This patent application claims priority benefit of United States Provisional patent application number 62/952,597 filed on December 23, 2019, and United States Provisional patent application number 63/198,219, filed on October 2, 2020.

FIELD OF THE INVENTION

[0001] This invention relates to devices for farming insects, and more particularly to a device for enhancing the yield of insect farms.

BACKGROUND OF THE INVENTION

[0002] According to the United Nations, it is widely accepted that by 2050 the world will host 9 billion people. To accommodate this rise in population, food production levels will need to increase substantially. Arable land is limited and expanding the area devoted to farming is rarely a viable or sustainable option. Insects offer a solution for providing a cost effective and sustainable source of nutrition. Until recently, the majority of edible insects have been gathered from forest habitats. Systems for mass-rearing of insects have only recently emerged. See, for example, US Patent 10,405,528 to Comparat et al and US Patent 10,264,769 to Leo, which disclose various processes relating to insect rearing. Although such known devices solve certain problems in rearing insects, much can still be done to enhance yield and improve the overall insect rearing process. Problems of scale are especially apparent in that known designs are most typically either highly labor intensive (and therefore costly), or broadly (and therefore inefficiently) applicable to a wide range of insects rather than a refined process for the production of a particular species or type of insect.

[0003] It would therefore be desirable to provide a device for the intelligent organization of insects, in particular, intelligently organizing and sorting the breeding adult populations which allows for an efficient high-volume operation.

SUMMARY OF THE INVENTION [0004] In accordance with a first aspect, an insect farm for farming insects, wherein the insects have an egg-laying stage, a larva stage, a pupa stage, and then become adults, comprises a first larva tray containing larvae wherein the larvae have previously been separated from parents of the larvae, wherein the larvae grow in the first larva tray for a first period of time during the larva stage, and at least one additional larvae tray, wherein a portion of the larva are transferred from the first larvae tray to the at least one additional larvae tray, and the larvae grow in the first larvae tray and the second larvae tray for a second period of time during the larva stage and after the first period of time. The insect farm may further comprise a modular adult container containing adult breeding insects, and a removable tray containing eggs laid by the adults positioned in the modular adult container, wherein the adults are present in the modular adult container for a breeding time based upon the age of the adults. The insect farm may still further comprise a pupae container containing insect pupae, an inspection device for inspecting pupae and generating a signal, wherein inspecting pupae is based on at least one of: a sex of the pupa, a color of the pupa, a size of the pupa, a length of the pupa, a uniformity of coloring of the pupa, and deformities of the pupa. A sorting device is operatively connected to the inspection device, and the sorting device sorts the next generation pupae based on a signal from the inspection device.

[0005] From the foregoing disclosure and the following more detailed description of various embodiments it will be apparent to those skilled in the art that the present invention provides a significant advance in the technology of insect breeding programs. Particularly significant in this regard is the potential the invention affords for providing a low cost, economically viable device for high volume continuous production of an enhanced, high quality breeding population. Additional elements and advantages of various embodiments will be better understood in view of the detailed description provided below.

BRIEF DESCRIPTION OF THE DRAWINGS [0006] Fig. 1 shows an adult modular container of an insect farm in accordance with one embodiment comprising a series of stacked modular adult containers, with each container having a removable tray defining an egg laying section.

[0007] Fig. 2 is an isolated isometric view of one of the modular adult containers and the removable tray with an egg-laying section in accordance with one embodiment.

[0008] Fig. 3 is a graph showing a representative graph of insect egg yield vs. time, showing a nonlinear nature of egg laying.

[0009] Fig. 4 shows a larvae stage portion of the insect farm in accordance with one embodiment, showing a series of stacked modular larvae containers.

[0010] Fig. 5 shows a schematic example of splitting the larvae from one modular larvae container to a pair of modular larvae containers.

[0011] Fig. 6 is a representative larvae growth curve in accordance with one embodiment.

[0012] Fig. 7 is a representative insect egg harvest schedule, showing how harvesting the eggs at different times affects a total yield.

[0013] Fig. 8 shows a schematic of one embodiment of a pupae sorter for selecting next generation adults.

[0014] It should be understood that the appended drawings are not necessarily to scale, presenting a somewhat simplified representation of various features illustrative of the basic principles of the invention. The specific design features of the insect farm as disclosed here, including, for example, the specific dimensions of the modular containers and trays will be determined in part by the particular intended application and use environment. Certain features of the illustrated embodiments have been enlarged or distorted relative to others to help provide clear understanding. In particular, thin features may be thickened, for example, for clarity of illustration. All references to direction and position, unless otherwise indicated, refer to the orientation illustrated in the drawings.

DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS [0015] It will be apparent to those skilled in the art, that is, to those who have knowledge or experience in this area of technology, that many uses and design variations are possible for the insect farm disclosed here. The following detailed discussion of various alternate elements and embodiments will illustrate the general principles of the invention with reference to an insect farm suitable specifically for beetles (order Coleoptera), and for high volume operations. Other embodiments suitable for other applications, such as other beetles and other insects, will be apparent to those skilled in the art given the benefit of this disclosure.

[0016] Turning now to the drawings, Fig. 1 shows an insect farm 10 in accordance with one embodiment raising beetles. Beetles (and insects generally) have several stages of life, egg, then larva, then pupa, and then adult. For the representative beetles used herein as an example insect, there are about 7 days from egg to larvae, about 77 days in the larval stage, and another 7 days in the pupae stage before becoming an adult. Each of these stages is enhanced in the insect farm disclosed herein.

[0017] Fig. 1 shows an egg laying stage 20, shown to comprise a series of modular adult containers 21 holding egg laying adult insects. The modular adult containers 21 may be formed as a stack 23 as shown, and may be mounted on a frame 24 with wheels 25 and a hitch 26 for ease of movement from one part of the farm to another.

The stacks 23 may be configured to include sensors and transmission/receiver capabilities via WIFI, Bluetooth, etc. This helps to ensure that environmental data is tracked for each modular adult container and/or for each stack, and a data point can be created and sent to a controller or processor for monitoring. Tags like NFC or RFID may be used to ensure monitoring can be done to enhance traceability. The tags may be scanned and logged or updated to track the progress of growth and egg laying of the insects the farm, and extrapolate meaningful data like larva production estimates, or to initiate actions, such as moving a modular adult container to a different location.

[0018] The modular adult containers have a floor 27 and side walls 22, and the side walls form side wall openings 28 which can allow air to reach the beetles. Preferably the side walls are of a relatively low stick surface such that the insect in the containers 21 cannot climb out. For insects which are more mobile the side openings may be screened off. Feed and water lines may be provided, although for some types of insects, a simple food mix introduced to the floor 27 may be sufficient for them to live. A removable tray 30 is shown on the top container 21. Preferably each container 21 has a similar removable tray 30. The removable tray can have a floor 31 and one or more sloped side walls 32 to allow the insects to enter the removable tray.

[0019] Important metrics. The particular species of beetle can be Tenebrio molitor. It will be readily apparent to those skilled in the art, given the benefit of this disclosure, that the insect farm disclosed herein can be broadly applicable to the order Coleoptera and any insect species in which pupation is a stage of metamorphosis, especially including Flermetia illucens or black soldier flies and related species. For a feed Conversion Rate (FCR), a ratio of 2.5 is acceptable but a range of 1.8-3.8 can be used, most preferably 2.5 or higher. Relative humidity of the surrounding environment can be 60-90%. Operating Temperature 25-38°C, preferably about 28-31 °C to achieve maximum growth rates, however their own metabolic energy and insulation due to high density needs to be taken into consideration. Also, higher temperatures induce higher levels of protein in species like Tenebrio molitor, but reduce growth rates to compensate. Density of insect adults, larvae, etc. Tenebrio molitor is not nearly as negatively impacted by density as others, and higher densities may induce higher growth rates. For our calculations we assume 10 larvae per square cm, and it is believed that high production rates are maintained even at higher densities than this. The beetle Tenebrio molitor is photosensitive, with light actually “hurting” them. Preferably the insect farm should be maintained with low light levels. For adults this effect is much less dramatic, and some light is necessary for their mating. For feed, in “cold” blooded animals, protein is the largest factor in growth rates. In “warm” blooded animals, it is energy that impacts growth rate the most. Therefore, insects fed higher quality and quantity protein will grow faster. Preferably feed is used with 18% protein. The range may be from 10 to 50%, preferably around 15-30% based on a cost-benefit analysis. Tenebrio molitor beetles can absorb most of the water they need from the air. However, animal density and other factors will delay the growth rate if water is expected to be entirely derived from the air (as this requires metabolic work for the insects). Therefore, liquid water can occasionally be provided. In some situations, and for some breeds of insect, no separate water line may be required, with the insects obtaining all the water they need from the food they are supplied. Oxygen, air particulate matter, and pressure may also be important metrics to control in the harvester. Other metrics useful to control in the insect egg harvester will be readily apparent to those skilled in the art given the benefit of this disclosure.

[0020] Fig. 2 is an isolated isometric view of one of the modular adult containers 21 of Fig. 1. The removable tray 30 has sloped side walls 32 which makes it relatively easy for the beetles to climb into the tray 30. The side walls can extend from the floor 27 of the modular adult container 21 to allow the adult beetles to walk up into the removable tray. The side walls of the removable tray may also extend to the floor of the removable tray to allow the beetles to walk down to the floor of the removable tray. In accordance with a highly advantageous element, the removable tray 30 is provided with a substrate 33 applied to a floor 31 of the tray 30 which the beetles are attracted to lay their eggs 34. Preferably there is no substrate 33 outside of the removable tray 30 (although there may be food for adults present on the floor 27 of the modular adult container 20). Thus, the removable tray 30 is an egg laying tray where beetles preferentially lay their eggs. The eggs preferably stick to the substrate to help separation of the egg from the egg laying adult. The eggs can be regularly separated from the adults, optionally through a sieving mechanism using a first sieve which separates larger sized adults from smaller larva. To further enhance separation, the diameter of the substrate 33 may be controlled so the substrate has a diameter smaller than a width of the eggs 34. In this way a second sieve can separate the larva from the substrate, frass and exuvium of the insects. Also, the removable tray 30 may be provided with a low friction surface to further help with separation of eggs 34 from the removable tray 30.

[0021] The exuvium or sheds of the insects deposited on the modular adult containers 20 and/or the removable trays 30 may be collected via vacuum or pressurized air. This step can be done cost effectively, and helps to ensure that the exuvium does not affect the final output product, and can itself be useful as a by-product. More specifically, the contents such as the insects (typically in the late larval stage of development), their frass and any remaining feed may be deposited into a filter device which separates the insects from their frass and uneaten food particles. Removing the frass (excrement) is a critical step in keeping sanitary conditions. The frass is a by-product, which can either be used directly as a product for fertilizer or even feed, or may be further refined to extract fat or protein or other nutrients that may be desirable. The frass collection system can be part of the sieving and be configured to reduce damage to the larvae, and even keep insects or larvae separate on a by-bin/unit basis, such that traceability of the bin may be maintained all the way through processing (in other words, the contents of the bins are not mixed during sorting/sifting). The frass can also be further refined through oil or screw press, or through use of chemical solvents and centrifuge and/or subsequent dehydration, depending on the desired output product.

[0022] Because beetles have an age dependent fecundity, it is necessary to program a separation step (including the sieving interval, where present) to increase as the birth rate decreases. As is shown in Fig. 3, the adult beetles experience a high rate of reproduction in the first couple weeks of maturity as an adult. Gradually, their reproductive rate decreases as their age increases. Therefore, the sieving/egg separation interval must increase as the age of the breeding adults increases. Other factors may also affect the birth rate of the adult beetles. These factors may include; temperature, humidity, light, feed, water, weight changes, sound. The invention may also include sensors to monitor these conditions to enhance egg laying efficiency.

Thus, the adults are present in the modular adult container 30 for a breeding time based upon the age of the adults. The breeding time is based upon an estimated amount of eggs the adults have produced and preferably is less than a period of time when the insects are fertile. In accordance with one embodiment designed to enhance yield, breeding time is limited to when an estimated amount of eggs to be laid is less than 85% or less than 80% or less than 75% of a total amount of eggs adults are expected to produce. Known insect farms may let their eggs accumulate and only segregate them once per week. However, as shown in Fig. 7, if the eggs are separated more frequently, preferably daily, yield can be enhanced as all of the eggs (and subsequent larvae) will be of the same age and same size at harvest. This is particularly advantageous for insects, as they typically have non-linear growth rates. This means that beetle larva just a few days different in ages can have noticeably different mass at harvest.

[0023] Fig. 4 shows a larvae growth stage 40 where the larvae 44 are transferred to after separation from where they are laid as eggs. A series of modular larvae containers or larvae trays 41 are provided, preferably stackable in a stack 49 as shown. Stackable trays help reduce floor space, which helps reduce overhead. Each tray contains larvae 44 and feed 43 to allow the larva to grow. Each larva tray 41 may have a floor 47, side walls 42 and openings 48 which allow air flow. The stack 49 may be provided with a pallet jack 54 on which the stack is positioned. The pallet jack can have a connecting element 46 and wheels 45 which allow for connection to a device for movement of the stack to and from a preferred location. Other connection assemblies to help which movement of a stack will be readily apparent to those skilled in the art given the benefit of this disclosure. Optionally the tray 41 can be the same as the modular adult container 21.

[0024] Known insect farms use a tray large enough to account for growth of the larvae during the larval stage, which can be considerable. This is done because they do not want the insects to overgrow and die off. Thus, space is wasted in the early part of the larvae growth stage to account for final growth. In accordance with a highly advantageous feature, Fig. 5 shows a space-saving embodiment where a first larva tray 41 contains larva previously separated from parents of the larva, and deposited in a relatively dense grouping. The larvae grow in the first larva tray for a first period of time during the larva stage. If left in that single tray the larvae would overgrow their tray and that can inhibit growth. Advantageously at least one additional larvae tray is provided, wherein a portion of the larva are transferred from the first larvae tray to the at least one additional larvae tray. As shown in Fig. 5 there are the larvae are separated with about half in each of two larvae trays. The larvae grow in the second larvae tray for a second period of time during the larva stage and after the first period of time. The first period of time and the second period of time are determined by a rate of projected growth of the insects in the larva stage. Fig. 6 shows representative larvae growth rates for beetles. The beetles do most of their growing in the final 7-10 days of their development period, growing about 50% of their final mass in that home stretch. This can be advantageously exploited by the dynamic density rotation method disclosed herein. Other insects show similar nonlinear growth rates, again with a first period of time to get to 40-60% of projected larva growth being substantially greater than the second period of time to get to about 100%. The larvae can be transferred to an additional larvae tray after 40-60%, most preferably 50% of the projected growth. Organizing the larva in this manner advantageously increases the efficiency of a high-volume operation. The removable trays 30 that are populated by eggs and recently hatched larva may have a tray density of 10 larva/cm ² , for example. The larva trays 41 may have a tray density of 40 larva/cm ² , for example, and can be halved when half of the larva are moved to a second larva tray. By varying the insect density in rearing trays, according to the size or condition of insects in various stages of development, output per tray may be significantly improved.

[0025] The beetles, when they are full grown, are approximately 150 mg, and have a preferred density of about 5 larva/cm ² . This results in a biomass density of about 750 mg/cm ² (150 mg x5). Using this biomass density figure on the larva when they are small, and not yet full size, weighing 25 mg or less, we could fit 30 larva/cm ² (750/25).

In accordance with one embodiment, and as an estimate of the value of the techniques disclosed herein, the larvae may be kept at 30 larva/cm ² for the first 7 weeks, and subdivided into a pair of trays 41 at the 7-week mark to be at 5 larva/cm ² in the final 4 weeks in advanced preparation for the adult mass. This gives an efficiency of 210 larva- weeks/cm ² + 20 larva-weeks/cm ² = 230 larva-weeks/cm ² . By comparison, it is believed that known farms may keep their density static the entire 11 weeks, for example 5 larva/cm ² , which is an efficiency of just 55 larva-weeks/cm ² . Therefore, the technique disclosed herein provides an approximately 400% efficiency increase over known technology.

[0026] Typically a breeding program in high volume insect rearing systems is done manually, which is inefficient for several reasons; the adults are often chosen by convenience, and are genetically identical to the average quality of the overall population. This leads to a stagnation in the population, and may force insect producers to manually undergo breeding programs based on inaccurate visual determinations. In high volume operations, this leads to significant loss in additional labor and feed costs. The pupae sorting stage can be a sorting mechanism that collects insect pupae in a hopper, distributes them with a conveyor system that orients them into a preferred position for the image capturing system which captures data that is intelligently analysed using machine learning to determine the sex and genetic quality of each individual pupa. Among other advantages, this innovation allows for an ideal sex ratio to be achieved in breeding bins for mass scale commercial insect farms. A preferred breeding ratio is about 1:1, which can yield up to a 50% increase in efficiency and cost reduction for commercial insect farms. Optionally the sex ratio may be modified as needed to help increase yield.

[0027] Fig. 8 shows an example of a device for intelligently sorting insect pupae. This pupae sorter stage 60 works well as pupae move very little and therefore can be readily inspected. The next generation of insect breeders can be selected based on one or more of several useful inspection parameters. These parameters include a sex of the pupa (preferably an even sex ratio - 50%M-50%F), a color of the pupa, a size of the pupa, a length of the pupa, a uniformity of coloring of the pupa, and deformities of the pupa (where a lack of deformities would make it more likely the pupa would be selected for next generation breeding). Additional traits of the pupa suitable for use as an inspection parameter may comprise, for example, a body shape of the pupae, evidence of disease, or proportion of or coloration of various body parts. Other physiological traits of the pupa, and/or a phenotypical trait of the pupa suitable for use as an inspection parameter will be readily apparent to those skilled in the art given the benefit of this disclosure. As an example, the inspection device can be an optical inspection device which optically inspects each pupa to determine its sex, size, color and if any deformities are present. When the inspected pupa passes a threshold for each of these inspection parameters, the inspection device sends a signal to the separator/sorting device which urges the inspected pupa into the next generation group area. Selected pupa (which will eventually spawn as adults) are then cycled back to the modular adult container 30, where they (the adults) will begin a career of breeding the subsequent generations. In a facility breeding insects of the order Coleoptera for example (depending on the egg rate), no more than 3%, preferably around 1%, of the total population should be necessary to maintain the total population. In the case of population expansion, then the desired population count will determine how many adults are necessary.

[0028] The pupae sorter stage 60 is shown to comprise a pupae 61 container containing insect pupae 64, a vibration device 62 to bias the pupae (which do not move otherwise) towards a conveyor 65 and inspection device 66. The pupae may be oriented for inspection, and placed in more or less single file. The inspection device 66 may optically inspect the pupae and generate a signal based on one or more of the preferred parameters mentioned above. A sorting device 67 can be operatively connected to the inspection device, wherein the sorting device sorts the next generation pupae based on the signal from the inspection device into a next generation group 68 and a rejection group 69. The sorting device can be a mechanical shunt which routes the pupae to one or the other collection areas.

[0029] Microcontrollers and optical sensors advantageously quickly and harmlessly can be used determine the sex of darkling beetle pupae. Determining the sex of the beetles at the pupae stage is important to provide a control for enhanced egg output in the rearing of insects. Optionally machine learning (ML) algorithms may also be used. For example the microcontroller may have a processor which is fed a dataset from the optical sensor input in order to rapidly identify desirable traits in mass insect rearing breeding programs. This is valuable for high volume insect production systems, as the breeder population may have a birthrate of 100X or more, or at least 200 eggs per female in a breeding period, this means that the breeder population may only represent 1 %-5% of the overall population, and thus the quality of genetics in adult breeding insects is highly influential on the overall colony genetic quality for high volume operations.

[0030] Currently, the preferred microcontrollers contain nRF52840 (Nordic Semiconductors) processors loaded with algorithms to process the live optical input. For prototyping purposes Arduino Nano 33 BLE Sense microcontrollers can be used, which contain the nRF52840 processor. Connected to the microcontrollers are OV7670 CMOS VGA Arduino Camera Modules (one camera module per microcontroller). The lenses on the camera of inspection device 65 can provide a focal range of 10mm - 25mm.

Lighting may be part of the inspection device. For example, fixed around the camera module lenses can be Adafruit NeoPixel Rings - 12 x 5050 RGB (Red-Green-Blue) LED with integrated drivers to control various light spectrums to provide enhanced sex organ recognition for the optical sensors. To control the RGB color output of the NeoPixel Rings an Arduino MEGA 2560 may be used. RGB color output value of R:255, G:255, B:0 (yellow spectrum) seems to provide the highest level of color contrast around the sex organs of the pupae, creating the largest visual delineation between the male and female sex organs. The first of several machine learning algorithms process an input dataset (of signals corresponding to pupae) against a set of features, weights or other reference standards to determine the probability of the image containing the abdominal apex ventral view, which is required to determine gender. If the input passes this first algorithm, it is then given as input to a second algorithm to determine female probability. The female probability is determined by parsing the input image as a dataset. Each parsed dataset’s image pixels are compared to the predefined algorithm features. The features for the female probability algorithm are defined by key ranges of pixel color derivatives over key linear pixel ranges. These ranges represent clusters of pixels that have been identified as gender specific clusters, that is, the data is compared to a reference image. The gender specific cluster identifications are a product of the machine vision training process. The training process is fed with images of pupae that are pre-determ ined to be female via stereo microscope inspection.

[0031] Female pupae have sex organs (ovipositor) which yield a subtle set of physical protrusions. Therefore, the model is trained to find the occurrences of those features, or pixel clusters that resemble the features. The probability of the female sex determination is in a direct relationship with the volume of positive female sex organ feature occurrences. The machine learning algorithm weighs each input dataset probability determination against a probability threshold to determine if the input dataset resembles that of a female pupa enough to pass as a female pupa; if this standard is not met the image dataset is then passed to a tertiary machine learning algorithm which is trained with images of male pupae sex organs as a reference. The tertiary machine learning algorithm features are trained on reference images in a similar manner as the secondary (female) algorithm features, although with different pixel color derivatives over different linear pixel ranges (different clusters), and also with datasets containing images of pupae sex organs that have been predetermined to be male via stereo microscope inspection. Like the female probability determinations, the male probability is also weighed against a probability threshold to determine if the features in the dataset resemble the features of a male sex organ enough to pass as a male pupa. If the machine learning algorithm trained with male pupae sex organ features does not pass the dataset as male, the dataset is given a default clarification of “Undetermined”. Thus, each machine algorithm contains a final probability threshold that can be manipulated with programming variables. These probability thresholds are the control element for manipulating the accuracy of the final determinations. Final determination classifications can include male, female, or undetermined. By including additional sensors, such as weight, color, or sizing determination sensors, etc., additional layers of classification may be added to the sorting intelligence, such that next generation pupae with excellent characteristics in more than one parameter may be achieved. Parameters inspected by the inspection device may be concurrent or sequential, and the parameters selected may be subjective traits or objective traits.

[0032] To further the accuracy of the probability determinations of the sex of a given pupae, a large dataset can be provided for the machine learning model to be trained. The datasets are acquired by capturing an escalating range of images of pupa sex organs. Each image in each dataset can be of a unique pupa to further diversify the training datasets. The machine learning features are trained by feeding these datasets to the training model by using libraries such as TensorFlow. To leverage faster training results, large cloud based virtual machines can be used for temporarily enhanced computing power.

[0033] The value of inspection and sorting for next generation adults is enhanced in high volume rearing operations of insects. As noted above, the automated pupae sorter stage may include a conveyor and physical sorting capabilities. The pupae can be fed into a hopper which gently transports them onto a conveyor system which then orients each pupa and passes it by an intelligent (machine learning coupled) optical sensor of an inspection device. Once the sex of the pupa is quickly and harmlessly determined, the sorting device is activated by a signal from the microcontroller to the sorting device to physically sort the pupae into bins based on the above referenced parameters. The pupae may also be sorted merely by their determined sex. These bins can then be further automated onto conveyor systems that populate new adult breeder groups. Currently, the inspection device may comprise a single 3-dimensional angle of view in order to train the sex-determining machine vision neural networks, combined with a preliminary machine vision neural network to determine the presence of the abdominal apex ventral view of the pupa in the image. The optical sensor feed sends the images to the first machine learning neural network (ML1 ) to determine if the image is of the abdominal apex ventral view, which is required for the subsequent machine learning algorithms to accurately determine the sex of the pupa. Once an image from the optical sensor feed is approved by ML1 , the same image is then sent to another machine learning algorithm (ML2) to determine if the sex organs are that of a female pupa. If ML2 does not flag the image as a female pupa, then the image is sent to the last machine learning algorithm (ML3) to determine if the sex organs are that of a male pupa.

[0034] To ensure that each pupa gets an accurate determination from the machine learning algorithms, an image of the correct angle and view must be supplied to the algorithms. In efforts to quickly and reliably capture this image for every pupa passed from the conveyor system, optionally a series of 9 optical sensors can be positioned around the pupa in a 180-degree range, creating a semi-circle of optical sensors around the pupa. Each sensor runs an input feed of images to ML1. The first image that is approved by ML1 can then be used for the remaining machine learning algorithms, and the input from the remaining optical sensors can be ignored for a period of time. By breaking up the machine learning into more modular algorithms that take less time and resources to train, a single-angle image can provide all of the input necessary to get an accurate male to female probability determination. Further, the additional, preliminary machine learning neural network reduces the minimum threshold of training images required to train a machine learning neural network for achieving male to female probability accuracy of at least 80%, alleviating required training resources and enhancing technical and commercial feasibility.

[0035] The microcontroller may be pre-loaded with software that dictates the signals the microcontroller emits. The microcontroller containing the nRF52840 processor runs on 3.3 Volts, which is good for its purpose of processing ML algorithms to make a decision to be incorporated into the signal. Once the microcontroller comes to a decision as a result of the machine vision neural network output, software dictates whether or not a mechanical action is necessary. If a mechanical action is necessary, the microcontroller emits the signal to a larger voltage (5 Volts) controller that is programmed to complete the loop of an electrical circuit containing a motor for a period of time. Once this circuit’s loop providing voltage to the electrical motor is completed, the motor activates and provides the necessary kinetic energy to actuate the sorting device. After a time interval equal to the product of the velocity of the surface of the conveyor system and constant width between each pupae has passed, the 5 Volt controller opens the electrical circuit that provides voltage to the motor, effectively terminating the motor’s locomotion. A separate retention mechanism (such as a spring or other biasing or return device) provides a constant retention pressure to the sorting mechanism in the opposite direction by which the motor is locomoting. The newton-meters of energy per square millimeter provided by the electrical motor can be twice that of the retention energy applied to the sorting mechanism. The trajectory the electrical motor provides to the sorting device can be opposite of that which the retention mechanism provides. This single directional locomotion opposed with a single directional retention mechanism provides a Boolean solution to mechanically sort pupae based on the output of machine vision neural networks.

[0036] A motor activates to move a sorting “fin” that allows pupae of a determined sex to take a unique path down the conveyor system; the sorting “fin” is then reset back to its original position due to the motor losing voltage combined with the presence of a mechanical spring that provides constant retention on the sorting “fin”. This entire process, from the pupae tray/hopper to the sorting device can be analyzed physically to reduce the impulse, friction, and forces acting on the pupae. In all methods, steps are taken to ensure little harm is incurred on the pupa as they pass through the various stages, preferably a mortality rate of less than 5%. Advantages of the pupae stage sorter comprise, for example, a variable sorting speed, to protect pupa physical integrity (not too rapid or fast). Automated catchment, as opposed to manual sorting of pupae at random to form adult breeding groups. A key advantage is ensuring that male to female (MF) ratios are kept in line with Fisherian principles, preferably 1:1 (an equal ratio). This ensures that maximum birthrates are achieved, as too many males in a population will lead to stress on males from over competition, and stress on females from overbreeding; while too few males in a population will result in stress on males from overbreeding and decreased egg rates from females from under breeding. Keeping the MF ratios preferentially controlled results in significant feed and labor savings. Additionally, the sensors may include the capability to sort pupae according to numerous desirable traits. This ensures that the overall colony is genetically controlled, as the most superior genetic pupae may be selected for the breeding program.

[0037] From the foregoing disclosure and detailed description of certain embodiments, it will be apparent that various modifications, additions and other alternative embodiments are possible without departing from the true scope of the invention. The embodiments discussed were chosen and described to provide the best illustration of the principles of the invention and its practical application to thereby enable one of ordinary skill in the art to use the invention in various embodiments and with various modifications as are suited to the particular use contemplated. All such modifications and variations are within the scope of the invention as determined by the appended claims when interpreted in accordance with the breadth to which they are fairly, legally, and equitably entitled.