV must be less than or equal to twice the standard deviation associated with the values determined during the reference analyses. • Development of quantitative models: external validation
[0191] The data from the external validation set samples are representative of the range of values determined during the reference analyses. These data allow us to evaluate the robustness of the models developed with the training set that present the best performances and by studying the following parameters:
[0192] - Multiple correlation coefficient R 2 : it corresponds to the adequacy between the data predicted by the model for the external validation set and the values determined during the reference analyses. It must tend towards 1.
[0193] - Standard error of prediction SEP: this value is a statistical parameter reflecting the residuals (or prediction errors) obtained by the prediction of the external validation set. The ideal case is SEP < SECV. If this is not the case (SEP > SECV), care must be taken to ensure that the SEP is statistically less than or equal to the SECV (i.e. SECV < SEP < 1.3 x SECV).
[0194] - Bias: Bias is the average value of residuals (or prediction errors) calculated from the reference values of the components of the validation set. Bias should be close to 0, which indicates that the deviations are randomly distributed. On the other hand, a large bias value indicates a systematic error, such as changes in equipment, or in the reference analyses. This bias can then be corrected by the software.
[0195] - RPD: RPD stands for "Relative Percent Difference" and is a statistical indicator comparing the accuracy of the calibration to the variability of the reference data used. This value therefore makes it possible to evaluate the quality of the prediction made by the calibration. It is considered that an RPD greater than 1.4, preferably greater than 2, is associated with a good prediction and therefore the model is considered sufficient. Conversely, an RPD value less than 1.4 indicates that the model is not suitable.
[0196] The model with the best performance is selected and then determines the modeling parameters of the final calibration.
[0197] Example 2: Calibration Test Results - Live Insects
[0198] • Development of the model for the quantification of living insects
[0199] Model developments for determining the quantity of living insects at a particular stage of evolution in a given complex environment (different samplings) were established according to the method described in Example 1.
[0200] Thus, an adequate mathematical preprocessing was applied to each spectrum obtained.
[0201] It was determined that the appropriate mathematical preprocessing for the performance indicator related to insects at the larval stage in sampling A is an SNV-type normalization followed by a first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 13-factor model (Fig. 1, A and B respectively).
[0202] It was determined that the appropriate mathematical preprocessing for the performance indicator for insects at the larval stage, in sampling B, is a first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 12-factor model (Fig. 2, A and B respectively).
[0203] Furthermore, for this same B sampling, it was determined that the adequate mathematical preprocessing for the performance indicator relating to insects at the nymph stage is an SNV-type normalization, followed by a second derivative followed over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 9-factor model (Fig. 3, A and B respectively).
[0204] It was determined that the appropriate mathematical preprocessing for the performance indicator for adult insects in sampling C is a first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of an 11-factor model (Fig. 4, A and B respectively).
[0205] In this same C-sampling, it was determined that the appropriate mathematical preprocessing for the performance indicator for insects at the nymph stage is an SNV-type normalization, followed by a second derivative over the spectral range from 1100 nm to 1600 nm. The statistical parameters of the training set and the external validation set allow us to approve an 8-factor model (Fig. 5, A and B respectively).
[0206] Finally, it was determined that the appropriate mathematical preprocessing for the performance indicator for adult insects in sampling D is a first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of an 11-factor model (Fig. 6, A and B respectively).
[0207] A summary of the results is presented in Table 4 below.
[0208] Table 4]
[0209] It will be noted that in case B above, if the larvae constitute the population of living insects within the meaning of the present invention then the nymphs are a constituent of the complex medium as insects at at least one stage of development different from at least one insect in the population of living insects; or vice versa. Similarly, in case C above, if the nymphs constitute the population of living insects within the meaning of the present invention then the adults are a constituent of the complex medium as insects at at least one stage of development different from at least one insect in the population of living insects; or vice versa. These two populations, however, consist of living insects.
[0210] In conclusion, for each model obtained, the correlation coefficients R 2 of the training set on the one hand and of the external validation set on the other hand are very high.
[0211] The correlation between the calculated values and the values determined by the reference analyses, according to each model developed, is very strong.
[0212] For each model relating to a given population of living insects, the RPD values are always greater than 2 and are very high, thus demonstrating the good quality of prediction of the models on the ranges developed.
[0213] Measurement uncertainty represents the maximum error associated with predicting the mass quantity of a given living insect population. The values obtained are relatively low.
[0214] These results confirm that NIRS spectrometry analysis is an adequate method for determining a quantitative value of a population of living insects in a complex environment including droppings, substrate and / or dead insects.
[0215] Furthermore, the results indicate that the method according to the process of the invention makes it possible to distinguish different populations of living insects although the latter are mobile in the complex environment, and this without prior manipulation of the sample.
[0216] The results of sampling D demonstrate that the method according to the process of the invention can be used when the insects are reared on metal grids and therefore that the presence of a metal grid does not affect the results. Example 3: Results of calibration tests - Excrements
[0217] • Development of the model for the quantification of excrement
[0218] The development of models to determine the quantity of droppings was done according to the method described in Example 1.
[0219] Adequate mathematical preprocessing is applied to each spectrum obtained.
[0220] In particular, it was determined that the appropriate mathematical preprocessing for the performance indicator related to droppings in sampling A is an SNV-type normalization followed by a second derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 9-factor model (Fig. 7, A and B respectively).
[0221] It was determined that the appropriate mathematical preprocessing for the performance indicator related to droppings in sampling B is an SNV-type normalization followed by a first derivative over the spectral range from 1100 nm to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 14-factor model (Fig. 8, A and B respectively).
[0222] A summary of the results is presented in Table 5 below.
[0223] Table 5]
[0224] In conclusion, for each model obtained, the correlation coefficients R 2 of the training set on the one hand and of the external validation set on the other hand are very high. The correlation between the calculated values and the values determined by the reference analyses, according to each model developed, is very strong.
[0225] The RPD values are always greater than 2 and are very high, thus demonstrating the good prediction quality of the models on the developed ranges.
[0226] The measurement uncertainty represents the maximum error associated with the prediction of the mass quantity of excrement. The values obtained are relatively low.
[0227] These results confirm that NIRS spectrometry analysis is an adequate method for determining the quantitative value of insect droppings in a complex environment including living insects at different stages of development, nutrient substrate and / or dead insects.
[0228] Example 4: Calibration Test Results - Nutrient Substrate
[0229] • Development of the model for the quantification of a nutrient substrate
[0230] The development of models to determine the amount of nutrient substrate in a complex medium was done according to the method described in Example 1.
[0231] Adequate mathematical preprocessing is applied to each spectrum obtained.
[0232] It was determined that the appropriate mathematical preprocessing for the nutrient substrate performance indicator in sample A is SNV-type normalization, followed by a second derivative over the spectral range of 1100–1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 10-factor model (Fig. 9, A and B respectively).
[0233] It was determined that the appropriate mathematical preprocessing for the nutrient substrate performance indicator in sampling B is SNV-type normalization, followed by a first derivative over the spectral range of 1100–1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 10-factor model (Fig. 10, A and B respectively).
[0234] Regarding C-sampling, after research, it was determined that the appropriate mathematical preprocessing is a first derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow to approve a 10-factor model (Fig. 11, A and B respectively).
[0235] A summary of the results is presented in Table 6 below.
[0236] [Table 6]
[0237] For each model obtained, the correlation coefficients R 2 of the training set on the one hand and of the external validation set on the other hand are very high.
[0238] For each model relating to the nutrient substrate, the RPD values are always greater than 2 and are relatively high, thus demonstrating the good quality of prediction of the models on the ranges developed.
[0239] Measurement uncertainty represents the maximum error associated with predicting the amount of nutrient substrate. The values obtained are relatively low.
[0240] These results confirm that NIRS spectrometry analysis is a suitable method to reliably determine the quantitative value of a nutrient substrate in different complex environments including live insects at different stages of development, droppings and / or dead insects.
[0241] Example 5: Calibration Test Results - Dead Insects
[0242] • Development of the model for the quantification of dead insects
[0243] The development of models to determine the mass quantity of dead insects in a complex environment was done according to the method described in Example 1.
[0244] Adequate mathematical preprocessing is applied to each spectrum obtained.
[0245] It was determined that the appropriate mathematical preprocessing for the spectra from sampling A is an SNV-type normalization, followed by a first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow us to approve an 11-factor model (Fig. 12, A and B respectively).
[0246] It was determined that the appropriate mathematical preprocessing for the spectra from B sampling is an SNV-type normalization, followed by a second derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 13-factor model (Fig. 13, A and B respectively).
[0247] It was determined that the appropriate mathematical preprocessing of the spectra acquired on C sampling is a second derivative over the spectral range of 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow to approve a 6-factor model (Fig. 14, A and B respectively).
[0248] Finally, it was determined that the appropriate mathematical preprocessing of the spectra acquired on D sampling is an SNV-type normalization, followed by a first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 10-factor model (Fig. 15, A and B respectively).
[0249] A summary of the results is presented in Table 7 below.
[0250] Table 7]
[0251] In conclusion, for each model obtained, the correlation coefficients R 2of the training set on the one hand and of the external validation set on the other hand are very high.
[0252] For each model relating to dead insects, the RPD values are always greater than 2 and are relatively high, thus demonstrating the good prediction quality of the models on the ranges developed.
[0253] Measurement uncertainty represents the maximum error associated with predicting the quantity of dead insects. The values obtained are relatively low.
[0254] These results confirm that NIRS spectrometry analysis is an adequate method for reliably determining the quantitative value of dead insects in a complex environment including live insects at different stages of development, nutrient substrate and / or droppings.
[0255] Furthermore, these results demonstrate that NIRS spectrometry analysis is an adequate method for distinguishing dead insects from live insects in a complex environment. Example 6: Calibration test results - Contaminants
[0256] • Development of the model for the quantification of contaminants
[0257] The development of models to determine the amount of contaminants in a complex medium was done according to the method described in Example 1.
[0258] It was determined that the appropriate mathematical preprocessing of the acquired spectra from E sampling is a Savinsky-type processing, followed by a first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow the approval of a 12-factor model (Fig. 16, A and B respectively).
[0259] A summary of the results is presented in Table 8 below.
[0260] Table 8]
[0261] In conclusion, the correlation coefficients R 2 of the training set on the one hand and of the external validation set on the other hand are very high.
[0262] The correlation between the calculated values and the values determined by the reference analyses is very strong.
[0263] In the contaminant model, the RPD value is greater than 2, which demonstrates the good prediction quality of the model on the developed ranges.
[0264] These results confirm that NIRS spectrometry analysis is a suitable method for reliably determining the quantitative value of contaminants in a complex environment including live insects, nutrient substrate and excrement.
[0265] Example 7: Calibration Test Results - IMM
[0266] • Development of the model for the determination of the average individual mass in the range from 1.03 mq to 42.86 mq
[0267] The development of models to determine the average individual mass of insects in a complex environment was done according to the method described in Example 1.
[0268] In particular, it was determined that the adequate mathematical preprocessing for the IMM indicator in the E sampling in the range of 1.03 mg to 42.86 mg is a first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow to approve a 14-factor model (Fig. 17, A and B respectively).
[0269] A summary of the results is presented in Table 9 below.
[0270] Table 9]
[0271] In conclusion, the correlation coefficients R 2of the training set on the one hand and of the external validation set on the other hand are very high.
[0272] Indeed, the correlation between the calculated values and the values determined by the reference analyses is very strong.
[0273] In the model for determining the average individual mass of insects, the RPD value is greater than 2, which demonstrates the good prediction quality of the model over the range of 1.03 mg to 42.86 mg.
[0274] • Development of the model for the determination of the average individual mass in the range of 40.85 mg to 173.46 mg
[0275] The development of models to determine the average individual mass of insects in a complex environment was done according to the method described in Example 1.
[0276] It was determined that the appropriate mathematical preprocessing for the IMM indicator in E sampling for model development in the range of 40.85 mg to 173.46 mg is SNV-type normalization followed by a first derivative over the spectral range from 1100 to 1600 nm. The statistical parameters of the training set and the external validation set allow for the approval of a 15-factor model (Fig. 18, A and B respectively).
[0277] A summary of the results is presented in Table 10 below.
[0278] Table 10]
[0279] In conclusion, the correlation coefficients R 2 of the training set on the one hand and of the external validation set on the other hand are very high. Indeed, the correlation between the calculated values and the values determined by the reference analyses is very strong.
[0280] In the model for determining the average individual mass of insects, the RPD value is greater than 2, which demonstrates the good prediction quality of the model over the range from 40.85 mg to 173.46 mg.
[0281] These results confirm that NIRS spectrometry analysis is a suitable method for determining the average individual mass of insects in a complex environment including living insects at different stages of development, nutrient substrate and / or dead insects, in a non-invasive manner.
Claims
DEMANDS 1. A non-destructive method for the quantitative determination of at least one performance indicator of a rearing of a population of live insects in a complex environment, said method comprising the following steps: (i) an irradiation step of the complex medium comprising the population of live insects with light comprising one or more wavelengths in a range from 420 nm to 2500 nm, (ii) a step of collecting the light reflected by the complex medium in order to obtain a reflection spectrum of the medium, and (iii) a step of correlating said spectrum to a quantitative value of at least one performance indicator.
2. A method according to claim 1, wherein the complex medium comprises at least two different constituents other than the population of live insects.
3. A method according to claim 2, wherein the at least two different constituents are chosen from a substrate and excrement.
4. A method according to any one of claims 1 to 3, wherein the complex medium comprises at least one insect at at least one developmental stage different from at least one insect in the live insect population.
5. A method according to any one of claims 1 to 4, wherein the complex medium comprises at least one dead insect.
6. A method according to any one of claims 1 to 5, wherein the performance indicator is chosen from a quantification of an insect population, a substrate, excrement, a contaminant, and / or an average individual mass of insects.
7. A method according to claim 6, wherein the performance indicator is a quantification of excrement.
8. A method according to claim 6, wherein the performance indicator is a quantification of an insect population.
9. A method according to claim 8, wherein the insect population consists of dead insects.
10. A method according to claim 8, wherein the insect population consists of insects living at a given stage of development.
11. A method according to any one of claims 1 to 10, further comprising, after the collection step (ii) and before the correlation step (iii), a mathematical processing step of the reflection spectrum comprising a first derivative, a second derivative, a Savinsky-type treatment and / or an SNV-type normalization.
12. Use of the method as defined in any one of claims 1 to 11, to carry out monitoring of the growth of the live insect population, to determine the efficiency of a sorting machine and / or to determine the efficiency of feeding equipment for the live insect population.
13. Use of a spectrophotometer for the quantitative determination of at least one performance indicator of a population of insects living in a complex environment.
14. Use according to claim 13, wherein the performance indicator is selected from a quantification of an insect population, a substrate, excrement, a contaminant, and / or an average individual mass of insects.
