Systems and methods for automatic and semi-automatic mosquito isolation, identification, counting and pooling

Abstract

A method of separating a batch of insects from a trap into individual insects, comprising pouring the batch into a container having at least one hole, the hole being sized to fit an individual insect; and moving the container to shake the insects therein so that the individual insects exit through the hole onto a collection surface, thereby providing the separated insects onto the collection surface. The insects from the trap can then be counted, and image recognition can be used to identify the genus, species or sex. The process can be performed for multiple traps, and the location data can be stored along with the insect identification to give a map of the insect distribution.

Description

[0001] Related Applications

[0002] This application claims the benefit of priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No. 62 / 935,414 filed on November 14, 2019, U.S. Provisional Patent Application No. 63 / 007,064 filed on April 8, 2020, and U.S. Provisional Patent Application No. 62 / 988,427 filed on March 12, 2020, the contents of which are incorporated herein by reference in their entirety.

[0003] Technical Field and Background Art

[0004] The present invention relates to insect monitoring and, more particularly, but not exclusively, to systems and methods for monitoring insect populations which may include isolating, identifying, counting and aggregating found insects.

[0005] There are many situations where it is necessary or desirable to monitor insect populations. Such situations range from surveying natural ecosystems and monitoring changes to public health situations where insects such as mosquitoes, black flies, midges and sand flies pose a public health hazard to humans or livestock.

[0006] Knowing how many mosquitoes are present in a given area is an important factor for authorities to make decisions about control, including spraying chemicals, quantities, locations, etc.

[0007] Today, measurements are made by spreading mosquito traps over an area and counting the insects caught.

[0008] When a mosquito trap is brought back to the laboratory, it usually takes several days to count and analyze the mosquitoes from the trap.

[0009] Traps brought to the laboratory were inverted onto petri dishes, roughly cleaned, and then counted.

[0010] During the counting process, the technician may place the petri dish under a microscope and then use forceps to remove the insects one by one, checking for species and sex as needed.

[0011] Another related problem with the process so far is that employees trying to correctly identify the insect species cannot always identify the species of all the types of insects collected.

[0012] In addition, it is usually necessary to randomly pick a certain number of insects of a specific type from the collection from the traps and store them in small tubes to be sent to a laboratory to test whether the insects carry any viruses.

[0013] Sometimes, if a sterile insect technique (SIT) program exists in place, it may also be important to count how many are wild and how many are insects that were previously released intentionally by people.

[0014] In summary, insect surveillance, and in particular mosquito surveillance, requires expensive and tedious work as professionals identify, count and collect mosquitoes one by one from hundreds of field traps. This work is very exhaustive and the available capacity limits the number of traps surveyed. As a result, there is a large human error rate and the overall surveillance is inefficient and does not achieve the desired effect. Summary of the invention

[0015] The present invention can provide a method and apparatus for automatic insect counting, for automatic species identification and aggregation, thereby reducing cost and time while improving accuracy and consistency.

[0016] The device can automatically separate, classify, identify, count and map processed insects (e.g., mosquitoes), including identifying for each insect an indication of the corresponding field trap from which the insect came, and updating real-world data based on the mosquito population at the field trap coordinates.

[0017] The solution can provide benefits such as eliminating tedious and repetitive tasks, saving time, and reducing the potential for human errors in manual processes.

[0018] According to one aspect of the present invention, there is provided a method for separating a batch of insects from a trap into individual insects, comprising:

[0019] pouring the batch of insects into a container having at least one hole sized for a single insect;

[0020] The container is moved to agitate the insects therein so that individual insects exit through the holes onto a collection surface, thereby providing separated insects onto the collection surface.

[0021] In an embodiment, the container comprises a bottom plate, the movement comprises vibration, and the at least one hole is in the bottom plate.

[0022] In an embodiment, the container comprises a circumference, the at least one hole is in the circumference, and the movement comprises rotation.

[0023] In an embodiment, the motion further comprises vibration.

[0024] In an embodiment, rotation and vibration alternate in a cycle.

[0025] In an embodiment, the container comprises an upper cone and a lower cone, the cones meeting at a common base, the base providing a maximum circumference, and the at least one hole is at the maximum circumference.

[0026] An embodiment may include pouring a batch of insects into a container through a funnel.

[0027] In an embodiment, said collecting surface is a moving surface.

[0028] According to a second aspect of the invention, there is provided an apparatus for separating individual insects from a batch of insects, the apparatus comprising a container for the batch of insects, the container being motorized to provide movement to the container and having at least one hole sized for an individual insect so that an individual insect from the batch of insects can be pushed out of the hole when pushed towards the hole by the movement.

[0029] In an embodiment, the container has a bottom plate with at least one hole in the bottom plate, and the movement is a vibrating movement.

[0030] In an embodiment, the container has a circumference, the movement comprises rotation on an axis perpendicular to the circumference, and the at least one hole is in the circumference.

[0031] In an embodiment, the motion further comprises vibration in at least one axis.

[0032] In an embodiment, the motion comprises vibrations in three axes.

[0033] In an embodiment, the at least one hole comprises an inner side facing the interior of the container and an outer side facing the exterior of the container, and its diameter is smaller at the inner side than at the outer side.

[0034] In an embodiment, the container comprises an upper cone and a lower cone, the cones meeting at a common base, said base providing a maximum circumference, and at least one hole is located at said maximum circumference.

[0035] Embodiments may have a guide for directing the exiting insects from the at least one aperture to the collecting surface.

[0036] Embodiments may include a funnel for pouring the batch of insects from the trap into a container.

[0037] Embodiments may include a motor with an eccentric weight to provide vibration.

[0038] According to a third aspect of the present invention, there is provided a method for picking up an insect and placing the insect on a first surface, the method comprising:

[0039] Imaging the collection surface from above;

[0040] determining the presence of insects on a surface for picking up from the imaging;

[0041] determining from the imaging a current position of the insect on the surface as a pick-up position;

[0042] Using the robotic arm, moving the pick tool to a position above the pick location;

[0043] lowering the pick-up tool to the pick-up position;

[0044] operating suction to pick up insects from the pick-up location into the pick-up tool;

[0045] Using a robotic arm to move a pick-up tool with the insect to a position above the deposition location; and

[0046] The suction is removed to deposit the insect, wherein one of the pick-up location and the deposit location is an identification location for imaging the insect for identification.

[0047] In an embodiment, the identification location is a pick-up location and the identification made at the identification location defines the deposition location.

[0048] The method may include switching from suction to blowing at the deposition location to deposit the insects.

[0049] In an embodiment, the picking tool comprises a porous surface in a tube leading to a vacuum source, the insects being held on said porous surface by suction.

[0050] In an embodiment, the pick-up tool has a central air duct and a peripheral air duct, through which said suction is applied, and by which said blowing is provided.

[0051] According to a fourth aspect of the present invention, there is provided a picking tool for insects, comprising a hollow tube having a first end and a second end, the tube being connected to an air pressure source at the first end and having a porous surface near the second end, the tool further having a robotic arm for positioning the tool in three-dimensional space, the tool being configured to work with an imaging system to position itself above coordinates provided by the imaging system as the position of the insect on the surface, the tool being configured to lower itself above the coordinates and apply suction to suck the insect against the porous surface, thereby picking up the insect.

[0052] In an embodiment, the distance between the net and the second end is the thickness of an insect.

[0053] The tool may have a central air duct and peripheral air ducts, through which the suction is applied so as to centrally position the picked-up insects on the web.

[0054] When the destination is reached, the tool can turn off the suction, thereby depositing the insects at the placement location.

[0055] Alternatively, when reaching the destination, the tool may switch off the suction and may replace the suction with a blast of air, which is applied through the central and peripheral air ducts, so as to deposit the insects at the placement location.

[0056] According to a fifth aspect of the present invention, there is provided a method for identifying and counting images obtained in batches from field traps, the method comprising:

[0057] receiving a batch of insects from a trap;

[0058] placing the batch into a separator which discharges insects individually from the batch;

[0059] Collecting insects expelled onto moving surfaces; and

[0060] capturing at least one image for each insect on the moving surface; and

[0061] For each individual insect found in the corresponding image, the count is incremented.

[0062] The method may include taking a series of images from different angles for each insect on a moving surface and providing the images to a neural network to identify the insect.

[0063] The method may include using the identifier to define a destination for placement of the insect.

[0064] The method may include using a first camera to locate the insect and using a second camera to capture images from different angles around the insect.

[0065] The method may include rotating the insect on a spinning disk to obtain images from different angles.

[0066] The method may include placing a second camera on the robotic arm and moving the second camera around the insect to obtain an image.

[0067] The method may include illuminating the insect with an excitation wavelength to induce fluorescence.

[0068] The method may include obtaining images at different focal depths.

[0069] The method may include recognizing a posture of the insect and obtaining an image of the body part according to the position defined by the posture.

[0070] The method may include defining species-defining features using a decision tree, and positioning a second camera to image the body part based on the decision tree.

[0071] The method may include operating the separator to discharge separated insects onto a section of a conveyor belt, then stopping the separator and identifying the insects on the conveyor belt, and repeating the operation and identification.

[0072] The method may include using the identification to define a destination for placing the respective insect.

[0073] Multiple batches can be obtained from multiple traps, each having a different location, and the method can use the insect counts from the corresponding traps to generate or update a report or geographic map of insect distribution.

[0074] In an embodiment, the identification is based on an insect database of insects expected in the area of ​​the trap.

[0075] In an embodiment, the identifying includes leaving some insects unclassified as unidentified, or identified below a threshold level of certainty, the method including forwarding images of the unclassified insects for manual identification by an operator.

[0076] According to a further aspect of the present invention, there is provided a method for automatically identifying the genus, species or sex of an insect at an imaging location, the method comprising:

[0077] Taking a first image of the insect from above to identify a direction of the insect; and

[0078] using the first image to find at least one location where a first given body part can be imaged, and sending a camera to that location to capture a second image; and

[0079] Continue with additional locations and additional body parts until enough images are available to identify the insect.

[0080] In an embodiment, the first image is taken using a first camera located overhead, and the second image is taken from a second camera on the robot arm.

[0081] In an embodiment, the first image is taken using a camera located overhead, and the second image is taken from a camera on track.

[0082] In an embodiment, the first image is taken using an overhead camera, and the second image is taken from either the overhead camera or the side camera.

[0083] Unless otherwise defined, all technical terms and / or scientific terms used herein have the same meanings as those generally understood by those of ordinary skill in the art to which the present invention belongs. Although materials and methods similar or equivalent to the methods and materials described herein can be used in the practice or testing of embodiments of the present invention, exemplary methods and / or materials are described below. In the event of a conflict, this patent specification (including definitions) will prevail. In addition, materials, methods and examples are only illustrative and are not intended to be necessarily restrictive. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Some embodiments of the present invention are described herein by way of example only with reference to the accompanying drawings. Reference will now be made in particular detail to the drawings, emphasizing that the details are shown by way of example and for the purpose of illustrative discussion of embodiments of the present invention. In this regard, the description in conjunction with the drawings makes it apparent to those skilled in the art how to practice the embodiments of the present invention.

[0085] In the attached picture:

[0086] Figure 1 is a simplified schematic diagram illustrating an eight-part process for separating, counting, identifying and pooling insects from a batch obtained from a trap and plotting the results according to an embodiment of the present invention;

[0087] Figure 2 is shown for implementing Figure 1 A simplified diagram of the equipment of the process;

[0088] Figure 3 yes Figure 2 A simplified schematic diagram of the control of the device;

[0089] Figure 4 yes Figure 1 A simplified flow chart of the process;

[0090] Figures 5A to 5C Batches of insects from a trap are shown;

[0091] Fig. 6A and 6B The separator according to the present invention is shown. Figures 5A to 5C batches;

[0092] Figures 6C to 6E is a view of the interior of a separator device according to an embodiment of the present invention;

[0093] Fig. 6F and 6G shows insects being individually discharged from a separator device according to an embodiment of the present invention;

[0094] Fig. 7A and 7Bis a side view of a double cone separator according to an embodiment of the present invention;

[0095] Fig. 8A and 8B is a simplified view of a separator according to an embodiment of the present invention looking into the space of a container; and

[0096] Fig. 9 is a simplified diagram showing the picking stage of picking up individual insects after separation according to an embodiment of the present invention;

[0097] Fig. 10A and 10B are two simplified diagrams showing a pick and place tool according to an embodiment of the present invention;

[0098] Fig.11A and 11B It is shown by Fig. 10A Two simplified diagrams of insects placed by a pick and place tool;

[0099] Fig. 12A and 12B is shown connected to a gas pressure source Fig. 10A Two simplified diagrams of the pick and place tool;

[0100] Fig.13 is a top view of a group of isolated mosquitoes provided according to an embodiment of the present invention;

[0101] Figures 14A to 14C It shows placing mosquitoes and taking images at different focal depths for identification according to an embodiment of the present invention;

[0102] 15A to 15E show a series of images taken at different angles to obtain features for identifying species according to an embodiment of the present invention;

[0103] Fig.16 is a simplified flow chart illustrating a procedure for obtaining and imaging insects for automatic identification according to an embodiment of the present invention;

[0104] Fig.17A and 17B is a simplified image showing the collection of insects in vials or test tubes based on identification using an embodiment of the present invention;

[0105] Figures 18A to 18H are different views including cross-sections of a pick-and-place tool according to a second embodiment of the present invention;

[0106] Fig.19A and 19B is a view of a screen for manual identification according to an embodiment of the present invention;

[0107] Fig. 20 is a simplified diagram showing how the distribution of insects found by an embodiment of the present invention may be displayed in map form;

[0108] Figures 21A to 21D is a view of a separator according to a second embodiment of the invention, wherein the outlet opening is located in the bottom plate of the separation chamber;

[0109] Fig. 22 is a simplified diagram illustrating an embodiment of the present invention in which insects are separated and counted on a moving conveyor belt;

[0110] Figures 23A to 23H is a simplified diagram showing another embodiment according to the present invention for separating, imaging and collecting insects from a batch;

[0111] Fig.24 is a simplified diagram showing the main anatomical features of mosquitoes for classification purposes;

[0112] Fig.25A and 25B is a simplified diagram showing an example of insect classification using an embodiment of the present invention; and

[0113] Fig.26 is a simplified diagram showing a separator according to a further embodiment of the present invention. DETAILED DESCRIPTION

[0114] In some embodiments of the present invention, the present invention relates to insect monitoring, and more particularly, but not exclusively, to systems and methods for monitoring insect populations, including devices for use in said systems and methods.

[0115] A method is provided for separating a batch of insects from a trap into individual insects, comprising pouring the batch into a container having at least one hole, the hole being sized for an individual insect; and moving the container to shake the insects therein so that the individual insects are discharged through the hole onto a collection surface, thereby providing the separated insects onto the collection surface. The insects from the trap can then be counted, and image recognition can be used to identify the genus, species or sex. The process can be performed for multiple traps, and the location data can be stored together with the insect identification to give a map of the insect distribution.

[0116] Before explaining at least one embodiment of the present invention in detail, it should be understood that the present invention is not necessarily limited in its application to the details of construction and arrangement of components and / or methods set forth in the following description and / or illustrated in the drawings and / or examples. The present invention is capable of other embodiments or of being practiced or carried out in various ways.

[0117] Now referring to the accompanying drawings, Figure 1An eight-part process for monitoring an insect population according to an embodiment of the present invention is shown, wherein conventional insect traps are used to capture the insects.

[0118] This embodiment may include up to eight different process sections 1 ... 8, such as Figure 1 shown, is used as the key to discuss the following figures.

[0119] Figure 1 The eight parts include methods for automatic or semi-automatic mosquito identification-counting-collection-mapping. Not all parts are mandatory. For example, only having parts 1-7 will only exclude mapping from the process, which may be sufficient in some cases.

[0120] Additionally, there are schematic diagrams of the apparatus and methods, and below that are descriptions of the processes, illustrating the role of the various elements in the system.

[0121] Additional embodiments for all or part of the process are presented, it being noted that the embodiments may be mixed and that parts of the process of different embodiments may be matched together in any manner deemed appropriate by a person skilled in the art, who is expected to perform a selection of the features found in this specification in order to match his precise requirements. As is known, field conditions may vary greatly, such that a combination that works well in one environment may work poorly or not at all in another. The eight parts of the process include 1 collecting insects from a trap. The trap may be a conventional low-cost trap. The collected insects may then be poured into an apparatus according to an embodiment of the present invention, 2 which then automatically separates out individual insects, 3. Individual mosquitoes are picked out, for example by a robotic arm, 4 for high-resolution visual identification 6. Mosquito pooling is automated to reduce human error, 7, which will be discussed in more detail below, and the data is visualized, 8.

[0122] Reference now Figure 2 , providing the implementation Figure 1A schematic diagram of a system for the method of. The flow is shown from right to left in the figure. No part related to collecting field traps is described, and it is assumed that the insects have been placed in the separator unit 11, the mosquitoes are separated into individuals, the insects are discharged as individual mosquitoes downward onto the conveyable surface 12, and the suction tube transfers them from the conveyable surface to the rotatable disk 14, which provides a constant distance and rotation angle between the insects and the imaging camera 15. The position of the insect can then be corrected as required to ensure that the insect is located at a certain position on the imaging disk, preferably with a tolerance of 0.5mm, and the correction is achieved by moving the imaging disk using an XY motorized system 16 according to the insect coordinates received from the top view camera 17. Once positioned within a tolerance of 0.5mm, the imaging camera 15, which can be located to the side (for example, at a 30 degree angle to the surface where the insect is located), takes multiple images of the insect as the disk rotates. When the required number of images are taken or the required number of rotations of the disk 14 are completed to provide observations from a sufficient number of different angles, the robot arm 18 is then used to transfer the insect towards a row of vials 19. Each insect is then placed in a vial that matches the species of the now identified individual mosquito or other insect. In addition, a generic vial can be assigned to all mosquitoes for which the model confidence, based on their species, is below a threshold confidence level, such as 70%, meaning that they were not clearly identified. Figure 1 The collection part 7.

[0123] As will be explained in more detail below with reference to the accompanying drawings below, the separator can hold a batch of insects (e.g., mosquitoes) within two cone-shaped structures. The two cones are connected together into a single structure to encourage the insects to move downward along the ramps on both ends toward the middle of the two cones.

[0124] The cones can rotate together, causing the insects to continuously fall toward the discharge hole. Optionally, a lateral movement of the two cones or a sudden change in the direction of rotation can be provided to generate vibrations. As a result of the vibrations, the insects may separate from each other to fall individually through the discharge hole.

[0125] The two cones can further serve as storage compartments. If the insects are alive, the compartments can also be kept at low temperatures to support the insects to remain stationary or immobile and not adhere to each other in a manner that could cause harm.

[0126] The separator compartment can continue to rotate to separate any large input batches of insects into small groups and to cause individual insects to fall and exit the discharge hole under the influence of the vibrations. The internal shape preferably supports the mosquitoes to continue to move towards the exit hole, so the preferred shape of an internal cone can be implemented, however other separator compartment shapes can also be implemented, such as separator compartment shapes with a circular perimeter or an external geometry, so that continuous separation and discharge of insects is provided.

[0127] The surface of the storage compartment can have one or more exit holes around the periphery of the two cones along their interface. The following is shown in more detail how the same function that vibration can separate insects and then allow insects to be discharged or fall as individuals can be achieved in other ways, such as using a flat vibrating surface with one or more exit holes. Other shapes besides two cones or flat surfaces can be implemented by experts based on embodiments of the present invention that involve using vibration to separate into individual insects and have exit holes. This type of vibration can cause separation by vibration and the discharge of separated individuals throughout the exit hole or opening.

[0128] Once the insects have fallen, they reach the movable or transportable surface 12 .

[0129] The problem that arises is the clogging of the outlet hole. When insects leave through the discharge hole, if they get stuck in the opening and cannot leave, they may cause a blockage. The diameter of the hole and how to use a similar phase outlet shape to prevent blockage will be discussed below. In one embodiment, there is an option to place an air source (e.g., an air flow from an air duct) at a fixed position just in front of the location where the discharge hole passes, so as to spray air toward the inside of the compartment, push the blocked insects back, and thus unclog the opening. Unclogging can also be solved by momentarily increasing the vibration force, or even by increasing the size of the outlet hole by taking out the outlet hole (preferably momentarily taking out the outlet hole).

[0130] The top view camera 21 can identify the location of insects on the conveyable surface 12.

[0131] Once the coordinates of the individual insects on the conveyable surface 12 have been identified, the pick and place robot arm 13 is directed to reach above the insects. The robot arm preferably moves toward the insects in the Z axis and collects the insects by holding the insects using suction. With the help of a mesh (i.e., a net or porous surface) toward the end of the suction tube, the insects are held against suction through the tube. A more detailed description of the tube and mesh structure is provided below.

[0132] A pick and place robotic arm 13 can transfer the insect towards a rotatable disk 14 which enables relative motion between the recognition camera and the insect in order to obtain a collection of consecutive images of the same insect from different angles.

[0133] It should be understood that the transfer stage of the individual separated mosquitoes can be a single stage in which the insects are moved directly toward the imaging location after being discharged from the separator exit hole; or it can be a multi-stage process as described above, which includes a linear movement of the surface where the insects are located, followed by a second transfer using a pick and place robotic arm guided by a camera, with the pick and place directed to the location where the insects are located and moved to the imaging location.

[0134] To place the insect on the disc 14, the controller 20 can stop the suction of the picking tool and the insect falls from the surface, net or porous surface where it is held under the pressure difference because there is no longer air pressure to hold the insect. To ensure accurate placement on the disc relative to the imaging camera, in a second stage, the motorized dual axis motor 16 to which the rotatable disc is attached can move the disc to a position to place the insect at a precise position or distance relative to the imaging camera, preferably to a tolerance of 0.5 mm.

[0135] After positioning, the controller 20 can command the imaging camera 15 to take successive images of the insect as the disk 14 rotates. The disk 14 can stop rotating before taking each image to ensure that all images are in focus and to reduce or avoid blur.

[0136] Other possibilities for transferring the insects could consist in letting the insects at the end of the conveyor belt fall along a funnel towards the tray, which would then be located at a lower position.

[0137] After the image is generated, a computer vision model (preferably a neural network model) can identify unique species identification features, as will be discussed in more detail below, and can then identify the insect species based on this information.

[0138] After the insect is identified, the insect is transferred toward a vial set 19 using the same or another robotic arm 18, preferably based on suction or air flow, and transferred into a vial matching that particular insect type.

[0139] Reference now Figure 3 , which is a schematic block diagram of the control system and the interfaces between the different modules.

[0140] like Figure 3 As shown, the controller 20 provides command and feedback control to various parts of the system and receives feedback. These parts include a separation module 30 with 1...n vibrating rotation axes, a rotating disk x and y axis motor 32. In addition, a pick and place insect module 34 has two single motor axes, and there may be more than one such pick and place module. A conveyor module 36 operates a conveying surface.

[0141] In addition to command and control, there are elements that provide triggers for operations such as images or user commands, including a user interface screen 38 and location and identification camera modules 40 and 42, where there may be more than one camera.

[0142] Reference now Figure 4 , which is a simplified flow chart illustrating the above-described embodiment.

[0143] Insects are collected from field traps and poured into a separator device, 50. Individual insects are then separated from the batch, 52. The separated insects are transferred to an imaging location, 54. Images are taken at different angles around the insect, 56. Using images and image recognition techniques, the species, sex, etc. of the insect are identified based on identification features that are found to be unique, such as wing vein patterns, dorsoventral patterns, dorsal thoracic patterns, mouth parts, shape of mouth parts, and leg segments, 58.

[0144] The system is updated with the number, sex, etc. of each identified species, and, if necessary, with the number of unidentified or provisionally identified insects, and in many cases this may be all that is needed.

[0145] However, in some cases, further testing may be required on the insect or some of the insects, for example, to find out how many of the insects carry a disease, whether it is a virus, bacteria or an animal parasite. According to 62, the insects are collected into vials corresponding to the identified species or sex. The vials can be stored at low temperatures, for example, to ensure that the viruses remain in the insects so that they are still present for later testing.

[0146] The process is now described in more detail.

[0147] Collect insects from field traps. The insects may include different types of mosquito species and possibly other insects caught in the traps. The traps may include insects of different sizes.

[0148] Traps can be left in the field for several days, or can be placed at night and removed the next morning to ensure that most insects are alive. In this way, many insects remain alive, which is sometimes important for testing the presence of vectors carried by the captured insects (e.g., mosquitoes).

[0149] according to Figure 5A , collect insects from the trap and can be conveniently poured into Figure 5B and 5C On the petri dish shown, Figure 5B and 5C are two different views of a petri dish containing Aedes albopictus mosquitoes.

[0150] In the prior art, counting and identifying species for such numbers of mosquitoes is a time-consuming and repetitive task.

[0151] like Fig. 6A As shown, insects are poured into a separator according to an embodiment of the present invention. Specifically, the contents of the trap overflow directly from the trap, or overflow from the petri dish into the separator 70.

[0152] like Figure 6B As shown, the separator 70 includes an insect storage compartment 72, which is open and ready for inserting insects, after which it can be closed. A group of mosquitoes are already inside waiting to be separated, and more mosquitoes are being poured in. The module includes a separation surface 74, an outlet hole 76 and a vibration motor 78.

[0153] Figure 6C The same top view is shown, and the closure 80 is shown.

[0154] The module ensures that the batch of mosquitoes is separated into individual insects, or at least into smaller groupings, in order to allow the number of mosquitoes to be counted, identified and or manipulated. For example, at the high end, the module manages to separate every single mosquito and all mosquitoes that fall from the separator onto the receiving surface are completely separated from each other and then identified one by one as will be described later. On the other hand, in different cases, for example, only 25% of the mosquitoes are successfully separated, which means that sometimes there are still insects that are in contact with each other, thus making it more difficult for the visual system to identify individual mosquitoes because important visual features may be blurred. In this case, when assuming that 25% of the mosquitoes are successfully identified, the user can extrapolate the results 4 to estimate the number of mosquitoes and their species for the entire batch. Although this will not be an accurate result, it does provide a result with a certain statistical reliability.

[0155] like Fig.6D The separator 70 shown shows a group of mosquitoes 82 moving toward the end of a shelf-like member that serves as a shelf 84. The shelf 84 picks up some mosquitoes from the bottom at each rotation and spills them downward toward an opening on one end from which individual mosquitoes are expelled onto the other end of the surface.

[0156] The storage compartment 72 may be rotated about its axis so that the individual units are ejected or fall out of the exit aperture.

[0157] Experts can use the concept of the present invention for other applications of separating mosquitoes that require separation of mosquitoes, such as for release systems in which a batch of mosquitoes waiting to be released is stored in a compartment and then a continuous stream of individual insects is provided from the storage compartment as separation begins. Such a module can be attached under a UAV (Unmanned Aerial Vehicle) to release insects over an area for biological control. Likewise, the module can also be attached to other release systems.

[0158] Fig.6D and 6E Successive stages in the use of the separator 70 are shown.

[0159] exist Fig.6D and 6EIn FIG. 1 , a group of mosquitoes is shown. As the machine rotates and preferably oscillates back and forth, for example perpendicular to the direction of rotation, individual mosquitoes are expelled from the outlet holes and received on the surface. Fig. 6F shows the insects being ejected one by one, and Figure 6G Shown are individual insects collected on a surface.

[0160] Reference now Fig. 7A and 7B , which respectively show a plan view and a 3D view of a separator 70 according to an embodiment of the present invention.

[0161] When the surface on which the exit hole 86 is located vibrates, the exit hole enables the insects to fall downward. As it rotates, the internal structure of the rack 84 is able to continuously separate a larger group into a plurality of smaller groups so that they are fed one at a time toward the exit hole in the compartment. The closure 80 is closed after the insects are dumped. The storage compartment 72 has a double cone shape. If necessary, the handle 90 can be used to rotate the motor and the storage compartment, for example when the field location lacks power, and the first rotary motor 92 provides the rotary motion in most cases. The second rotary motor 94 can be used to provide a linear motion to provide vibration, thereby allowing the trapped mosquitoes to fall out.

[0162] It should be noted that the rotating unit with the outlet aperture can be realized without an internal shelf, however the rotating unit preferably has an internal shelf and thus the main embodiment comprises an internal shelf.

[0163] Reference now Fig. 8A and 8B , which are respectively a cross-sectional view and a perspective view of the opened storage compartment from above and showing the motor at the rear.

[0164] The rotating storage compartment 72 has at least one outlet hole 82 around the perimeter and a rack 84 inside holds batches of mosquitoes which push or pick up the mosquitoes and let them fall toward the hole. A first rotary motor 92 and a second rotary motor 94 for rotation and vibration are shown at the back.

[0165] Thus, the initial batch of insects is divided into smaller groups which are picked up in small quantities by the mechanical element as the compartment 72 rotates and which, as the rotation continues, eventually fall towards the middle section of the two cones where there are outlet holes 82 for the individual discharge of the insects. The separation process can be improved if vibration of the surface is provided, by moving the entire drum structure back and forth at regular intervals.

[0166] If the insects need to be kept at low temperatures, another advantage of the above design is that because it is a relatively closed compartment it can be kept at a lower temperature, for example by introducing insulating walls into the structure.

[0167] The vibrations may be provided in a cycle, for example, the control program commands the motor to rotate the storage device clockwise, and after the cycle, the vibrations are momentarily increased in an axis perpendicular to the rotation, causing mosquitoes near the hole to pass through the hole and thereby separate from other mosquitoes to which they may be attached.

[0168] The speed of rotation, the duration of the vibrations and their direction, as well as other parameters, can all be varied to optimize the individual separation of the mixture of insects obtained.

[0169] As the drum rotates, the airflow duct can clear any potential blockage from the outlet of the separator unit as follows Fig. 23B -H.

[0170] An alternative embodiment for avoiding clogging includes creating a hole that is narrower on the inside but widens outwards, ensuring that no insect that is too large to fit through will manage to get in. For example, the entrance side could be 4mm and the exit hole 6mm, or whatever is appropriate for the insect in question.

[0171] It should be noted that the rotating compartment has been described so far as being drum-shaped and may have zero or more shelves to catch insects as the compartment rotates and allow them to fall back toward the bottom of the drum and then exit from the exit hole. The drum may have exit holes of different sizes, allowing insects of different sizes to be deposited. During operation, the drum may rotate both clockwise and counterclockwise (e.g., 1 revolution clockwise, then 1 revolution counterclockwise, or any other combination).

[0172] Once the mosquitoes are discharged from the separation module, they are Fig. 9 The insects land on a surface where they are transported and / or picked up by a pick-up unit on a robot arm and transferred to the imaging position.

[0173] Specifically, the expelled mosquitoes 100 fall onto a conveying element or conveyable surface 12. It can be a conveyor belt, or it can be a tray placed on a conveyor belt (or other conveyor belt structure). The surface on which it lands is preferably synchronized with the operation of the separator 70.

[0174] The transfer may be performed in batches. Thus, each time a batch of a preset number of mosquitoes is present on the surface, the transfer is performed. Alternatively, a single mosquito is transferred each time.

[0175] A tray may be located above the conveyor belt to directly receive small quantities of mosquitoes as they are discharged from the separator unit.

[0176] To avoid accumulation of insects on the tray, the tray can be moved away from the separator when it is filled. The speed of movement is adjusted according to the speed at which the insects fall. A conveying element moving at a suitable rate can be used to form a line of insects.

[0177] In an embodiment of the present invention, the rate of insect expulsion is related to the rotational speed of the separation module, which expel mosquitoes while rotating.

[0178] The conveying surface can, for example, move the just collected insects laterally, say 30 cm from the center of where they fell, after which they are taken away for the imaging process.

[0179] The position camera 21 is located above the surface where one or more mosquitoes are located so that it can provide information about the coordinates of the insects on the surface. The coordinates are then sent to the pick and place robot arm 13, which picks up a single mosquito and places it at an imaging station (hereinafter also referred to as an identification station). The identification station is where the species and or sex of the insect is identified automatically or manually.

[0180] Reference now Fig. 10A and 10B , the pick-up and place element 110 comprises a tube having a porous surface 112 (e.g. a net) located just behind the end 114 of the tube to form a container 116. An air tube 118 connects the tube to a vacuum source. In use, the tube approaches the insect and sucks it up so that it remains against the net.

[0181] Fig. 11B and 11A A tube connected to a vacuum generator 120 is shown via an air tube 118. A porous surface 112 is located at the end 114, the mesh allowing relatively free airflow but not allowing insects to pass through. The size of the holes in the mesh can be selected according to the type of insect being monitored.

[0182] The pick and place element 110 is connected to a moving element on at least two axes (X and Y) to be positioned just above the mosquito. A piston or other movement on the Z axis then lowers the tube to suck the insect onto the web.

[0183] The pick and place element can now be moved to transfer one or more picked insects from one point to a second point.

[0184] Fig. 12A and 12B It is shown how a pick and place element 110 releases a single mosquito 100 onto an imaging surface. Fig. 12A The element is shown just after the mosquito is released, and Fig. 12B The element is shown immediately afterwards, when the element is pulled up and the insect is left behind for identification.

[0185] As mentioned above, the net is not at the end of the tube of the pick and place element. Instead, the end of the tube may be below the net at a certain distance, for example 2 mm, so as to damage the insect when approaching the surface on which the insect is to be released. The net may alternatively be located at the very end of the tube, in which case the insect is released before reaching the surface, or the net may be higher in the tube so that the pick and place element can actually reach the surface, thereby ensuring accurate placement on the imaging tray.

[0186] Therefore, the pick and place element comprises an air duct having a net on one end and connected to a valve on a second end to pick up individual insects.

[0187] The switch can be controlled by the controller software to alternate the air duct between suction mode or blowing mode or off.

[0188] In suction mode, the element sucks the insects and holds them securely just below the net. When in blowing mode, the insects fall from the net, since the airflow is now directed inside the duct towards the net and outside. It should be noted that having the blowing mode is not mandatory and that when switching to off, mosquitoes held on the net may now fall downwards.

[0189] In order to pick up an insect using suction, its coordinates are identified (eg by a position camera 21) and provided to a mobile arm holding a suction unit.

[0190] As previously mentioned, an alternative to pick and place using suction may include transferring the insects from the surface they fall from the separator to the identification location by allowing the insects to fall from a funnel at the end of the conveyor belt and thus directly to the imaging location.

[0191] Once the insect is positioned, in embodiments, an enhanced precision positioning system may use dual axis motion to ensure that the insect is located in the same position relative to the imaging camera each time.

[0192] In an embodiment, in order to drop the insects as close as possible to the surface of the identification area, the height of the individual insects can then be estimated to determine how high above the surface the insects need to be released to avoid injury, based on the size of the mosquitoes as seen from above by the camera 21. This distance is preferably minimized so that the insects are placed on the surface as accurately as possible.

[0193] Reference now Fig.13, which is an image of separated insects of different sizes. For the largest insect in the image, the tube can be approached at a higher height than when approaching the smallest insect in the image, in order to pick up the insect. The same is true when placing the insect, and the advantage is that when the insect falls on the surface, if an air flow is used to blow the insect downwards, then the minimized distance reduces the chance that the insect will be blown to the side and thus located inaccurately. Additional correction means may be provided to move the disk on which the insect is located, so the insect is located in a fixed position relative to the recognition camera.

[0194] Various forms of fixation may be employed to keep the live insects in place. For example, cold air or CO2 may be applied to an area such as a storage compartment, or the entire separator may be in a closed environment at a low temperature. Alternatively, the insects may be fixed to a specific surface, such as using a pressure differential (suction) below the conveying surface 12, thereby ensuring that the live insects do not wander off.

[0195] Individual mosquitoes are positioned for imaging to facilitate accurate imaging and successful identification.

[0196] When the pick and place element 110 is moved to the placement position, a valve operated by the system controller is switched so that instead of suction, the air flow is now directed downward, down and away from the tube so that insects held on the web by suction are now released and fall.

[0197] As mentioned above, it is possible to repel mosquitoes, but it is not mandatory, and the mosquitoes may simply fall away once the suction is turned off.

[0198] The placement of the insect at the imaging position can be accurate to the order of 0.5 mm in order to ensure that the entire insect body is in view and in focus. This means that the camera sensor and lens can be selected and the camera and its lens positioned at a distance from the mosquito so that most of the insect body is in focus, with a depth of field of about 0.5 mm, so that if the object or part of it (such as the edge of a leg) is located more than 0.5 mm from the center of focus, it is no longer in focus. It is therefore desirable to be able to locate objects with such accuracy.

[0199] It should be noted that in order to manage and image the different features of the insect, while it is preferred to have them all in focus, this is not mandatory in order to achieve the described solution, and in fact in some cases the size of the insect may be larger than the depth of field of the chosen optics. In this case, a collection of images may be generated, each at a different focus, such as a collection of captured images, each at a different focus. The collection of images may then be reconstructed to create a single focused image of the insect's body, or even a 3D model. It is also possible to look for different unique features for classifying the insect and set the focus to the features that best show them, however this may slow down the process.

[0200] Fig. 14B and 14C An example is shown in which a mosquito is rotated and then photographed with a different focus on a different area each time. Specifically, Fig. 14B Focus on the wings, and Fig. 14C Focus on the abdomen.

[0201] The level of localization accuracy may depend on the optical system used by the recognition camera, making 0.5 mm only a reference value. We will now explain how to localize a mosquito with a resolution of 0.5 mm.

[0202] Initially, the suction tube holding the mosquito is moved to a predetermined position and the suction is stopped and / or the mosquito is gently blown while the net is not too close to the surface to squeeze the insect, but on the other hand close enough to prevent the insect from moving along the surface when it detaches from the holding surface.

[0203] The top view camera 17 is positioned to observe the imaging location and can send the coordinates of any identified objects at the imaging location to the controller 20 .

[0204] The control software in turn sends correction commands to the motors of the dual axis motor 16 of the disk 14 at the imaging position to move in either axis so that the mosquito on the surface is repositioned in the desired position within a 0.5 mm tolerance.

[0205] In another embodiment, first, the robotic arm receives the coordinates of the insect on the surface where the separated insect arrives. The robotic arm can hold a camera to locate the insect on the surface. Alternatively, another camera can provide such an image in an intermediate stage, and then the robotic arm rotates around the mosquito, generating multiple images at different angles around the insect and even above. Thus, by using a robotic arm, the requirement of having a specific distance between the object and the camera sensor or lens is obtained, only using a different implementation.

[0206] Thus, imaging can be performed around individual insects at various angles, and additionally, but not necessarily from above individual insects. Now, as mosquitoes are identified and manipulated at an individual level, automatic pooling becomes possible so that all individuals identified as belonging to the same species, sex, etc. can be placed together in the same vial.

[0207] In another embodiment, a camera attached to the robot rotates around each of the different insects located on the surface after being separated, without the need to transport them to a designated identification station. In this case, once identified, the insects can be moved directly from the surface to the appropriate vials if automatic collection is also required.

[0208] Once one or more insects have been imaged, the surface on which they were received after separation is cleared of insects to allow a new batch of insects to be discharged from the separator unit and received on the receiving surface, which is also the imaging surface in this other embodiment. This is particularly applicable when collection or collection is not required, or when collection or collection is not required for all found insects, but only for certain species, but may always require collection or collection for partial specimens or small amounts of dust. Cleaning is applied by having a robotic arm pick up and remove individual insects with a suction tube, or by using a blower to blow air on the surface and blow away any objects on the receiving surface. Thus, an air flow is used to remove insects from the surface.

[0209] This cleaning method of blowing air onto the receiving surface is applicable to all embodiments herein.

[0210] Reference now Fig.15A , which shows an image used in a classification process to identify species, etc. The classification may be performed using a trained neural network using three or more layers.

[0211] The method includes obtaining images of the same insect at different angles or different perspectives, for example from consecutive frames, to determine whether a particular feature is present in any or multiple images.

[0212] The method includes obtaining a plurality of insects, imaging the insects, and identifying at least one of the group consisting of wing vein patterns, dorsal-ventral patterns, dorsal-thoracic patterns, and leg segments from the imaging, thereby identifying the species of the insects.

[0213] like Figures 15B to 15E As shown, the above features can be visualized due to rotating the insect after isolating it and imaging it at different angles and different viewing angles, thereby ensuring that at least one such feature is observed. The different views show the light gray stalk on the head area, in this particular case, light gray hair near the head area, and different unique features on the thorax, abdomen, etc., which help identify the species of the mosquito.

[0214] As discussed, the mosquito is imaged in multiple images from different angles by rotating the surface on which the mosquito rests, or more generally by creating a relative motion between the lens and the insect. Thus, imaging can also be achieved by moving the camera around the insect. Another possibility is to move the camera 180 degrees around the insect and rotate the disk 180 degrees in order to make the entire machine more compact in space. Once the images are available, computer vision algorithms identify the species by locating some or all of the features in the body parts from the multiple images.

[0215] As the disk is rotated, images can be taken from a fixed camera, viewing the insect at a particular angle (e.g., 30 degrees above the surface). In an embodiment, the disk stops rotating before each image. For example, the disk can be rotated in 45 degree increments, resulting in 8 different images. Different angle increments are possible. For each angle, a single image can be taken, or multiple images of different focal lengths can be taken.

[0216] The image can then be sent to a computer which identifies the genus, species or sex as required based on the visual model. That is, in the case of disease control, the presence of a specific species is often of interest and is often sex specific. For general environmental monitoring, population mix is ​​of interest and any level of detail data such as genus may be required.

[0217] As described above, computer vision can use a neural network, and the network can implement a trained model. The training can include a human operator labeling one or more insects of the genus, species, or sex of interest, the label specifying which mosquito species it is.

[0218] One embodiment provides an insect identification method for release and capture operations. Insects with fluorescent markers are released, and a camera can be provided with a light that activates the marker and can then be identified. The scene of the camera can be illuminated using a wavelength selected to excite the fluorescent markers on individual insects from traps that were previously released and marked. This type of data is valuable to researchers trying to understand the flight distance and other behaviors of the insects they have released.

[0219] After the identification process, if a sample is desired, the insects can be transferred again, preferably to a specific vial, tube, or storage compartment designated to hold that particular species.

[0220] Insects that were not successfully separated or that were not successfully identified can be placed into a separate vial to include all those insects that were not automatically identified or separated.

[0221] The number of rotations and the angle of each rotation can be preset, for example, one image every 15 degrees. In addition, one or more images can be taken from above the insect (i.e., above the rotating disk), and the software guides the camera according to the precise coordinates of the insect in a predetermined coordinate system. The images from above can be used to provide information, such as the posture of the insect, the corresponding position of the head and abdomen, and based on the understanding of where the unique features are, the software can guide the system to move the camera or disk to those positions and speed up the process. This process can alternatively take an image every 15 degrees, for example, regardless of the posture of the insect.

[0222] As described above, in the case where the camera depth of field enables only a portion of the entire object to be viewed in focus, since the object width is greater than the depth of field, a collection of images with different focal points is obtained to ensure sufficient data for different features of the object, such as Fig. 14B and 14C When only part of the image is in focus, the analysis process becomes easier because the out-of-focus part can be ignored.

[0223] During the recognition process, different body parts can be cropped, and the different body parts that are recognized, such as the head, abdomen, wings, etc., can be sent separately to a visual processing model (such as a neural network trained to recognize species based on body parts alone). For example, the Aedes albopictus has a very distinctive white stripe on its head, so an algorithm can identify and crop the area of ​​the head, and then analyze the head image by locating the white stripe.

[0224] Reference now Fig.16 , which shows the procedure in one possible embodiment of the identification process. An initial orientation image is taken - 130 - to locate the insect and, if necessary, determine its position - 132 .

[0225] The insect or the camera is moved so that the insect reaches a desired initial position relative to the camera 134. One or more images are then captured, and the camera is moved to successive imaging positions to capture more images, ultimately obtaining a sequence of images 136.

[0226] After imaging the features, the images are sent - 138 - to the model, which provides a score for each image, which class (species, sex, etc.) it matches best. The total score for the sequence of images then gives the identification result for the entire collection of images, thereby identifying the insect. If a mosquito is not identified with a high enough confidence by the computer, it can be marked as "other" or "unknown" for manual identification later by an operator, who can manually mark all such unknown images.

[0227] In an embodiment, the controller may direct the robotic system to move the camera or disk until one or all of the set of features appear in the image.

[0228] Considering the model in more detail, in an embodiment, for a given top view image of a mosquito, the goal of the real-time machine learning is to significantly reduce the number of side view images required to correctly detect the mosquito species. A deep learning model can be used to exploit morphological features of the mosquito in a manner similar to that used by human experts to identify mosquitoes.

[0229] Images of mosquitoes with specific poses and body parts (e.g., flat wings, legs, abdomen, thorax, nose) are required to achieve good classification performance.

[0230] Finally, by acquiring only a few optimal side views, machine performance can be significantly improved and time and memory consumption can be reduced.

[0231] Classification of mosquito species with high interspecific similarities and intraspecific variations is challenging.

[0232] One approach is to prepare sparse typical top-view and some side-view photos covering various poses / stances / viewpoints and certain body parts.

[0233] For each mosquito species, we acquire sparse N top-view images with various poses / stances / viewpoints as well as specific body parts.

[0234] N = multiple typical top views of the mosquito body in different poses / postures / viewing angles, [estimated N ~ 2-3 back, 2-3 side, 2-3 abdomen]

[0235] For each top view pose / gesture of the mosquito body, the camera can be manually rotated 45 degrees and a picture taken. There may be up to 8 top view images in total.

[0236] For each original top view, M side view images with important morphological features such as wings / legs / abdomen / thorax / head, etc. required to achieve good classification performance can be obtained without rotation.

[0237] M = some typical side view images with notable morphological features like wings / legs / abdomen / thorax / head. See feature table file. These side views are only taken once, as they are the same for all 8 top rotations, [estimated M ~ 3 side view images]

[0238] The side view angle may have an elevation angle fixed at 27 degrees above horizontal. The azimuth angle range is 0, 360 degrees. The tilt angle is fixed at 0 degrees.

[0239] The object-camera distance is fixed.

[0240] Each acquired index image is labeled by the following parameters:

[0241] Top or side view

[0242] Top view with body pose and camera relative pose in azimuth [in case of 8 views; 0, 45, 90, 135, 180, 225, 270, 315] where body pose is dorsal / lateral / ventral as described in feature table file

[0243] Side view, where direction = (elevation, azimuth, tilt) angles, camera-object distance and its typical organs [wings / legs / abdomen / thorax / head]. In our case, the tilt is 0 and the elevation is 27, so the direction = (27, azimuth, 0) and the distance is fixed.

[0244] For a given top view image, the model can find the best match with the database top view and thus obtain a side pose view with unique features.

[0245] Next, the camera is moved to the optimal position and orientation and images of the insect's surroundings are acquired.

[0246] For a given insect, we can take a top-down view [image source].

[0247] The model then finds the best top view match by running a similarity algorithm on all (source, destination) image pairs, where the source is the current input top view image, and the destination is any top view image in the database. For example, assuming there are two species, each with 64 top views, the total number of matching pairs is 128. This can be run in batches / in parallel, and the end result is the best top view with the highest probability match.

[0248] From the best top view, the best side view direction / distance can be extracted from the labeled database.

[0249] For example, suppose the best top view is a top side view with a 45 degree angle, but in the database we have a best side view which corresponds to a top view with a lateral angle of 0 degrees.

[0250] Therefore, the azimuth angle is increased by 45 degrees in the side view, in other words, the side view is taken at a position where the azimuth center is (27, 45, 0), and one image is on the left and the other image is on the right. The output is the best one top view image and its corresponding 3 best side view images.

[0251] Finally, for genus / species / sex classification, we used the top view and the three best side view images obtained.

[0252] The second approach uses deep learning.

[0253] For a given top-view image, the best matching angle-view [back / side] and pose / posture-view with unique features can be found.

[0254] The data is initially prepared for training.

[0255] For each mosquito vector species, sparse N top-view images with various poses / stances / viewpoints and specific body parts are acquired.

[0256] Each acquired index image is labeled by the following parameters:

[0257] Each image is labeled with dorsal or lateral or ventral angle view.

[0258] Each image is labeled with a skeletonized polyline of the body from the tail to the head / antennae / beak. This skeleton represents the angular pose. For example, if the tail is at coordinates (5, 4) and the head is at coordinates (23, 20), then the polyline is at {(5, 4), (23, 20))

[0259] Each image is labeled with important organ morphological features, such as wings / legs / abdomen / thorax / head, etc.

[0260] Each image is labeled with optimal side view parameters and genus / species / sex. N = at least 10,000 labeled images.

[0261] The algorithm is learned to extract body pose-posture [3D shape] from top-view images [2D images].

[0262] The real-time process predicts the best pose-posture view and extracts the best side view from the labeled parameters. A side view image is acquired. Finally, for genus / species / sex classification, one top view and the best side view image obtained are used.

[0263] Reference now Fig.17A and 17B , which shows pooling, i.e. placing the mosquito in a vial for later inspection or testing, e.g., for a virus test. Pooling is an option, and can be omitted if not desired, so that once the mosquito is identified, it can simply be dumped, e.g., by rotating or tilting the surface on which the mosquito rests, so that it falls, possibly guided by air blowing from the side, or a suction tube can pick up the mosquito and remove it to a common removal location that is the same for all.

[0264] However, when a collection is required, then in order to realize an automatic collection process, a plurality of vials 140.1 ... 140.n are located together at position 142, so that a moving arm that picks up a mosquito from an imaging position can bring the mosquito to position 142. Precise control, for example using a Z-axis motor (e.g. a pneumatic motor using an air piston or an electric piston), enables movement toward the vials and placing the mosquitoes into the vials by blowing them inside the vials and / or by closing the suction. In an embodiment, all axis movements can be realized by a multi-axis robot (e.g. an articulated commercial robot).

[0265] Each vial 140.1...140.n can be assigned a unique identifier, for example using a barcode, and mosquitoes from the same species can be transferred to the corresponding vials so that by the end of the pooling process, there is a collection of vials, each vial having mosquitoes of the same species (e.g., 50 mosquitoes per vial), or the same species from the same location.

[0266] As mentioned above, there may also be another common vial into which all mosquitoes that are not automatically identified can be transferred for additional analysis later by other means.

[0267] When pooling is required, the machine or at least the vials can be placed in a cooling area, ensuring a preferred temperature of 4 degrees Celsius, more generally, a temperature close to but above freezing to allow smooth operation of electronic components, robotic arms, motors, cameras, etc.) to ensure that the samples remain cold for virus testing. In one embodiment, the entire solution including the separator, transfer conveyor element, imaging station and vials can be kept in a closure under cold conditions, preferably with an active cooling system. Alternatively, only the part of the system that keeps the insects on or in it is kept cold, such as the separator unit, the conveyor system, the imaging station or only the vials.

[0268] Holding the insects and moving them from the identification station to the corresponding vials and moving them to the corresponding vials can be achieved by different methods that may occur to the skilled person.

[0269] One method is to use a pick and place device as described above, where a pressure differential is applied across the net at the end of a suction tube to hold the insects while a robotic arm moves the tube from a pick position to a place position.

[0270] When the tube reaches the coordinates of the corresponding vial, it can move down, and when it is above the target vial, it can close the pressure difference, causing the insect to fall from the net. Optionally, the pressure difference can be reversed, and the air flow can blow the insect toward and into the vial.

[0271] The tube holding the insect can be lowered to approximately 1 mm above the vial opening, or until it is almost flush with the opening surface. In an embodiment, the diameter of the tube is smaller than the vial opening, in which case the tube can enter the vial and then cause the insect to fall out.

[0272] In other embodiments, instead of or in conjunction with air flow, a mechanical gripper such as a mechanical tweezer can grip the insect and, when the gripper is positioned over the vial, open the gripper and allow the insect to fall into the vial. This opening and guiding of the mechanical gripper is controlled and managed by the system controller.

[0273] The entire system or portions thereof may be maintained at cryogenic conditions to support immobilization of live organisms as well as support storage conditions for dead organisms, thereby preserving any potential viruses therein.

[0274] Reference now Figures 18A-18H , which shows an alternative embodiment of a suction tube. As explained, the insects are transferred to the vial, and the tube can use an adapter (also referred to herein as a connector) between the vial and the suction pick and place unit (referred to herein as the tube). The adapter can be part of the vial, or can be placed on top of the vial, or can be connected to the suction unit as needed. For example, the insects are dropped from a certain height into the vial, and the adapter can be connected to the tube, as will be described in more detail below.

[0275] A suction arm based pick and place tool picks up the insect, transfers it, and then places it at the target location. Insects can be delicate, and mosquitoes are fragile, and handling may require care not to crush the insect.

[0276] Reference now Fig.18A , the tube 150 connected to the air inlets 152, 154 and 156 is shown in longitudinal section. During suction, the air flow comes from the middle area of ​​the net at the bottom of the device, and the holding net 158 ​​keeps the insects in the center of the net as much as possible. The air flow during suction passes through the center 60 of the tube 150, because only the central air inlet 154 allows suction, even if all three air inlets 152, 154 and 156 are connected to the vacuum generator through air ducts. The air flow does not flow to the outside of the middle area 160, that is, does not flow through the external channel 162, because the air inlets 152, 156 on both sides have a one-way air flow valve, allowing the air flow to flow only downward to the net. Fig.18B and 18C The area surrounding the web is shown in more detail.

[0277] Thus, when a mosquito or sand fly or other insect of interest is placed in the vial, the airflow is reversed to be directed downward toward the net. The three air inlets 152, 154 and 156 are used together, possibly connected to the same outlet from a vacuum source. The air flows downward through all 3 inlets and blows the insects off the net and into the vial.

[0278] Connector 164 facilitates placement of the suction unit to contact vial 168 when blowing air, and includes vent holes 166 to allow air to escape. The diameter of the vent holes is smaller than the smallest insect that may be of interest, e.g., 0.5 mm diameter, 1 mm diameter, 2 mm diameter, etc. In an embodiment, all holes have the same diameter, e.g., they are all 0.5 mm).

[0279] If connector 164 is not used, if the tube is placed flush with the vial, then when air is blown into the vial it has nowhere to go and will cause undesirable turbulence interfering with successful placement of the insect. Therefore, the tube is positioned some distance above the vial, or the tube touches, requiring connector 164. The connector is preferably resistant to static charge.

[0280] Reference now Fig.19A , which shows a screen 170 for manually tagging insects. If automatic tagging does not produce results, the insect can be submitted to a human operator who receives the image of the insect and manually identifies the species, sex, etc. and tags it accordingly. In an embodiment, the reference is made while the insect is still at the imaging station, and the software can allow the human operator to visually explore the insect by rotating the insect or rotating the camera around the insect (or both).

[0281] The screen contains a current image 172 of the insect, and arrows below 174, 176 to enable the human operator to move and take the next image at intervals that may be pre-selected or that may be selected by the operator.

[0282] In further embodiments, all images are recorded, such as on a cloud service, which may enable an operator to gain later access to each image and alter the software decision based on insect species, or perform manual tagging as described above. For example, a human may wish to review any decisions with a confidence level below a certain threshold, or any species identification that appears unusual in some way.

[0283] Fig.19BA screen 180 is shown with a plurality of images 182.1 ... 182.n from a sequence of insects rotating relative to a camera, and an operator can use the images to manually identify the species. A drop-down list 184 of potential selections may be provided. If unsure or the operator needs help or a second opinion in making an identification, a "Ask an Expert" or similarly named button 186 may be used to allow for later identification. The expert is able to categorize and filter the images to display only those images marked with button 186. A radio button group 188 is provided for gender classification.

[0284] In embodiments, identification of insects may use images from parts of the spectrum other than light, or via sonic or ultrasonic sensors.

[0285] In embodiments, at the identification station, imaging may include hyperspectral imaging, reading reflections from a laser beam fired toward the insect, using reagents that react to specific materials.

[0286] Once the insect or other material is identified, it can then be transferred to the corresponding vial.

[0287] In the case where it is already known that a batch of insects is of the same species, for example due to the type of trap, all that needs to be done is to place them into vials. The operator places or deposits the batch of insects onto a separator unit for separation into individuals, and then the robot tube takes out each insect and transfers them to the corresponding vials using the pick-and-place suction unit described above. In this case, the method comprises separating the insects, positioning the individual insects on a separation surface, transferring the separated insects into a storage compartment by suction, and then subsequently dropping or blowing them into a storage area.

[0288] It should be noted that the process can be used for different species of insects, such as mosquitoes, sand flies, fruit flies, and in particular mixtures of insects, where the goal is, for example, to study a specific ecosystem.

[0289] Reference now Fig. 20 , which shows the mapping phase. Insects may be collected from multiple traps at different locations, and information is needed not only on the total number of insects, but also on their distribution.

[0290] Mapping may be necessary, but is not mandatory, and can occur in parallel with pooling so that different vials are used to collect insects from different locations. Thus, information can be obtained that the distribution of disease-carrying organisms is confined to a certain part of the insect's geographic distribution.

[0291] Once the mosquitoes from a given trap are identified and counted, the location data of the trap can be included in the insect count because they all come from the same trap whose physical coordinates are known. Therefore, the map 190 can be updated with information showing the number of insects in each trap. For example, a symbol 192 can visually indicate the number of insects of a given species at the location of the trap. Separate maps can be provided for different species, or different symbols can indicate different species, and the symbols can be colored or covered with percentages to indicate the presence or proportion of disease carriers. Alternatively, the report can provide the number and geographic location of the insects. The database can store images of the insects and their locations.

[0292] A report or map may show the evolution over time of the insect population at a particular trap or over a geographical area. The evolution over time may allow predictions to be made, for example, regarding the spread of an infectious disease.

[0293] The operator may enter the coordinates of the trap as each batch is emptied into the separator, or the separator may be present in situ and its GPS coordinates simply checked each time it is filled, or the location may be entered in any other convenient manner. Additional data may be entered regarding environmental conditions at the trap, such as wind speed when the batch was obtained, altitude of the trap location, and any other information the user deems relevant.

[0294] Thus, the operator can automatically update the information on the geographical area represented on the map by bringing in updated count data. As each trap is processed, data can automatically flow out of the separation and identification machine.

[0295] Isolation of insects:

[0296] Reference now Fig.21A ...21D, which shows another embodiment of a separator. It is described above that the function of separating insects is to vibrate the surface on which the insects are located. The vibration is synchronized with the rotation, so that the machine rotates and then moves back and forth on a vertical axis, causing vibration. The vibration can be performed on one axis, but it can also be performed on two axes to effectively separate the insects. One or more outlet holes allow the insects to be discharged from the separator one by one. In an embodiment, the opening at its outlet hole can have a phase shape, which means that the diameter at the outside of the hole is larger than the diameter at the inside to prevent blockage as described above.

[0297] In an embodiment, a typical inlet diameter may be 4 mm, wherein the outlet diameter is preferably 6 mm. For insects of different sizes, other diameters are possible.

[0298] There may be one or more such outlet holes on the vibrating surface. As mentioned above, the surface may be flat, or may be conical or circular, as in the case of the rotating storage compartment described above.

[0299] The surface itself may include pores, or the surface may be porous.

[0300] By vibrating the surface, the insects are separated and by having exit holes on at least a portion of the vibrating surface, the separated insects are allowed to exit through the surface.

[0301] Adapters with different outlet sizes can be connected to the outlet holes and quickly replaced in order to use the same outlet holes but make the machine suitable for smaller insects.

[0302] When several insects of different sizes are found together in a trap, an adapter may first be initially applied to obtain the smaller insects, and then the adapter may be manually changed to provide a larger exit hole for the larger insects. This process will typically be repeated as necessary until all or most of the insects are separated. The remaining insects may be manually identified.

[0303] like Fig.21A 21D, a vibrating surface 200 of a separator according to an embodiment of the present invention has an exit hole 202 through which insects fall when the surface is vibrated by a vibrating motor 204. The vibrating surface is the floor of a housing 206 having a cover 208. Three vibrating motors 204, 210, 212 with eccentric weights 214 may be provided to apply vibrations on each axis.

[0304] Reference now Fig. 22 , shows an embodiment of a situation where it is not necessary to identify the exact mosquito species. In a variation, only a single image, such as a top view image, is sufficient to perform the required identification. For example, it may only be necessary to determine the total number of mosquitoes without identifying species or gender.

[0305] In the device 220, a sensor (e.g., a top camera 222) is positioned to observe the insects after they are discharged from the separator. The camera 222 can be located above a conveyable surface 224 that conveys the discharged insects away from the separator 226, or the camera can be located below the discharge hole 228 of the separator 226, so that the camera 222 can count the insects that fall in flight. The camera can be replaced by a simple optical detector (e.g., an IR detector) that only counts the number of falling objects, regardless of whether they are mosquitoes. However, the camera can provide an image to the computer vision software to verify what has fallen.

[0306] It should be noted that the transfer of the individual elements towards the identification position can also be achieved by introducing an articulated robotic unit, or by moving the xy (and possibly z) motorized axes of the pipette, in either case with the pipette and guided by a control system.

[0307] Regarding the primary separator embodiment, in the case of a rotating element, the specialist may use multiple rotating elements on a single conveyor system and instruct the operator to place batches of insects of different sizes in each rotating drum, each rotating drum having openings of different sizes.

[0308] In an embodiment, the separator unit may be mounted within a conventional mosquito trap to be filled with insects and provide the insects for real-time imaging.

[0309] In an embodiment, a pick and place device can be used directly on a batch of insects from a trap. The tube is sized to pick up the insects individually or in very small quantities and transfer them to the imaging location. If a large number of insects are inadvertently picked up by the pick and place device, the insects may pile up and further attempts to pick up individual insects may be made.

[0310] Reference now Figures 23A-23H , shows an embodiment in which a mechanical separator 230 comprises a hopper 232 for pouring insects into an insect compartment 234. The compartment is rotated by a motor 236 and vibrated by a motor 238, and the insects are released one by one through release holes 240 onto a conveyor belt 242, where the insects are conveyed to an imaging area 244, where they are imaged by a top camera 246 and a rotating side camera 248, which moves around the insects as the top camera is positioned to obtain images at different angles. A light 249 illuminates the imaging area 244. The pick and place device consists of a suction head 250 and a robotic arm 252, and moves the insects from the imaging area 244 to a vial 254. An inlet 256 blows air to free blocked holes by blowing captured insects back into the separator interior. Fig.23D , 23G 23H shows a single motor 258 that rotates and vibrates the storage compartment 234. The rotating side camera 248 is supported by a holding structure 260, which may also support the top camera 246. A frame 262 may hold the separator, and a guide 264 may be provided to observe the dislodged insects reaching the conveyor belt 242. Fig.23E Motors 266 , 268 , and 270 are shown for the three axes of the robotic arm 252 .

[0311] The combination of the mechanical separator 230 and the conveyor belt 242 provides the ability to separate individual insects from each other because the individual insects are discharged from the discharge holes as the rotating unit rotates. As they are discharged, they have their own inertia due to the rotation, which causes each insect to fall on the collection surface (e.g., conveyor belt 242) at a different location and at a different time, supporting the physical separation between the insects as they travel toward the imaging station.

[0312] The level of confidence in recognition has been mentioned above. When an image is sent to a neural network model, there is always an answer. However, there may be no body parts in the image that are relevant to what we are looking for, so that the result provided from the model (expressed as the probability of identifying a specific species in the image) should be very low. However, sometimes there are errors and the model can provide high probabilities even for images that do not have the relevant body parts in the image.

[0313] In order to improve speed, computation time and accuracy, one embodiment can send only images estimated to be from an angle relative to the insect for detection to include relevant body parts with distinguishing features. One method of doing this is to use a top view camera to identify the posture of the insect. Therefore, only images of angles corresponding to angles that may display relevant information are stored. One method of doing this is to rotate the camera (or insect) and take multiple images as usual, but then only use specific images associated with specific angles relative to the insect's body axis.

[0314] In another embodiment, 3D imaging can be applied. In this case, the camera can not only rotate around the insect in the same plane and at the same angle relative to the surface the insect is placed on, but it can also rotate between planes and move up and down, covering the 3D surface around the insect.

[0315] In one embodiment, the process of mechanically obtaining an image is as follows: Move the conveyor belt 242 at a high speed A. Activate the mechanical separator 230 by rotating and or vibrating the drum for X seconds. Once the separator stops, the conveyor belt 242 moves toward the imaging position 244 at a low speed B. This is because separation may require very high speeds at the conveyor belt, making it advantageous to interrupt and work in a pulsed mode. The mosquitoes are now separated along, for example, about 10 cm of the conveyor belt. The length of the conveyor belt is now slowly moved, scanned, and after completion, enter the next cycle, starting with high-speed separation to produce another batch and produce another 10 cm long area of ​​separated mosquitoes for scanning.

[0316] When the camera identifies an insect in the top view and sends an image to identify the insect in the image, the conveyor belt stops, the insect is located on the conveyor belt based on the coordinates found by the top camera, and an image is taken by the side camera to classify the insect.

[0317] Optionally, a high-resolution camera can take a top view over the indicated coordinates, and a side-view camera can rotate in a circle centered at the given coordinates or at specific coordinates associated with the insect (eg, the insect's head).

[0318] Images are taken at different angles between the camera and the insect (e.g., by rotating the camera relative to the insect). At each angle, the camera can take multiple images at different focus planes so that different body parts are focused at each angle, or, in the case of a top-view camera, the exact coordinates of the feature to be focused on can be taken into account when calculating the coordinates. Thus, when the camera takes an image, the focus may be at a position calculated from the pose. Changing the focus can be controlled using, for example, a liquid lens.

[0319] The conveyor belt is then restarted at a lower speed until all insects on the conveyor belt have been identified, and additional batches of insects are then separated and placed on the conveyor belt.

[0320] The insects are usually dead by the time they are imaged. However, the conveyor belt can be cooled regardless of whether the insects are alive or dead, making virus testing, etc., still feasible.

[0321] The pulse process described above provides the ability to first separate the insects on the conveyor belt so that they are sufficiently separated, i.e., on average, not touching each other, and then switch to an imaging mode where the separator stops and the conveyor belt moves slowly to identify the insects on it, and then potentially stops completely to allow images from multiple positions around it. Alternatively, the conveyor belt can have continuous slow motion and have multiple camera mounts around positions above the conveyor belt so that multiple insects can be identified in parallel.

[0322] Reference now Fig.24 , which is a general diagram showing interesting features about a mosquito that can become relevant when finding the species or sex of the mosquito. These include the head, antennae, palps, proboscis, vertex, thorax, abdomen, wings and their pattern, legs and leg parts, and cercariae.

[0323] Now refer to Table 1, which shows a decision tree for finding the species of a mosquito. The neural network model receives an image where a body part is supposed to be located and then applies this decision scheme to decide which species it is. Typically, a decision tree tells the network which body part to look for. It either identifies or does not identify the body part in the image and uses these identifications to navigate through the decision tree to the decision.

[0324] Other decision schemes are possible. In addition, a neural network can be trained to identify species based on a collection of images of each body part it receives.

[0325]

[0326] Table 1 - Decision tree for mosquito species

[0327] Fig.25A An example showing how a decision tree can be used to identify samples of Culex Pipiens.

[0328] In dorsal view of the hind femur, a dark stripe is present that extends almost 100% of the segment length.

[0329] The lateral chest is pale and usually lacks small white scaly spots, but sometimes has such spots.

[0330] Fig.25B An example showing how a decision tree can be used to identify Aedes Caspius.

[0331] There are approximately 3,500 different species of mosquitoes in the world, but in each specific place (country, region of country, city, neighborhood) there is a subset of species that are commonly found, so during system calibration, and as an important upfront process, the input model needs a collection of species to compare it to the picture. Thus, the system can be calibrated based on the species expected in a trap or batch.

[0332] In one embodiment, the separator may be omitted and individual insects may be placed on an imaging stage by a human operator for imaging in the manner described above.Such a process is particularly advantageous for initial training of a neural network.

[0333] Variations on the process discussed above include fusing the classification of body parts and using identification of emitted fluorescence to obtain classification of species.

[0334] A light source is provided that emits a wavelength that can stimulate a fluorescent response from an insect at an imaging station. Specific body parts of the insect are excited and fluoresce. In the process, imaging is initially performed to obtain a top view using visible light to determine whether an insect is present in the image. If an insect is present, the top view image can be used to identify at least one body part, such as the head, and give its coordinates.

[0335] Imaging is then performed, again using a top view of light with an appropriate excitation wavelength to excite the body part so that it emits light at its own fluorescing wavelength. The fluorescing wavelength can be detected on a camera, for example with a suitable filter.

[0336] The coordinates of the fluorescence identified in the image are compared to the location of the body part identified in the top camera view to correlate and identify that a specific body part is emitting light. In some cases, if the insect emits a specific wavelength, then the classification of the insect is concluded, but in other cases, it may be necessary to count how many insects emit a specific wavelength from a specific body part. Thus, for example, 10 male mosquitoes of the Aedes albopictus type with red eyes may be identified, and 3 female mosquitoes of the Aedes albopictus type with green fluorescence from the head may be identified.

[0337] The method may optionally use a side camera to provide support when identifying insect species using the method described above for using a side camera, and in fact the side camera may be used to perform classification of the body part that emits light.

[0338] Another option is to completely separate the identification using visible light and the identification using fluorescence. This separation can optimize the process so the insect can be imaged using a rotating camera as described above, and then can be imaged again using the excitation wavelength used for fluorescence excitation.

[0339] The process can use the above method to separate the insects and drop them on a conveyor belt, first taking a top view image, then a side view image, and sorting using normal light. The conveyor belt is then moved to a second imaging fluorescence station. At the second imaging station, the insects are imaged using visible light to locate specific body parts and their coordinates, such as the head. An excitation frequency is then used, and the fluorescence wavelength is detected by a camera. The coordinates of the detected fluorescence are compared to the coordinates of the body part that is expected to produce the fluorescence.

[0340] Different body parts may emit different wavelengths, and classification can then classify insects based on both their species and the wavelengths they emit.

[0341] Insects may also be optionally picked up and placed into vials or storage containers based on fluorescence identification along with species, for example, Aedes aegypti with orange fluorescent heads may be placed in one vial, and Aedes aegypti with green fluorescent abdomens may be placed in another different vial.

[0342] Reference now Fig.26, which is a simplified schematic diagram showing an apparatus for separating insects of different sizes simultaneously. Apparatus 260 is a variation of a separator in which the outlet holes are in frame 262 and are of different sizes. Outlet holes 264 and 266 are of different sizes and a mesh 268 moves between the different holes. The bottom plate vibrates and moves the mesh between the holes of different sizes so that at part of the time the larger hole is in operation and at other times the smaller hole is in operation. More than two holes of different sizes may be used.

[0343] It is expected that during the life of a patent from this application, many related imaging and neural network technologies will be developed, and the scope of these and other terms herein is intended to include all such a priori new technologies.

[0344] The terms "comprise", "comprising", "include", "having" and their conjugations mean "including, but not limited to".

[0345] The term "consisting of" means "including and limited to."

[0346] As used herein, the singular forms a, an, and the include plural references unless the context clearly dictates otherwise.

[0347] It should be understood that, for the sake of clarity, certain features of the invention described in the context of separate embodiments may also be provided in combination in a single embodiment, and the description of the invention should be interpreted as if such embodiments were explicitly set forth herein. Conversely, for the sake of brevity, various features of the invention described in the context of a single embodiment may also be provided individually or in any suitable sub-combination, or may be suitable as modifications to any other described embodiment of the invention, and the description of the invention should be interpreted as if such individual embodiments, sub-combinations, and modified embodiments were explicitly set forth herein. Certain features described in the context of various embodiments are not considered essential features of those embodiments unless the embodiments do not function without those elements.

[0348] Although the present invention has been described in conjunction with specific embodiments of the present invention, it is apparent that many alternatives, modifications and variations will be apparent to those skilled in the art. Therefore, it is intended to cover all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims.

[0349] All disclosures, patents and patent applications mentioned in this specification are incorporated herein in their entirety by reference in their entirety, to the extent that each individual disclosure, patent or patent application is specifically and individually indicated to be incorporated herein by reference. In addition, the citation or identification of any reference in this application should not be construed as an admission that this reference can be used as prior art of the present invention. With respect to the section headings used, they should not be construed as necessary limitations. In addition, any priority document of the present application is incorporated herein by reference in its entirety.

Claims

1. A method for separating a batch of insects from a trap into individual insects, comprising: pouring the batch of insects into a container having at least one hole sized for a single insect; moving the container in a shaking motion to shake the insects therein so that individual insects exit through the holes onto a collection surface, thereby providing separated insects onto the collection surface, Therein, the container comprises an upper cone and a lower cone, the upper cone and the lower cone meet at a common base, the base provides a maximum circumference, and the at least one hole is located at the maximum circumference.

2. The method of claim 1, wherein the container comprises a floor, the movement comprises vibration, and the at least one hole is in the floor.

3. The method of claim 1, wherein the container comprises a circumference, the at least one hole is located in the circumference, and the movement comprises rotation. The method of claim 3 , wherein the motion further comprises vibration. The method according to claim 4 , wherein the rotating and the vibrating are alternated in a cycle.

6. A method according to any one of claims 1 to 5, comprising pouring a batch of insects into the container through a funnel.

7. The method according to any one of claims 1 to 5, wherein the collection surface is a moving surface.

8. An apparatus for separating insects from a batch of insects into individuals, the apparatus comprising a container for the batch of insects, the container being motorized to provide motion to the container so as to shake the insects in a shaking motion, the apparatus having at least one hole and having guide means for guiding the exiting insects from the at least one hole to a collecting surface, the hole being sized for individual insects so that individual insects from the batch of insects can be pushed out of the hole when pushed towards the hole by the shaking motion, in, The container comprises an upper cone and a lower cone, the upper cone and the lower cone meeting at a common base, the base providing a maximum circumference, and the at least one hole is located at the maximum circumference.

9. The apparatus of claim 8, wherein the container has a bottom plate, the at least one hole is located in the bottom plate, and the shaking motion comprises a vibrating motion.

10. The apparatus of claim 8, wherein the container has a circumference, the motion comprises rotation about an axis perpendicular to the circumference, and the at least one hole is located in the circumference.

11. The apparatus of claim 10, wherein the motion further comprises vibration in at least one axis.

12. The apparatus of claim 11, wherein the motion comprises vibrations in three axes.

13. The apparatus according to any one of claims 10 to 12, wherein the at least one hole comprises an inner side facing the interior of the container and an outer side facing the exterior of the container, and a diameter at the inner side is smaller than a diameter at the outer side.

14. Apparatus according to any one of claims 8 to 12, comprising a funnel for pouring the batch of insects from the trap into the container.

15. Apparatus as claimed in any one of claims 8 to 12, comprising a motor having an eccentric weight to provide vibration.

16. A method of picking up insects and placing them on a collection surface where separated insects provided by the method of any one of claims 1 to 7 have arrived, the method comprising: imaging the collection surface from above; determining from said imaging the presence of said insect on said surface for picking up; determining from the imaging a current position of the insect on the surface as a pick-up position; Using the robotic arm, moving the pick tool to a position above the pick location; lowering the pick-up tool to the pick-up position; operating suction to pick up the insects from the pick-up location into the pick-up tool; using the robotic arm to move the pick-up tool with the insect to a position above a deposition location; as well as The suction is removed to deposit the insect, wherein one of the pick-up location and the depositing location is an identification location for imaging the insect for identification, wherein the pick-up tool includes a porous surface in a tube leading to a vacuum source, and the insect is held on the porous surface by the suction.

17. The method of claim 16, wherein the identified location is the pick-up location and the identification performed at the identified location defines the deposit location.

18. A method according to claim 16 or claim 17, comprising switching from the suction to blowing air at the deposition location to deposit the insects.

19. The method of claim 18, wherein the pick-up tool has a central air duct and a peripheral air duct, the suction is applied through the central air duct, and the blowing is provided by the central air duct and the peripheral air duct.

20. An insect picking tool for use in the method according to any one of claims 16 to 19, the picking tool comprising a hollow tube having a first end and a second end, the tube being connected to an air pressure source at the first end and having a porous surface near the second end, the tool further having a robotic arm for positioning the tool in three-dimensional space, the tool being configured to work with an imaging system to position itself above coordinates provided by the imaging system as the position of the insect on the surface, the tool being configured to lower itself above the coordinates and apply suction to draw the insect against the porous surface, thereby picking up the insect.

21. The pick-up tool of claim 20, wherein the porous surface is a mesh, the mesh being located a distance from the second end equal to the thickness of an insect.

22. A picking tool as claimed in claim 20 or claim 21 having a central air duct and peripheral air ducts, the suction being applied through the central air duct to position the picked up insects centrally on the web.

23. The pick-up tool of claim 22, configured to turn off the suction when reaching a destination, thereby depositing the insects at a placement location.

24. The picking tool according to claim 23 is configured to turn off the suction when reaching the destination and replace the suction with a blast of air, which is applied through the central air duct and the peripheral air duct to deposit the insects at the placement location.

25. A method for identifying and counting images obtained in batches from field traps, the method comprising: receiving a batch of insects from a trap; placing the batch into a separator comprising a container having at least one hole sized for a single insect; moving the container in a shaking motion to shake the insects therein, thereby individually discharging the insects from the batch; collecting said insects expelled onto a moving surface; and capturing at least one image for each of the insects on the moving surface; as well as For each individual insect found in the corresponding image, increment the count, Therein, the container comprises an upper cone and a lower cone, the upper cone and the lower cone meet at a common base, the base provides a maximum circumference, and the at least one hole is located at the maximum circumference.

26. The method of claim 25, comprising taking a series of images from different angles of each insect on the moving surface and providing the images to a neural network to identify the insect.

27. The method of claim 26, comprising using identification to define a destination for placement of the insect.

28. The method of claim 27, comprising using a first camera to locate the insect, and using a second camera to capture images from different angles around the insect.

29. The method of claim 28, comprising rotating the insect on a rotating disk to obtain the images from different angles.

30. The method of claim 28, comprising placing the second camera on a robotic arm and moving the second camera around the insect to obtain the image.

31. A method according to any one of claims 25 to 30, comprising illuminating the insect with an excitation wavelength to induce fluorescence.

32. A method according to any one of claims 28 to 30, comprising obtaining images at different depths of focus.

33. A method according to any one of claims 27 to 30, comprising identifying a posture of the insect and obtaining an image of a body part according to a position defined by the posture.

34. A method according to any one of claims 28 to 30 comprising using a decision tree to define species defining features and positioning the second camera to image a body part according to the decision tree.

35. A method according to any one of claims 25 to 30, comprising operating the separator to discharge separated insects onto a section of a conveyor belt, then stopping the separator and identifying insects on the conveyor belt, and repeating said operation and identification.

36. A method according to any one of claims 25 to 30, wherein the separator is an apparatus according to any one of claims 8 to 16.

37. A method according to any one of claims 25 to 30 comprising picking up the insect using a picking tool according to any one of claims 20 to 24 and placing the insect.

38. The method of claim 37, comprising using an identifier to define a destination for placement of the respective insect.

39. A method according to any one of claims 25 to 30, comprising obtaining a plurality of batches from a plurality of traps, each trap having a different location, the method comprising using the insect counts from the respective traps to generate or update a report or geographic map of insect distribution.

40. A method according to any one of claims 25 to 30, wherein the identification is based on an insect database of insects expected in the area of ​​the trap.

41. A method according to any one of claims 26 to 30, wherein the identification includes leaving some insects unclassified due to being unidentified, or identified as being below a threshold level of certainty, the method including forwarding images of the unclassified insects for manual identification by an operator.

42. A method of automatically identifying the species or sex of insects on a moving surface onto which insects are discharged by providing discharge using the method of any one of claims 28 to 30, the method comprising: capturing a first image of the insect from above to identify a direction of the insect; and using the first image to find at least one location for imaging a first given body part, and sending the second camera to the location to capture a second image; as well as The second camera is sent to additional locations identified by the first camera to image additional body parts until sufficient images are available to identify the insect.

43. The method of claim 42, wherein the first image is taken using a first camera located overhead and the second image is taken from the second camera on a robotic arm.

44. The method of claim 42, wherein the first image is taken using a camera located overhead and the second image is taken from a camera on orbit.

45. The method of claim 42, wherein the first image is taken using an overhead camera and the second image is taken from an overhead camera or a side camera.

46. ​​A method of separating a batch of insects from a trap into individual insects, comprising: pouring the batch of insects into a container having at least one hole sized for a single insect; moving the container in a rotational motion, the container having a circumference, the rotational motion thereby moving the insects therein so that individual insects exit through the apertures onto a collection surface, thereby providing separated insects onto the collection surface, Therein, the container comprises an upper cone and a lower cone, the upper cone and the lower cone meet at a common base, the base provides a maximum circumference, and the at least one hole is located at the maximum circumference.

Images