Pest detection system and method

Abstract

The subject matter of the present invention relates to a pest detection system employing infrared emitters and detectors for searching for changes in incident light indicative of various pest wing oscillations. The emitter and the detectors are placed on one or more printed circuit boards such that infrared light projected from the emitter can be received by the one or more detectors. Depending on the signal generated by the detector, the system of the subject matter of the invention can determine whether a pest flies between the emitter and the detector. To save power, the transmitter, detector, or both may be driven by pulse width modulation. Methods of the subject matter relate to determining the presence of pests by signal filtering and interpretation.

Description

[0001] This application claims priority to U.S. Patent Application Serial No. 17 / 725,753, filed on April 25, 2022. All extrinsic materials noted in this application are incorporated herein by reference in their entirety. Technical Field

[0002] The field of the invention is light-based pest detection systems and methods. Background Art

[0003] The background description includes information that may be helpful in understanding the present invention. This does not constitute an admission that any of the information provided in this application is prior art or relevant to the present invention, nor does it constitute an admission that any publication specifically or implicitly referenced is prior art.

[0004] Agriculture is both one of humanity’s oldest and most important industries. Humankind has historically sought to reduce the impact of pests and diseases in order to improve the health and nutritional content of food and increase crop yields. Pest detection is critical when it comes to increasing crop yields. Pests are often the greatest threat to successful crop production, and early pest detection can reduce yield losses by up to 20-40% according to some estimates, depending on crop type and growth stage.

[0005] Pest detection in the past required farmers to walk the fields to visually inspect for pests that damage their crops. But with the growth of industrialization, farmland has increased to the point where walking around the farm is impractical. In fact, even driving around the farm is impractical because it requires farmers to spend an inordinate amount of time on a menial task. As a result, efforts have been made to develop automated pest detection systems.

[0006] In some efforts, visual detection systems have been developed. However, these systems are not perfect. In visual detection systems, a camera is used to periodically take pictures or record videos of pests attracted, for example, to a trap. However, many of these systems still require a human operator to view the images for pest detection, requiring a human user to, for example, count the number of pests caught in a trap. Acoustic pest detection has also been implemented, in which an electromechanical listening device is used to determine when pests exceed a certain threshold by listening for increases in noise levels associated with various pests.

[0007] However, these systems have drawbacks in terms of power consumption, accuracy, and the absence of the need for manual inspection. Therefore, it has not been realized that pest detection can be achieved using infrared light through a system designed to automatically detect the wing beat frequencies associated with flying pests. Summary of the invention

[0008] The present invention provides devices, systems and methods for pest detection using infrared light. In one aspect of the subject matter of the present invention, a pest detection system includes: a first printed circuit board (PCB) portion and a second PCB portion, wherein there is a space between the first PCB portion and the second PCB portion; at least one emitter disposed on the first PCB portion; at least one detector disposed on the second PCB portion, wherein the at least one emitter is oriented to face the at least one detector, and wherein the at least one detector is oriented to face the at least one emitter; a pulse width modulation controller configured to drive at least one of the at least one emitter and the at least one detector.

[0009] In some embodiments, the first PCB portion and the second PCB portion are part of a single PCB, and in some embodiments, the single PCB is V-shaped, having a first arm and a second arm, wherein the first arm includes the first PCB portion and the second arm includes the second PCB portion. The single PCB is annular, having a first side and a second side, wherein the first side includes the first PCB portion and the second side includes the second PCB portion. In some embodiments, the pulse width modulation controller operates at a duty cycle of 10% to 90% and a frequency greater than 240Hz. At least one detector may include a photodiode, and at least one emitter may be configured to project infrared light (e.g., a wavelength of 700nm-1400nm). In some embodiments, at least one detector is configured to detect infrared light (e.g., a wavelength of 700nm-1400nm).

[0010] In another aspect of the inventive subject matter, a pest detection system comprises: a first printed circuit board (PCB) portion and a second PCB portion, wherein there is a space between the first PCB portion and the second PCB portion; a wide angle emitter disposed on the first PCB portion; a group of detectors disposed on the second PCB portion, wherein the wide angle emitter is oriented to face the group of detectors, and wherein each detector in the group of detectors is oriented to face the wide angle emitter; and a pulse width modulation controller configured to drive the wide angle emitter and the detectors.

[0011] In some embodiments, the first PCB portion and the second PCB portion are part of a single PCB, and the pulse width modulation controller can be configured to operate at a duty cycle of 10% to 90% and a frequency greater than 240 Hz. Each detector in the set of detectors can be a photodiode, and the wide-angle emitter can be configured to project infrared light (e.g., wavelength 700nm-1400nm). Each detector in the set of detectors can be similarly configured to detect infrared light (e.g., wavelength 700nm-1400nm).

[0012] It should be appreciated that the disclosed subject matter provides a number of advantageous technical effects, including the ability to detect the presence of pests while eliminating false positives and with greater accuracy than existing systems.

[0013] Various objects, features, aspects and advantages of the present subject matter will become more apparent from the following detailed description of preferred embodiments and the accompanying drawings, in which like numerals represent like components. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A circular pest detection system is shown.

[0015] Figure 2 A circular pest detection system disposed within a fly trap is shown.

[0016] Figure 3 A cross-sectional view of a fly trap is shown.

[0017] Figure 4 A V-shaped pest detection system is shown.

[0018] Figure 5 A V-shaped housing of a pest detection system is shown.

[0019] Figure 6 Components of the housing are shown.

[0020] Figure 7 A reflective pest detection system is shown.

[0021] Figure 8 A perspective view of a portion of a reflective pest detection system is shown.

[0022] Fig. 9 A wide angle emitter is shown.

[0023] Fig.10 A pest detection system implementing a wide angle transmitter is shown.

[0024] Fig.11 The relative radiation intensity of a wide-angle emitter at different angles is shown.

[0025] Fig.12 is a flow chart of how pests can be detected.

[0026] Fig.13 An example signal is shown with a threshold for detecting signal amplitude.

[0027] Fig.14 A schematic diagram of the onboard signal classifier is shown.

[0028] Fig.15 shows how a signal waveform can be measured.

[0029] Fig.16 An example signal with several measured waveforms is shown.

[0030] Fig.17 A schematic diagram showing how pests are detected by the system of the subject invention. DETAILED DESCRIPTION

[0031] The following discussion provides exemplary embodiments of the subject matter of the present invention. Although each embodiment represents a single combination of elements of the present invention, it is believed that the subject matter of the present invention includes all possible combinations of disclosed elements. Therefore, if one embodiment includes elements A, B and C, and a second embodiment includes elements B and D, then even if not clearly disclosed, it is believed that the subject matter of the present invention includes other remaining combinations of A, B, C or D.

[0032] As described in the specification of this application and the following claims, unless the context clearly dictates otherwise, the meanings of "a", "an" and "the" include plural forms. In addition, as described in the specification of this application, unless the context clearly dictates otherwise, the meaning of "in..." includes "in..." and "on...".

[0033] Furthermore, as described in this application, unless the context dictates otherwise, the term "connected" includes direct connection (where two elements connected to each other are in contact with each other) and indirect connection (where at least one other element is located between the two elements). Therefore, the terms "connected" and "connected with..." are synonymous.

[0034] In some embodiments, the numbers representing the number of components, properties (such as concentration, reaction conditions, etc.) used to describe and claim certain embodiments of the present invention should be understood to be modified by the term "about" in some cases. Therefore, in some embodiments, the numerical parameters proposed in the specification and the appended claims are approximate values, which can be varied according to the desired properties sought to be obtained in a specific embodiment. In some embodiments, the numerical parameters should be interpreted according to the number of significant figures reported and by applying common rounding techniques. Although the numerical ranges and parameters of the wider range of some embodiments of the present invention are approximate values, the numerical values ​​described in the specific examples are reported as accurately as possible. The numerical values ​​given in some embodiments of the present invention may include some errors, which are necessarily caused by the standard deviations found in their respective test measurements. In addition, unless the context otherwise provides, all ranges proposed in this application should be interpreted as including their endpoints, and open ranges should be interpreted as only including commercial practical values. Similarly, unless the context otherwise provides, all numerical lists should be deemed to include intermediate values.

[0035] It should be noted that any language related to a computer should be understood to include any suitable combination of computing devices, including servers, interfaces, systems, databases, agents, peers, engines, controllers, or other types of computing devices that operate alone or in combination. It should be understood that computing devices include processors configured to execute software instructions stored on tangible, non-transient computer-readable storage media (e.g., hard disk drives, solid-state drives, RAM, flash memory, ROM, etc.). Software instructions preferably configure computing devices to provide roles, responsibilities, or other functions, as discussed below with respect to disclosed devices. In particularly preferred embodiments, various servers, systems, databases, or interfaces use standardized protocols or algorithms that may be based on HTTP, HTTPS, AES, public-private key exchange, web service API, known financial transaction protocols, or other electronic information exchange methods to exchange data. Data exchange is preferably carried out through packet switching networks, the Internet, LAN, WAN, VPN, or other types of packet switching networks. The following description includes information that may be helpful in understanding the present invention. This does not mean that any information provided in this application is recognized as prior art or relevant to the present invention, nor does it mean that any publication specifically or implicitly cited is prior art.

[0036] Embodiments of the subject matter of the present invention generally relate to pest detection by electromagnetic radiation emission and detection. When a pest (e.g., a fly) passes between an emitter and a corresponding detector, the wing beats of the fly are detected, thereby allowing the system of the subject matter of the present invention to determine the wing beat frequency. Therefore, each system includes at least one emitter / detector pair (although different emitter to detector ratios are also considered and described below). For a variety of reasons, for the detector, a photodiode is used instead of a phototransistor. For the embodiments of the subject matter of the present invention, a photodiode provides an improved linear response, resulting in a more stable system than a system established using a phototransistor, and the power consumption of the photodiode is lower than that of a phototransistor. Therefore, compared with a phototransistor, a photodiode is more sensitive to the response of incident light, has a better linear response in a wider range of light, and can pass less current. For these reasons, the embodiments of the subject matter of the present invention implement photodiodes instead of phototransistors (although it is still envisioned that a poorer system can be established using phototransistors). In order to further save power consumption, pulse width modulation can be implemented on the emitter and detector so that they are turned on synchronously at the same time.

[0037] For the reasons mentioned above, photodiodes are more accurate than phototransistors in detecting pests of various sizes. Because of their slow response time, phototransistors are not suitable for detecting small insects, such as those with wingbeat frequencies exceeding 700 Hz.

[0038] The emitters used in the embodiments of the present subject matter include LEDs configured to emit infrared light. The infrared wavelength range considered is 700nm-1400nm. In a preferred embodiment, infrared light of about 940nm is emitted, so the spectral response range of the photodiode used as a detector can be 840nm-1100nm. The spectral response range of the photodiode of the present subject matter can be matched to the spectral response range of any emitter used, so the range can also include 700nm-1400nm. Some advantages of implementing infrared light emission include that infrared emitters consume less power than, for example, laser emitters, and visible light can cause unwanted changes in pest behavior. In addition, visible light can attract unwanted pests. In addition, visible light systems can be affected by solar radiation, which can generate noise and hinder the normal operation of the system.

[0039] The embodiments of the subject matter of the present invention can be arranged in various different ways. For example, Figure 1 A circular configuration of the pest detection system 100 is shown. A set of emitters 102 is disposed on the opposite side from a set of detectors 104, both of which are disposed on the same printed circuit board (PCB) 106. Although Figure 1 106 is shown on a single PCB, but it is contemplated that the PCB 106 may be divided into multiple sections without departing from the subject matter of the present invention. The emitters on one side are configured to project infrared light toward the detectors so that the signals generated by one or more detectors can be interpreted by a microprocessor to determine whether a pest is detected. According to the two-dimensional representation of the light propagating from each emitter, the infrared light projection is as follows: Figure 1 This visualization is for demonstration purposes only, as infrared light is invisible to humans.

[0040] The annular configuration can be used in embodiments such as fruit fly detection, and the pest detection system 100 can be placed in a funnel-based fly trap, for example. The pest detection system 100 has four transmitters and five detectors, although it can be as few as one transmitter and one detector. There is no upper limit to the number of transmitters or detectors, although design and functional considerations may result in a number between 1-20 (e.g., depending on the ring diameter, required sensitivity, etc.). Transmitters closer to the detector and higher density of transmitters and detectors can improve sensitivity, which is very important for some pests such as small insects. By minimizing the blind spots that exist (e.g., pests can pass between the transmitter and the detector but not through the blind spots of the emission cone of one of the transmitters), sensitivity can be improved. Since the transmitter and the detector are physically closer (e.g., 3-10cm, preferably about 4.5cm), signal integrity remains very high (e.g., compared with shorter distances, signal strength drops more at longer distances). Gain and filter adjustments can be made from the default design to be more sensitive to smaller insect sizes and higher frequencies. In addition, since the transmitter is closer to the detector, the transmitter current can be reduced.

[0041] For example, filtering can be done in the frequency domain by adding a bandpass filter. The bandpass filter has a lower limit and an upper limit. The lower limit can range from 50-300 Hz and the upper limit can range from 1000-2000 Hz. In a preferred embodiment, the lower limit is about 120 Hz and the upper limit is about 1500 Hz (+ / - 10% for each limit). The wingbeats of smaller insects are usually within this range. Additional filtering in the time domain can be used to eliminate other noise.

[0042] There is a correlation between gain and current, where the light emission from an emitter is a function of current. Higher gain can amplify noise, but can also improve the function of a system where a low current drives one or more emitters (e.g., produces less light). On the other hand, increasing the current of the emitters to produce more infrared light can cause problems for smaller insects because too much infrared light emitted can reduce sensitivity. The system of the subject invention is configured to measure the relative change in light intensity received at one or more detectors as a pest passes between the detectors and the emitters. If too much light is emitted, the detectors can saturate, making them unable to detect small changes. Thus, the gain and current of a given system depend on a variety of factors, including pest size, the distance between the emitters and detectors, the saturation level of the detectors, etc.

[0043] like Figure 2 As shown, the pest detection system 100 is disposed around an entrance 202 through which pests (e.g., flies) enter a trap 200. The pests enter the bottom of the trap and pass over the entrance as they enter the interior of the trap. Figure 3 yes Figure 2A side cutaway view of a trap 200 is shown. The cutaway view shows that the trap 200 has been modified to include a landing point 204 for, for example, a fly to land on. A problem with this type of trap configuration is that sometimes pests will walk rather than fly through the entrance 202 to enter the trap 200, and because the system of the subject invention relies on wing beats to determine if a pest is present, if the pest walks, it will not be counted. The landing point 204 improves counting accuracy by providing a place for pests such as flies to land just outside the entrance 202, making it more likely that the pest will fly rather than walk through the entrance.

[0044] In another embodiment, a pest detection system may include a PCB configured in a V-shape. Figure 4 Pest detection system 400 is shown with a PCB configured at a right angle, but other angles of 10°-170° may be implemented without departing from the subject matter of the present invention. The right angle configuration provides a balance so that emitter 402 can be paired with detector 404 in a manner that minimizes blind spots. Figure 4 As shown, each side of the V-shaped PCB includes an emitter-detector-detector pattern. Each emitter corresponds to two adjacent detectors on the opposite side of the V-shape. The emitters can be configured to project infrared light according to, for example, a cone or a portion of a cone. Because each emitter emits infrared light that is projected outward, propagating away from its source, each emitter is paired with two detectors. The number of detectors paired with each emitter can vary depending on the diffusion angle associated with the emitter (e.g., the vertex angle of the emitted light cone of the emitter) and the distance between the emitter and the detector. Each of the two detectors paired with a single emitter is configured to receive mainly or only infrared light from the emitter. In some embodiments, more than two (e.g., up to 10 or more) detectors are paired with a single emitter. Although there is no theoretical limit to the number of detectors per emitter, practical limitations may arise taking into account, for example, the emitter cannot project enough light to reach the detector with sufficient intensity, or the detector is located outside the projected light area of ​​the emitter. As described above, the intensity of the infrared light received at the detector affects the system performance.

[0045] Figure 5An example housing 500 is shown that can be used with a V-shaped PCB (of the pest detection system 400). Pests enter an opening 502 on one side of the housing 500, and the V-shaped PCB is disposed therein, and the pests fly within the V-shaped portion and between the emitters / detectors. Because the housing 500 is formed with a triangular top portion, the V-shaped PCB can be disposed so that it fits inside the top. In some embodiments, the V-shaped PCB is located near the opening 502, but it can be located at any desired position along the length of the housing 500. The advantage of the V-shaped PCB being located near the opening is that pests are more likely to fly through rather than walk through the entrance of the detector / emitter, thereby improving accuracy. Figure 5 Also shown is a solar panel 504 disposed on top of the top portion, which may optionally be used to power the system.

[0046] Figure 6 Shows from Figure 5 Disassembled housing 500. Several components can be seen, including a body 506, an electronics bay 508, a solar panel 504, a rear panel 510, and a lid 512 covering the electronics bay 508. The electronics bay 508 forms an interior space surrounded by the lid 512, and is configured to house, for example, a battery 514 and a microprocessor 516, which has sufficient I / O for a V-shaped PCB. A switch 518 and an indicator LED 520 pass through the wall of the electronics bay 508. The switch 518 can be configured to turn the system on or off, and the indicator LED 520 can be configured to perform one or more different functions, such as turning on when the system is turned on, flashing or turning off when counting pests, etc. The solar panel 504 can be secured to either side of the top of the housing 500. In some embodiments, when the lid 512 is attached, the electronics bay 508 can be watertight to prevent water from damaging sensitive electronics. In embodiments where the electronics compartment 508 is not completely waterproof, the electronics compartment 508 can be located at the bottom of the housing 500 to minimize exposure to water. Finally, the back panel 510 is triangular in shape to match the shape of the housing 500. It includes a mesh portion 522 to prevent pests from fully passing through and exiting on the other side.

[0047] In some embodiments, the emitter can be paired with a detector having one or more reflective surfaces to improve the detection area. For example, Figure 7 A pest detection system 700 is shown having two parallel reflective surfaces—a first reflective surface 702 and a second reflective surface 704—which are configured as plates (e.g., assuming Figure 7 (Drawn in the xy plane, they extend in the z direction). The emitter 706 and detector 708 may be located at midpoints along the height of the reflector. Figure 8A perspective view of the PCB 710 is shown with the second reflective surface 704 connected thereto. From this figure the positions of the emitter 706 and the detector 708 can be seen. Figure 7 In the illustrated embodiment, the emitter 706 and the detector 708 are connected to the same PCB 710, and the second reflective surface 704 is disposed between them. The emitter 706 projects infrared light to the detector at a certain angle, wherein the angle θ measured from the second reflective surface 704 may be 10 to 80° (preferably 40 to 60°), and the detector 708 receives the infrared light after the infrared light is reflected one or more times between the first reflective surface 702 and the second reflective surface 704. Figure 7 As shown, light is reflected twice from the first reflective surface 702 and once from the second reflective surface 704. The reflective surface of the subject invention needs to be able to sufficiently reflect infrared light so that for a given configuration, it can reflect light from the emitter back to the detector at least once, but Figure 7 The multiple reflections shown do not depart from the present invention.

[0048] In some embodiments, a single emitter specifically designed to project a wider cone of infrared light than a typical emitter in an LED package can be used in conjunction with a set of detectors. Fig. 9 An emitter 900 is shown, which is configured to project a wide and flat cone of infrared light. In order to increase its projection angle, the emitter 900 includes a wide-angle lens portion 902.

[0049] Fig.10 It shows that the Fig. 9 1000 is a pest detection system 1000 with a dedicated transmitter 1002. PCB 1004 is a single piece, but in some embodiments, the PCB portion can be divided into multiple pieces as needed (e.g., one piece for the transmitter and another piece for the detector). Fig.11 The figure shown casts light, Fig.11 An emitter is described that is configured to project infrared light of sufficient intensity to provide pest detection over a range of 150° (-75° to 75°). The range may vary depending on the emitter configuration. For example, an emitter in a similar package may be configured to project over a range of 180° to 45°. Thus, emitter 1002 projects infrared light at 150°, which is received by detector 1006.

[0050] To accommodate the wide projection angle (e.g., 150°) of the emitter 1002, Fig.10Seven detectors 1006 are shown disposed relative to the emitter 1002 so that the detectors are positioned to receive light from the emitter 1002. The detectors 1006 are shown to be evenly spaced, but this is not required, and the detector density can be varied depending on the light intensity at different projection angles. In addition, although seven detectors are shown, fewer detectors (e.g., as few as two or three) are contemplated, and there is no theoretical maximum number of detectors per emitter.

[0051] The linear density of the area spanned by the detector along the projection of the emitted infrared light can vary. The ratio of the detector to the emitter can be a function of the angle of the emitter, the distance between the emitter and the detector, the size of the pest, etc. Like the emitter, the detector also has an angle at which the incident light can be received. For example, a wide-angle receiver can detect light fluctuations caused by pests that are not directly in front of it. The detection performance can also be affected by the pest passing in front of the detector. For example, if the pest passes near the detector, the pest usually needs to be close to the front of the detector, but if the pest passes at an intermediate distance (for example, farther from the detector than the previous example), the detector can indicate the presence of the pest in a wider range of positions. Because the detector detects light basically according to a detection cone (similar to the light cone projected from the emitter), if the pest is closer to the detector, the pest is closer to the apex of the cone, making the area where the pest can be detected smaller than when the pest is farther away. In some embodiments, the distance between the detectors can be less than 1.5 times the size of the pest to be detected. This configuration can minimize the number of pests that should be counted but not counted.

[0052] As described above, embodiments of the subject matter of the present invention are configured to detect the presence of pests (i.e., insects). This is achieved by irradiating infrared light from one or more emitters to one or more detectors. In some embodiments, some or all of the emitters and detectors are activated by pulse width modulation (PWM). Driving the emitters by PWM results in energy savings by ensuring that they are only turned on at certain times rather than always on. In addition, driving the emitters and detectors by synchronous PWM ensures that the detectors are only activated when the emitters produce light emissions, thereby saving electricity.

[0053] Duty cycle parameters (e.g., duty cycle and frequency) can be static or adaptive. For example, if the system of the subject of the present invention is configured to monitor moths with a wing beat frequency of 60 Hz, the detection sampling frequency needs to be at least 120 Hz. For a detection sampling frequency of 120 Hz, there are at least 120 "on" periods, so a duty cycle frequency of 240 Hz should be implemented. Duty cycles of 10% to 90% can be implemented, and the duty cycle percentage can be dynamically changed as needed. In some embodiments, a higher duty cycle (e.g., ≥50%) can be associated with a lower duty cycle frequency, while a lower duty cycle (e.g., ≤50%) can be associated with a higher duty cycle frequency. The duty cycle frequency should be at least twice the detection sampling frequency, and the detection sampling frequency should be at least twice the wing beat frequency of the pest to be detected. In an adaptive system, the duty cycle parameters can be changed according to various environmental factors to detect different pests. For example, factors such as location, time of day, time of year, relative humidity, temperature, etc. can be considered when setting the duty cycle, as all of these factors contribute to the likelihood of a particular type of pest being present. In a static system, the duty cycle can be set and then left unchanged. In such a system, the duty cycle frequency should be set high enough to facilitate the detection of pests with high wing beat frequencies, such as fruit flies, mosquitoes, and gnats.

[0054] Adjustments to the duty cycle parameters can be made by the user through software (e.g., providing input to set the duty cycle or on / off ratio) or hardware (e.g., turning one or more adjustment dials to set the duty cycle parameters). In some embodiments, the duty cycle parameters can be adjusted based on environmental factors as described above, but the duty cycle frequency can also be adjusted remotely by the user through software. In some embodiments, the duty cycle parameters can vary based on the location and the current pests known to be present at the location, for example, by extracting information from a database maintained by a local environmental service department (public or private). Dynamically defining duty cycle parameters based on pest type can improve efficiency.

[0055] Fig.12 The steps shown describe how the information from the detector can be used to determine whether a pest is present. In step 1200, the analog signal from the detector is monitored. The detector of the subject matter of the present invention is a photodiode, which generates an electrical signal (e.g., voltage, current, or both) when receiving light. Specifically, these photodiodes can be selected based on sensitivity to infrared light. Therefore, when the emitter projects infrared light, the detector receives the infrared light, thereby generating a raw analog signal.

[0056] In step 1202, as described above, analog signals that are not within the target range are largely eliminated by hardware frequency filtering using, for example, a bandpass filter. Certain frequencies or frequency ranges are blocked because the wing beat frequencies of pests are generally not within or near these ranges. For example, for moths whose wing beat frequencies are generally in the range of 50 to 80 Hz, the hardware filtering is set to eliminate frequencies below 40 Hz and above 100 Hz, while for smaller insects such as fruit flies and mosquitoes whose wing beat frequencies are generally in the range of 200 to 1200 Hz, a bandpass filter can be implemented to eliminate frequencies below 200 Hz and above 1500 Hz. It may be desirable to maintain frequencies from 40 Hz to about 1500 Hz, and it is clearly contemplated in this application that all ranges therebetween are feasible in the embodiments of the subject matter of the present invention, so that the range can be determined based on the wing beat frequencies of the pest type to be detected. Once the range of wing beat frequencies is known, appropriate filtering can be implemented. Although bandpass filtering is useful, in some embodiments, only frequencies below a threshold need to be eliminated, so each of the above lower limits should be interpreted as also disclosing the upper limit of the filter that removes all frequencies below that value.

[0057] After passing through the hardware filter, the signal is digitized in step 1204. As the signal is digitized, a multi-step filtering process can be implemented to minimize false positive pest detections while reducing processing costs. Although these steps are described in order, some or all of the steps can be completed in an order different from the order in which they are written. First, the analog signal from the detector is sent (e.g., continuously when generated) to an analog-to-digital converter (ADC). The digitized sampling frequency of the ADC of the subject of the present invention can be, for example, 4kHz-10kHz. In some embodiments, a sampling frequency of about 8kHz is sufficient.

[0058] Then in step 1206, the digitized signal is continuously monitored for potential pest signals. To monitor for potential pest signals, two signal amplitude thresholds are set, one high and one low. For example, it has been found that when the signal amplitude is normalized to a range of -1 to 1, setting the thresholds to -0.6 and 0.6 is sufficient for pest detection. Amplitude normalization is not required, although normalization can make the system of the subject invention more robust (e.g., better able to handle detectors that output different signal amplitudes). In embodiments where normalization is not performed, these thresholds can still be implemented by performing some basic mathematics based on the measured amplitude range of a given detector or detector group (e.g., the maximum amplitude can be associated with 1 and the minimum amplitude can be associated with -1).

[0059] When the digital signal value from the ADC exceeds any threshold a certain number of times in consecutive digitized blocks (for example, exceeds the threshold 2-3 times in a row, or exceeds 2-6 times in a row in some embodiments), this indicates the presence of potential pests. Looking for consecutive digital signal values ​​that exceed any threshold can improve performance by eliminating false positives. Once it is observed that the digital signal value has exceeded a certain threshold a specified number of times in a row, a certain number of digital signal values ​​will be recorded. In some embodiments, 1024 samples are recorded, but the number of samples recorded can range from 256 to 8192. The number of digital signal values ​​to be recorded can vary depending on several factors, including the ADC sampling rate. Fig.13 An example recorded signal is shown, plotted as amplitude versus digital signal block (in this case the recorded samples are 1024 long). The graph shows where the digital signal values ​​exceed a normalized amplitude threshold (eg + / - 0.6).

[0060] Another way to express how many digital signal values ​​should be recorded is to state the duration of the recording. Because the number of digital signal values ​​captured for a certain time depends, for example, on the sampling frequency of the ADC, stating the captured time may be more robust. For example, if the sampling rate of the ADC is about 8kHz and 1024 digital signal values ​​are recorded, this means that the system records an elapsed time of about 0.125 seconds. This can be expressed as follows:

[0061] Number of digital signal values / ADC sampling rate (Hz) = recorded elapsed time

[0062] Thus, the elapsed time should be 0.05-0.3 seconds (preferably about 0.125 seconds), but up to 1-3 seconds may be recorded without departing from the subject matter of the present invention. These ranges of elapsed times improve the ability of the system to detect pests, because sufficient time must be captured to confirm that the pest has indeed passed the detector and that the signal is not just capturing noise. It is worth noting that any value for the number of digital signal values ​​recorded should be an integer, and higher ADC sampling rates generally improve system performance (although performance and power consumption should be balanced). For example, if an ADC with a sampling rate of 4kHz is used, the number of digital signal values ​​to be recorded should be about 512.

[0063] Therefore, when the recorded signal includes consecutive digital signal values ​​that exceed the threshold normalized amplitude value, this indicates the presence of a pest, and not just some random electrical noise. Although random electrical noise can exceed the threshold, it is much less common for the noise to breach the threshold in two consecutive samples. However, in some cases, the noise can exceed the threshold in multiple consecutive digital signal values. Further filtering can be implemented to eliminate these types of false positives.

[0064] Thus, in step 1206, the system determines whether the signal in the recorded sample is indicative of pests or noise. Fig.14 How to use an onboard classifier to achieve this is shown. The onboard classifier can be implemented as software, hardware, or some combination of software and hardware. First, the recorded sample is passed to the onboard classifier, which must select which type of classifier to use based on, for example, the expected type of pest. The expected type of pest can be determined in a variety of ways, including by location, observation, etc. (as described above with respect to the duty cycle parameter). This involves feature extraction and model selection. In feature extraction, the system examines one or more peaks from the recorded signal, calculates the properties of these peaks, and then calculates the statistical results of the peak properties. Next, an appropriate model is applied to determine whether there are pests. For example, if moths are indicated as possible pests, a moth model can be used to analyze the features to determine whether to count moths. The results of this analysis are then saved to storage (for example, whether the pests are counted or the noise is ignored, the results will be saved) and can be used to improve detection through, for example, machine learning. During step 1208, it is determined whether the signal indicates pests or noise. Once the above results are saved to storage, the number of detected pests can be counted according to step 810. If any redundant counts exist, they may be eliminated according to step 812, although this step is optional as redundancies may be eliminated throughout the entire process of determining the presence of pests as described above and in more detail below.

[0065] Thus, classification involves feature extraction as described above, which is calculated in the time domain. In some embodiments, the goal of classification is to identify all peaks in the signal that exceed an upper threshold of normalized amplitude, and then combine consecutive peaks whose valleys between them do not exceed a lower threshold. From there, the peak characteristics are calculated, which involves determining the peak width. The peak width can describe the distance along the x-axis between the previous valley and the next valley of a particular peak, where the units along the x-axis represent discrete digital signal values. The height is also calculated, where the height is measured based on the signal amplitude (e.g., normalized to -1 to 1), and the height of the peak is measured in units along the y-axis between the valley and peak being analyzed. Fig.15 shows how to measure width and height.Finally, the sharpness can be calculated, where the sharpness is the height of the peak divided by the width of the peak.

[0066] Thus, the system of the subject invention can be configured to calculate the total number of peaks in the recorded signal as well as the maximum sharpness, average sharpness, maximum height, average height, maximum width, and average width. All or part of these values ​​can be used with a machine learning algorithm, as, for example, a training set, to improve classification. Fig.16 An example signal is shown with the different peaks labeled, and the maximum sharpness, average sharpness, average height, average width and peak threshold labeled. In this figure, the amplitude is not normalized to the -1 to 1 range.

[0067] Fig.17A schematic diagram of a system is shown, wherein the system is configured to perform the above and Fig.12 All or part of the steps shown in . All the details described above are also relevant here. The analog sensor describes the pulse width modulated infrared detector described above. The signal from the sensor is converted from analog to digital using an ADC sampling at, for example, 8kHz. From there, a comparator (e.g., ADS1115) monitors for consecutive blocks of digital signals that exceed an amplitude threshold. If no consecutive digital signal values ​​are detected, monitoring continues. When some number of consecutive digital signal values ​​exceed the amplitude threshold (e.g., 2 to 6, inclusive), the system wakes up and receives 1024 recorded digital signals, in Fig.17 The time, temperature, and relative humidity may then be recorded, and an onboard classifier may then be used to determine if the recorded digital signal value is indicative of a pest. Humidity, time, and temperature may all be used to improve detection, and may be saved along with other data, such as to aid machine learning. When a pest is found the information is stored to the pest storage, and when noise is found the information is stored to the noise storage. Both the pest storage and the noise storage may be saved to any type of computer readable medium, whether locally or remotely over a network connection.

[0068] Therefore, specific systems and methods related to pest detection have been disclosed. However, it will be apparent to those skilled in the art that, without departing from the inventive concept in the present application, more modifications may be made in addition to those described. Therefore, unless in accordance with the spirit of the present disclosure, the subject matter of the present invention is not limited. In addition, when interpreting the present disclosure, all terms should be interpreted in the broadest possible manner according to the context. In particular, the term "comprising" should be interpreted as referring to elements, components or steps in a non-exclusive manner, indicating that the elements, components or steps mentioned may exist, be used or be combined with other elements, components or steps not explicitly mentioned.

Claims

1. A method for detecting pests, the method comprising: emitting infrared light from at least one emitter driven by a pulse width modulated signal; generating an analog signal based on infrared light by using at least one detector driven by a pulse width modulated signal; filtering the analog signal based on the expected pest wing beat frequency; Convert analog signals into digital signals; waking up the microcontroller upon detecting at least two consecutive digital signal values ​​exceeding a threshold; Use a microcontroller to record a series of digital signal values; as well as Use the onboard classifier to determine whether a range of digital signal values ​​indicates the presence of pests. The method of claim 1 , wherein the threshold comprises an amplitude or a normalized amplitude. The method of claim 1 , wherein the series of digital signal values ​​comprises at least 1024 samples. 4 . The method of claim 1 , wherein the number of digital signal values ​​in the series of digital signal values ​​is a function of a sampling rate of the analog-to-digital converter and a duration to be recorded. The method according to claim 4 , wherein the duration is between 0.05 seconds and 0.3 seconds. The method of claim 1 , wherein at least one detector comprises a photodiode.

7. The method of claim 1, wherein at least one detector is configured to detect infrared light having a wavelength of 700 nm-1400 nm.

8. The method of claim 1, wherein at least one emitter is configured to project infrared light having a wavelength of 700 nm-1400 nm.

9. The method of claim 1, wherein the filtering step comprises using a hardware bandpass filter to allow frequencies between 40 Hz and 1500 Hz to pass.

10. The method of claim 1, wherein the filtering step comprises using a hardware bandpass filter to allow frequencies between at least one of: 40 Hz and 100 Hz and 200 Hz and 1500 Hz.

11. A method for detecting pests, the method comprising: emitting infrared light from at least one emitter driven by a pulse width modulated signal; generating an analog signal based on infrared light by using at least one detector driven by a pulse width modulated signal; filtering the analog signal based on the expected pest wing beat frequency; Convert analog signals into digital signals; waking up the microcontroller upon detecting at least two consecutive digital signal values ​​exceeding a threshold; Use a microcontroller to record a series of digital signal values; Select an onboard classifier based on the expected pest type; Use an onboard classifier to determine whether a series of digital signal values ​​indicates the presence of pests; as well as The number of pests is calculated based on the number of times the onboard classifier determines that the pests are present. The method of claim 11 , wherein the threshold comprises an amplitude or a normalized amplitude.

13. The method of claim 11, wherein the series of digital signal values ​​comprises at least 1024 samples.

14. The method of claim 11, wherein the number of digital signal values ​​in the series of digital signal values ​​is a function of a sampling rate of the analog-to-digital converter and a duration to be recorded.

15. The method of claim 14, wherein the duration is between 0.05 seconds and 0.3 seconds.

16. The method of claim 11, wherein at least one detector comprises a photodiode.

17. The method of claim 11, wherein at least one detector is configured to detect infrared light having a wavelength of 700 nm-1400 nm.

18. The method of claim 11, wherein at least one emitter is configured to project infrared light having a wavelength of 700 nm-1400 nm.

19. The method of claim 11, wherein the filtering step comprises using a hardware bandpass filter to allow frequencies between 40 Hz and 1500 Hz to pass.

20. The method of claim 11, wherein the filtering step includes using a hardware bandpass filter to allow frequencies between at least one of: 40 Hz and 100 Hz and 200 Hz and 1500 Hz.

Images