WO2023247209A1

WO2023247209A1 - Apparatus and method for measuring insect activity

Info

Publication number: WO2023247209A1
Application number: PCT/EP2023/065457
Authority: WO
Inventors: Michael Stanley Pedersen; Knud Poulsen; Emily BICK; Frederik ELBÆK
Original assignee: Faunaphotonics Agriculture & Environmental A/S
Priority date: 2022-06-21
Filing date: 2023-06-09
Publication date: 2023-12-28

Abstract

A computer-implemented method of determining a resulting measure of insect activity in a geographic area, the method comprising obtaining a first plurality of sensor data from a movable detection system comprising a plurality of sensors concurrently traversing the geographic area, the plurality of sensors being coupled physically to each other, each sensor being configured to record data indicative of insect activity, the first plurality of sensor data comprising a first set of sensor data from a first sensor and a second set of sensor data from a second sensor, the first set of sensor data comprising one or more digital images, the second sensor being configured to obtain data indicative of an insect in flight, determining a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data, and calculating a resulting measure of insect activity based at least on the first and second measure of insect activity, the resulting measure of insect activity being indicative of insect activity in at least a portion of a target area traversed by the movable detection system, the target area being comprised in the geographic area.

Description

Apparatus and method for measuring insect activity

Technical field

The present disclosure relates to an apparatus and method for measuring insect activity.

Background

It is generally desirable accurately to determine insect activity within a geographic area, such as an agricultural area, a forested area, or another geographic area of economic or biological interest.

For example, knowledge about insect activity in a geographic area may serve as valuable input for supporting various decisions, e.g. decisions about which types of crops or plants to plant in order to support biodiversity, and to optimize the use of insecticides, etc.

When distributing insecticides across a field of crops or other geographic area on which insects are to be controlled, it is generally desirable to apply the right types of insecticides in the right amounts in order to obtain an efficient insect control while not applying unnecessary, useless or environmentally harmful amounts of insecticides. Therefore, an accurate measurement of local insect activity is desirable.

Previous work has focused on detection of insects, and particularly on detecting an insect in flight, which may be done using an insect sensor such as any disclosed in WO 2021/165479. In this way one or more insects of interest may be detected and possibly identified. However, an approach primarily detecting insects during certain life stages or primarily detecting insects in flight has certain limitations.

For example, an insect of interest may fly primarily in specific weather and at certain times during a day. Further, an insect of interest may have one or more life stages during which they are unable to fly and/or one or more life stages during which they are primarily detected by their effect on their surroundings.

It is thus generally desirable to provide a reliable measurement of insect activity that does not rely on recording an insect during a specific period or life stage. Summary

The present disclosure relates to different aspects each yielding one or more of the benefits and advantages described in connection with one or more of the other aspects, and each having one or more embodiments corresponding to the embodiments described in connection with one or more of the other aspects and/or disclosed in the appended claims.

According to one aspect, disclosed herein are embodiments of a computer-implemented method of determining a resulting measure of insect activity in a geographic area.

The method comprises obtaining a first plurality of sensor data from a movable detection system comprising a plurality of sensors concurrently traversing the geographic area, each of the sensors in the plurality of sensors being coupled physically to each other, each sensor being configured to record data indicative of insect activity, the first plurality of sensor data comprising a first set of sensor data from a first sensor and a second set of sensor data from a second sensor, and the first set of sensor data comprising one or more digital images. The second sensor may be configured to obtain data indicative of an insect in flight.

The method further comprises determining a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data. Further, the method comprises of calculating a resulting measure of insect activity based at least on the first and second measure of insect activity. The resulting measure of insect activity is indicative of insect activity in at least a portion of a target area traversed by the movable detection system, the target area being comprised in the geographic area. Thus, the method allows for fusion of sensor data from multiple sensors, such as fusion of different types of sensor data.

The second sensor may be configured to acquire sensor data indicative of at least one insect signature. An insect signature is a measurable attribute, such as an optically detectable attribute, which can be utilized in the classification of the detected insect. Examples of an insect signature is: a wing beat frequency, a trajectory, a body-wing ratio, a relative or absolute total size, a relative or absolute body size, a relative or absolute wing size, a glossiness measure, a melanisation measure, etc. For the purpose of the present description, an insect signature is a measurable attribute different from a digital image. In some embodiments, the sensor data from the second sensor is indicative of a time-resolved quantitative measurement. Other examples of an insect signature are: a characteristic modulated electric field pattern, such as a characteristic modulated electric field pattern correlated or associated with one or more fundamental wing beat frequencies, and/or a characteristic modulated electric field pattern correlated or associated with one or more harmonics of a fundamental wing beat frequency. A characteristic modulated electric field pattern or a wing beat frequency may be determined at least in part from modulated electric field data. A wing beat frequency may be a free flight wing beat frequency, i.e. the frequency of wing beats of a free flying insect, or a perching wing beat frequency, i.e. the frequency of wing beats of an insect sitting on a surface. Thus, the second sensor may allow for detection, and possibly classification and/or identification of insects, and does so supported by an analysis of acquired sensor data. In some embodiments, the second set of sensor data is different from one or more digital images. Accordingly, the first and second sensors may be different types of sensors configured to produce different types of sensor data.

The sensor data may be used as input for a classification algorithm so as to arrive at a type, e.g. species, of an insect detected. The sets of sensor data may be used as training datasets for a machine-learning model. For example, the one or more digital images may be used with image recognition software that can determine the type, e.g. species, of an insect and from this, the machine-learning model can be trained to recognise the type, e.g. species, of insect based on the sensor data provided by the second sensor.

The movable detection system may comprise one or more processing units configured to receive the first and/or second sensor data from the first and/or second sensor, respectively, and to process the received sensor data to extract information therefrom, for example an insect signature. The one or more processing units may be configured to determine the first measure of insect activity based at least on the first set of sensor data, and/or the second measure of insect activity based on the second set of sensor data. In other embodiments, some or all the processing steps are performed by a data processing system external to the movable detection system, or the processing steps may be distributed between a local processing unit of the movable detection system and a remote data processing system, separate from the movable detection system.

Here and in the following, the term processing unit is intended to comprise any circuit and/or device suitably adapted to perform the functions described herein. In particular, the term processing unit comprises a general- or special-purpose programmable microprocessor unit, such as a central processing unit (CPU) of a computer or of another data processing system, a digital signal processing unit (DSP), an application specific integrated circuits (ASIC), a programmable logic arrays (PLA), a field programmable gate array (FPGA), a special purpose electronic circuit, etc., or a combination thereof.

Determination of the first measure and/or of the second measure and/or of the resulting measure may comprise identification based on the respective sensor data which may be in the form of a pre-processed sensor signal. For example, identification of one or more insects or of crop alteration. Thus, in some embodiments of the method, the process is capable of detecting individual insects and/or identify, i.e. distinguish different types, e.g. different species, of insects. Accordingly, any of the measures of insect activity may be a total insect activity indicative of the amount of all detectable insects. The first and second measures of insect activity may each be indicative of a number of insects in the same part of the geographic area, and the resulting measure of insect activity may be indicative of the insect activity in said same part of the geographic area. Alternatively, any measure of insect activity may be a specific insect activity indicative of the number of insects of one or more specific types, e.g. one or more specific species, life stages, and/or the like. Yet alternatively, any measure of insect activity may be indicative of a local biodiversity index, i.e. a numerical estimate of the diversity of insect at a specific level such as species, family, or order. In some embodiments, the detected insects are or include airborne insects moving above a ground surface and/or insects located on the vegetation and/or the ground surface. Examples of airborne insects include flying insects, jumping insects and insects that hover, glide or float in the air. The ground surface may be the upper surface of the soil, an upper surface of a vegetation canopy or another reference surface.

Here, different types of insects may refer to different species, functional groups, or to other insect categories of a suitable taxonomy. Alternatively or additionally, different types of insects may refer to different life stages of insects and/or to other classifications.

The identification of either an insect or of a crop alteration may be correlated, whereby the identification of one or the other is used in the identification of the remaining one. The identification may at least in part be based on a look-up table, a decision tree, a neural network, a support vector machine, and/or the like. Alternatively or additionally, the identification may directly be based on the sensor data. The identification of respective types of insects may be based on an indicator feature extracted from the first and/or second set of sensor data by a trained machine-learning algorithm. Such identification may e.g. be performed by a trained machine-learning model, e.g. a model configured to receive a representation of the sensor data and to classify the received data, such as classify the receive data into one of a known type of insects or one of a known type of crop alteration.

Examples of suitable machine-learning models include convolutional neural networks. For example, in some embodiments, the sensor data may be fed into a neural network, such as a convolutional neural network, or another type of machine-learning model. The neural network may be a feed-forward neural network that includes one or more input layers receiving the sensor data, e.g. a time series of detected light intensities at one or more wavelengths, and/or one or more digital images. The neural network may optionally receive additional inputs. The neural network may include one or more hidden layers and an output layer. The neural network may be trained, based on a dataset of training examples, to classify the sensor data. To this end, the training examples may include sensor data that have been obtained previously. The hidden layers of the trained neural network may represent automatically extracted features of the sensor data that are fed into the neural network. The output layer may represent a classification received sensor data based on feature values of the automatically extracted features as extracted by the one or more hidden layers.

Insects may fly only for short periods of time during the day, such as e.g. 0.5 - 4 hours per day and/or mostly during particular periods of the day such as morning, noon, and/or evening. This means that the absolute number of insects detected at a given circadian time may vary greatly. Some species of insects are known to fly much of the time, such as fruit flies, while other species hardly fly at all, such as Japanese beetles. Only adult insects fly, while all other life stages are non-flying. Some insects have one or more life stages during which they are primarily detectable by their effect on their surroundings, rather than as individuals.

The combination of sensor data from a sensor configured to record one or more digital images with sensor data from at least one other sensor configured to record data indicative of insect activity can advantageously mitigate the limitations of a single sensor recording as basis for a calculation of a measure of insect activity in a target area.

When the movable detection system records sensor data at a given time, an insect in the target area may be flying, or sitting on part of the vegetation, or sitting on the ground. It may also be that the only indication of the insect being, or having been, present in the target area is crop alteration. Crop alteration may be a change in the physical appearance of a crop, such as e.g. damage to one or more leaves, and/or change in colouring of at least part of the crop, and/or change in leaf shape such as rolling of a leaf. By comprising a plurality of sensors, the functionality of the movable detection system may be increased as the plurality of sensors expand the time of day the detection system is able to detect an insect; this expanded time allows for greater inferences in pests activities, population dynamics, and insect diversity. For example, with the detection system being able to detect both flying and non-flying insects, the detection system may be able to detect an insect during time periods when individuals are in flight and during time periods when individuals are located on the crop or ground. Further, having the plurality of sensors may increase the range of insect species that the detection system is able to detect; For example, by the detection system being able to detect insects in multiple life stages and/or being is able to detect both insects that are frequently in flight, minimally in flight, and anything in-between. Additionally, the detection system may be less susceptible to being disadvantaged by poor timing due to its ability to detect an insect during more than one, or during any of multiple, life stages and/or by being able to detect crop alteration or earth alteration caused by an insect. Further, the detection system with its plurality of sensors will be suitable for detection of different types of insects, when compared to insect detection systems which are configured and optimized for detection of a single type of insect, e.g. mosquitos. In some embodiments, the one or more digital images are recorded in at least a part of the visible spectrum. In some embodiments, the first sensor is a digital camera, such as a commercial digital camera. In some embodiments, the first sensor is configured to record in the visible spectrum, I R, and/or UV. The first sensor may be a high-resolution image sensor, such as a high-resolution digital camera. The one or more digital images may be one or more single images or a time sequence of images/video frames. The time sequence of images may be configured in a video format. The video recording may comprise audio recorded at the same time. The time sequence may be a high-speed video clip.

In some embodiments, at least one sensor in the plurality of sensors, in particular one of the first and second sensors, is configured to obtain data indicative of an insect in flight, and at least one sensor in the plurality of sensors, in particular one of the first and second sensors, is configured to obtain data indicative of an insect not in flight and/or of crop alteration or earth alteration caused by one or more insects. Both the first and second sensor may be configured to obtain data indicative of an insect in flight, and/or both the first and second sensor may be configured to obtain data indicative of an insect not in flight and/or of crop alteration caused by one or more insects.

The first and second sensor each have a respective sensory field, where the respective sensory field is the area or volume within which the sensor is capable of recording data indicative of insect activity, i.e. it represents the area or volume within which the sensor is able to "see". For example, for a camera, the sensory field is the Field-of-View of the camera. The first and second sensors may be configured such that their sensory fields overlap, partially overlap, or do not overlap. The sensory fields of the first and second sensors may be different, such as be different types of sensory fields, or have different shapes. The first and second sensor may be configured such that their sensory fields point in substantially the same direction, or they may point in different directions. The first and second sensors may be configured such that their sensory fields have a predetermined, in particular constant, spatial relationship relative to each other while the movable detection system moves within the geographic area.

In some embodiments, the second sensor is an insect sensor such as any embodiment of insect sensor disclosed in WO 2021/165479. In some embodiments, the second sensor is an optical insect sensor configured to detect light from a probe volume, which is the optical insect sensors sensory field. An optical sensor may acquire sensor data within a probe volume extending outside the insect sensor device by detecting light from the probe volume. Some embodiments of an optical insect sensor comprise an illumination module configured to illuminate the probe volume and further comprise one or more detectors configured to detect light from the probe volume, in particular light emitted, in particular reflected or backscattered, by insects responsive to being illuminated by the illumination module. The detector module may thus output a sensor signal indicative of the detected light, e.g. indicative of a detected light intensity as a function of time. An optical insect sensor may be configured to optically detect one or more attributes associated with insect detection events in the probe volume of the insect sensor, in particular in a probe volume outside and in a proximity of the insect sensor. In some embodiments, the probe volume is located in a proximity of the optical insect sensor. In particular, the probe volume may extend between a proximal end and a distal end of the probe volume, relative to the optical insect sensor, e.g. relative to an aperture or other optical input port of a detector module comprising a detector configured to detect light from the probe volume. In some embodiments, the distal end may be no more than 5 m from the optical insect sensor, such as no more than 4 m, such as no more than 3 m. The proximal end may be separated from the optical insect sensor, e.g. from an aperture or other optical input port of the detector module, by 1 cm or more, such as by 10 cm or more, such as by 20 cm or more, such as by 30 cm or more.

An optical insect sensor is particularly useful for recording data indicative of insect activity in a geographic area. In particular, the insect sensor is non-intrusive to the environment in the sense that it does not rely on and, consequently, is not biased by pheromones, light, colour, or other means of attracting, trapping or killing insects. In particular, insects may be detected in their natural environment regardless of their affinity to a certain lure or trap technology, thus reducing the biases of the measurement results to different trapping techniques for different insect species. To this end, the probe volume is preferably an enclosure-free void/space allowing unrestricted movement of living airborne insects into and out of the void/space. Moreover, an optical insect sensor can be easily moved across a large target area and can perform measurements in relatively short measurement times. In other embodiments, other types of insect sensors may be used, that provide data directly or indirectly indicative of the presence of insects in a local vicinity of the sensor. In some embodiments, the method comprises identifying from the second set of sensor data, one or more types of insects, and/or determining respective amounts or numbers of the one or more type of insects detected in the probe volume of an optical insect sensor.

In some embodiments, the second sensor is an electric field (E-field) sensor configured to acquire electric field data, such as a sensor configured to acquire data on electric field strength. The E-field sensor may comprise one or more electric field probes, and/or one or more radio antennas. Insects may modulate electric fields in their surroundings, such as modulate the amplitude of an electric field around them, e.g. of the Earth's static electric field. Such modulations may be measured by an electric field sensor. In some embodiments, the E-field sensor is configured to acquire modulated electric field data. In some embodiments, one or more E-field sensors are each configured to acquire frequency-modulated electric field data, wherein the frequency of the modulation is in the frequency range between 0.01 kHz and 22 kHz, such as between 0.01 kHz and 5 kHz, such as between 0.01 kHz and 2 kHz, such as between 0.01 kHz and 1 kHz. The E-field sensor may be configured to detect a near-field electric field modulation from an insect. In some embodiments, the insect sensor comprises two or more E-field sensors each configured to acquire data on electric field strength. An E-field sensor may be further configured for at least partly passive detection. Thus, the E-field sensor may be configured to receive, but not transmit signals during at least part of the sensors operating time. This may provide for an energy efficient sensor. The E-field sensor may be configured such that the probe volume is substantially cylindrical around a receiver of the sensor or the sensor may be configured such that the probe volume is substantially spherical around a receiver of the sensor. The radius of the extent of a probe volume that is substantially cylindrical or spherical may be in the range of 0.1 m to 10 m, such as in the range 0.15 to 5 m, such as 0.2 m to 3 m, such as 0.25 to 2 m, such as up to 20 m, or larger. In some embodiments, the probe volume of the E-field sensor may be e.g. cuboid, spherical, cylindrical, or other geometric shape, or a shape characteristic of an antenna pattern. In some embodiments, the second sensor is comprised in an insect sensor system comprising an electric field generator configured to generate an electric field at a point or area of measurement of at least one E-field sensor in the insect sensor system. In some embodiments, the insect sensor system comprises an electric field generator configured to generate an electric field in a volume surrounding at least one E-field sensor. Thus, the electric field generator may be configured to generate and shape a probe volume, where there is no significant electric field present, or in addition to an existing probe volume, e.g. an existing electric field. The probe volume generated by the electric field generator may be e.g. cuboid, spherical, cylindrical, or other geometric shape, or a shape characteristic of an antenna pattern. The insect sensor system may advantageously be configured to provide an electric field probe volume that is an enclosure-free void/space allowing unrestricted movement of living airborne insects into and out of the void/space.

The movable detection system may comprise a propulsion mechanism, e.g. a motor and wheels, belts, a propeller, or other type propulsion system. The movable detection system may thus be self-propelled. The movable detection system may be part of, such as integrated into or mounted on, a manually driven, semi-autonomous or autonomous vehicle. For example, the movable detection system may be part of, such as integrated into or mounted on, a tractor, a movable farming machine, a spraying boom, or other agricultural vehicle, on an unmanned aerial vehicle, a self-driving robot, or the like. The movable detection system may be part of, such as integrated into or mounted on, a ground vehicle or an aerial vehicle. The movable detection system may be mounted on, or be mountable on, an autonomous vehicle or an operator-controlled vehicle, such as a remote-controlled vehicle or a manned vehicle. The first and second sensors may be mounted on the same vehicle or they may otherwise be configured to move together through the geographic area. In this way, the first and second sensor are physically coupled to each other as they traverse the geographic area. The first and second sensor may be positioned to have a predetermined distance between each other and/or with their respective sensory fields oriented in a predetermined manner. The first and second sensor may be positioned adjacent to each other, such as right next to each other or with a small distance in-between, or they may be positioned further apart. The first and second sensor may be positioned to have a predetermined distance between them of 0 cm to 200 cm, such as of 0 cm to 150 cm, such as of 0 cm to 100 cm, such as of 0 cm to 75 cm, such as of 0 cm to 50 cm, such as of 0 cm to 30 cm, such as of 0 cm to 10 cm, such as of 0 cm to 5 cm, such as of 0 cm to 2 cm.

The first and second sensors traverse the geographic area, while being coupled physically to each other. The first and second sensor may be coupled to each other via a rigid structure, such as a spray boom of a tractor. The first and second sensor may have a constant distance and/or spatial configuration relative to each other. Alternatively, the first and second sensor may have a distance and/or spatial configuration to each other that at most varies within predetermined constraints. For example, any of the sensors may be configured to move within predetermined constraints, such as a camera moving as part of a sweeping motion. This allows each of the sensors in the plurality of sensors to record their data indicative of insect activity in a known correlated manner. For example, the sensor data recorded by each of the sensors in a plurality of sensors may be correlated both temporally and spatially as the geographic area is traversed by a movable detection system. With each of the sensors in a plurality of sensors being spatially correlated in a known manner, the relative position and extent of the sensory field of each sensor may also be known to a high accuracy during recording of data indicative of insect activity. This may be particularly advantageous when different types of insect activity are to be related to each other such as e.g. the detection of crop alteration and the detection of one or more insects.

Embodiments of the method described herein do not rely on a large number of stationary sensors or traps placed in a grid or other formation over the target area of interest, thus reducing the costs of the measurements and allowing the acquisition of relevant data without time-consuming collection and integration of data from many sensors or traps.

Embodiments of the method described herein thus facilitate an efficient measurement of the spatially resolved insect activity, which is relatively inexpensive, uses only limited hardware and power, involves only limited disruption to the target area and allows integration of the measurements with other activities already requiring a vehicle to move over the field. Additionally, using a movable detection system allows larger areas to be mapped than would be feasible or reasonable with a larger number of stationary sensors.

The speed at which the detection system traverses the target area may be substantially constant or vary. The movable detection system may be configured to record and/or process sensor data, while the detection system is either moving or standing still. The movable detection system may be configured to have a stop-measure-go mode in which the movable detection system stops to allow the plurality of sensors, i.e. at least the first and second sensor, in the detection system to record data at a standstill before the detection system moves on. The movable detection system may be configured to vary its speed, such as slow down, speed up, or stop, to allow the plurality of sensors in the detection system to record data at an appropriate speed. The appropriate speed may be dependent on factors determining what data indicative of insect activity the first and/or second sensor should be able to capture. For example, the appropriate speed may be dependent on which insects are anticipated to be present in the target area, and/or crop(s), and/or weather, and/or time of day.

Thus, the movable detection system may move at different speeds, it may repeatedly revisit some portions of the target areas or otherwise traverse the target area in a non- uniform manner. Therefore, the actual number of insects detected by the detection moving system in different parts of the target area may not be equally representative for the actual insect distribution. The amount of time spent recording sensor data at a location may therefore be a parameter in the calculation of the resulting measure of insect activity. In some embodiments, the movable detection system may repeatedly traverse some or even all portions of the target area, e.g. so as to increase the accuracy of the acquired data. This process may be performed according to a predefined schedule and/or it may be performed adaptively, responsive to the already acquired data.

In some embodiments, the method receives position information indicative of respective sensor positions of the first and/or second sensor within the target area having been traversed by the movable detection system. Based on the received position information, the process may determine respective local observation times for the portion of the target area traversed. In some embodiments, the position information may be received from the movable detection system or from a vehicle on which the movable detection system is mounted. Examples of suitable position sensors include a GPS sensor or another suitable sensor employing a suitable satellite-based positioning system. Alternatively, the positions of the first and/or second sensor may be detected in another manner, e.g. by monitoring the movement of the movable detection system by a stationary sensor, from an unmanned aerial vehicle and/or the like. The position information may be time-resolved position information indicative of the sensor positions of the first and/or second sensor at respective times. In some embodiments, the process only records sensor position information of the first and/or second sensor when the sensor is active, i.e. when the sensor is acquiring sensor data.

In some embodiments, calculating the resulting measure of insect activity is responsive to input associated with expected insect behaviour. That is, one or more steps in the calculation process producing the resulting measure is configured such that it can be affected by input associated with expected insect behaviour. Input associated with expected insect behaviour may be input that quantifies or qualifies the sensor data. Input associated with expected insect behaviour may be information obtained at substantially the same time as the sensor data is recorded or it may be information obtained at an earlier time.

For example, the calculation process may be responsive to the cumulated temperature units termed 'degree days' as an input. The calculation process may be responsive to being given the calendar date as an input, for example to describe a seasonal variation in insect behaviour such as e.g. a variation in which life stages of one or more insects to expect. The calculation process may be responsive to being given the circadian time as an input so as to describe a variation in insect behaviour over the course of a day such as e.g. the variation of circadian flight activity of one or more insects. The calculation process may be responsive to being given location and/or climate information, such as a GPS location, e.g. to correlate to expected presence of one or more insects. The calculation process may be responsive to being given weather information as an input so as to describe a variation in insect behaviour due to the weather. In some embodiments, the calculation process is responsive to a plurality of input, such as all of the above examples. The calculation may be responsive to a formula inclusive of days post an event such as initial monitoring. The calculation may be in responsive to field collected or trap calibration data from current or previous years and/or current or previously evaluated fields.

The first and second measure of insect activity are based on the sensor data indicative of insect activity, i.e. are correlated with one or more instances of insect activity detection. The calculation of the resulting measure being responsive to an input may act to adjust what conclusion may be drawn from the acquired sensor data. Being responsive to an input may mean e.g. applying a weight to the first and/or second sensor data, possibly relatively to each other, and/or increasing or decreasing the resulting measure.

The method may further comprise obtaining a second plurality of sensor data from a movable detection system comprising a plurality of sensors. The movable detection system recording the second plurality of sensor data may be the same movable detection system as the one that recorded the first plurality of sensor data or it may be a different, but similar movable detection system. The second plurality of sensor data comprises a third set of sensor data from a first sensor and a fourth set of sensor data from a second sensor, the second plurality of sensor data being recorded at a later time during the same day or on a later day relative to the first plurality of sensor data, and the method further comprises determining a third measure of insect activity based on the third set of sensor data, and a fourth measure of insect activity based on the fourth set of sensor data, and calculating a resulting measure of insect activity is further based on the third and fourth measure of insect activity.

The difference in time between the recording of the first plurality of sensor data and of the second plurality of sensor data may be seconds, minutes, hours, days, years, etc. For example, the difference in time between the recording of the first plurality of sensor data and of the second plurality of sensor data may be more than 30 minutes, more than 2 hours, more than 4 hours, more than 8 hours, more than 12 hours, more than 16 hours, more than a day, more than 2 days, more than 4 days, more than a week, more than 2 weeks, more than a month, more than 2 months, more than 4 months, more than 6 months, more than 8 months, more than 10 months, more than a year, more than 2 years, more than 4 years, more than 6 years, more than 10 years.

The second plurality of sensor data may be obtained in the same geographic area, in the same target area, or substantially in the same GPS location as the first plurality of sensor data. Alternatively, the second plurality of sensor data may be obtained in a similar, but different, geographic area or target area to the first plurality of sensor data such as in an adjacent geographic area or adjacent target area. The second plurality of sensor data may be obtained in a different geographic area having a similar climate as the first plurality of sensor data.

In some embodiments, the method further comprises obtaining target information indicative of one or more anticipated insects in the target area, and wherein calculating a resulting measure of insect activity is responsive to the target information.

Information indicative of one or more anticipated insects in the target area traversed by the mobile detection system may have numerous advantages, such as facilitating highly sensitive identification of one or more insects and/or limiting the scope of insect activity indications to scan for. Target information indicative of one or more anticipated insects in the target area may be explicit, such as a list of one or more target insect(s) expected in or near the target area, or implicit, such as e.g. type of crop(s) present in or near the target area. It is known that specific types of crops either in the target area or near the target area, or a combination of crops in the target area and in a nearby area, will be indicative of one or more anticipated insects in the target area. Further, it is known that there is a correlation between insects, whereby the presence, or anticipated presence, of one insect is indicative of one or more further anticipated insects. The target information may be a plurality of data such as e.g. a GPS location and date, which may be used to narrow the scope of anticipated insects in the target area. The target information may be provided by a user, by a data storage system, or by a data information system.

In some embodiments, the method further comprises obtaining locality information relating to the locality conditions such as e.g. degree days, time of day, date, GPS location, weather information, speed at which the movable detection system was moving during the recording of the sensor data, relative position, amount of time spent recording sensor data at a location, etc., and wherein calculating a resulting measure of insect activity is responsive to the locality information. The relative position may be a relative position in the crop field, such as e.g. distance to edge of crop field. The locality information may be provided by a user, by a data storage system, by a data information system, or by a device configured to measure the locality information.

In some embodiments, the method further comprises obtaining historical data, e.g. data recorded at an earlier time than the first and/or second plurality of sensor data, and wherein calculating a resulting measure of insect activity is responsive to the obtained historical data. Historical data is thus data having an earlier time stamp than the first and/or second plurality of sensor data. Historical data may be e.g. locality information, and/or target information, and/or sensor data, and/or one or more measures, such as one or more resulting measures, of insect activity. Historical data may have been recorded by the same or a similar movable detection system, or it may originate from an outside source, such as a user or a database. Historical data may e.g. be the type of crop planted in the field at a previous time, for example in the case of crop rotation.

The historical data may have been recorded in the same geographic area, in the same target area, or substantially in the same GPS location as the first and/or second plurality of sensor data. Alternatively, the historical data may have been recorded in a similar, but different, geographic area or target area to the first and/or second plurality of sensor data such as in an adjacent geographic area or adjacent target area. The historical data may have been recorded in a different geographic area having a similar climate as the first and/or second plurality of sensor data. The historical data may be data relating to the same crop, but in an earlier season, and possibly at a different geographic area. For example, the target area may be a soy field and historical data retrieved may be data from the previous year relating to a soy field in a different geographic location, maybe with a similar climate.

The calculation of the resulting measure being responsive to an input, target information, locality information, and/or historical data may act to adjust what conclusion may be drawn from the acquired sensor data.

In some embodiments, the method further comprises determining one or more local insect control measures in at least part of the target area based on the resulting measure of insect activity. Insect control measures may be e.g. the release of beneficial insects and/or biological agents for reducing or suppressing undesired insect activity. In some embodiments, the method further comprises controlling an insect activity control device to perform one or more of the determined local insect control measures in the target area. Accordingly, in one aspect, disclosed herein is a method for controlling an insect activity control device for controlling insect activity based on the calculated resulting insect activity. The insect activity control device may be a sprayer configured to perform spraying responsive to the position of the sprayer within the target area. Alternatively, the insect activity control device may be another vehicle, such as framing vehicle, or other movable device for performing a suitable insect control measure.

For example, a precision sprayer may be controlled to selectively spray insecticide only in a target area having an insect activity measure above a predetermined threshold, or the dosing of the insecticide may otherwise be controlled in dependence of the local insect activity, where the local insect activity is the insect activity in the geographic area, or a part thereof. When certain species of insect are identified, the type of insecticide or other insect control measure may automatically be selected to selectively target the identified insects. In some embodiments, biological control agents of insect pests may be released in response to specific insect activity. In other embodiments, a general insect activity may be used as a proxy indicator for specific insect activity.

Having a more accurate knowledge of local insect activity may help farmers with precision spraying, e.g. only spraying the areas with high activity, and also potentially help identify problem areas, or areas which consistently show early signs of infestation before others, or areas which are relatively more frequented by beneficial insects.

Spraying in fewer areas and/or only in areas, where one or more insects have been detected, will facilitate a lowering of the insects' ability to increase their resistance to a pesticide, as targeted spraying will act to limit the insects exposure to survivable quantities of the pesticide. Ideally, an insect is only exposed to a pesticide in a quantity that is large enough to kill it, and not to any smaller, survivable quantities that allow for evolutionary selection for those individuals who fare better and thus increase resistance to pesticide from generation to generation. In some embodiments, a precision sprayer may be controlled to selectively spray and/or dose insecticide in dependence of the local insect activity, for example in dependence of the local insect activity as represented by an insect distribution map.

Accordingly, in an aspect disclosed herein are embodiments of a computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area, the method comprising:

- obtaining a first plurality of sensor data from a movable detection system comprising a plurality of sensors coupled physically to each other and each configured to record data indicative of insect activity, the first plurality of sensor data comprising a first set of sensor data from a first sensor and a second set of sensor data from a second sensor, the first set of sensor data comprising one or more digital images;

- determining a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data;

- calculating a resulting measure of insect activity based at least on the first and second, measure of insect activity, the first resulting measure being indicative of insect activity in at least a portion of a target area traversed by the movable detection system, the target area being comprised in the geographic area;

- creating, based on the resulting measure of insect activity, an insect distribution map of the target area, the insect distribution map representing local insect activity in respective parts of the target area.

The process may compute an insect distribution map indicative of spatially resolved insect activity across the target area, for example in subareas defined within the target area. The measured spatial distribution of insect activity in a geographic area may have a number of uses, such as adapting a local spraying plan. For example, accurate knowledge of the spatial variation of insect activity may help farmers with precision spraying, e.g. only spraying the areas with higher than threshold insect activity, such as only spraying when the insect activity is high enough to reach a threshold, e.g. where the expense of spraying is outweighed by the economic gain, and also potentially help identify problem areas, or areas which consistently show early signs of infestation before others, or areas which are relatively more frequented by beneficial insects. Insects are often non-uniformly distributed across an area and hot spots of locally high insect concentrations may occur. However, the location of such hot spots may change over time making it desirable to update the insect distribution map, possibly at regular intervals.

The computation of the insect distribution map is based on the resulting measure of insect activity, which in turn may be based on locally detected insects and/or on detecting how the insects affect the vegetation based on sensor data recorded by the movable detection system while it traverses the target area. As the process acquires sensor data by traversing the area by a movable detection system comprising a plurality of sensors each configured to record data indicative of insect activity, a high resolution of detection may be achieved.

Various decision-making processes, such as spraying decisions, decisions about the use of fertilizers, types of crops, irrigation and/or the like may be based on an insect distribution map as described herein, optionally in combination with other information, such as environmental data, crop data, information about soil quality etc. When one or more specific insect distribution maps are created, specific to one or more particular insect types, e.g. insect species, life stages, etc., the decision making may further be based on knowledge about these specific insect types.

The computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area may make use of more than one plurality of sensor data. For example, a second plurality of sensor data as described herein.

Accordingly, disclosed herein are embodiments of a computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area, the method comprising:

- obtaining a second plurality of sensor data from a movable detection system comprising a plurality of sensors coupled physically to each other and each configured to record data indicative of insect activity, the second plurality of sensor data comprising a third set of sensor data from a first sensor and a fourth set of sensor data from a second sensor, the third set of sensor data comprising one or more digital images, the second plurality of sensor data having a different time stamp than the first plurality of sensor data;

- determining a first measure of insect activity based at least on the first set of sensor data, a second measure of insect activity based on the second set of sensor data, a third measure of insect activity based at least on the third set of sensor data, a fourth measure of insect activity based on the fourth set of sensor data,

- calculating a resulting measure of insect activity based at least on the first, second, third and fourth measure of insect activity, the first resulting measure being indicative of insect activity in at least a portion of a target area traversed by the movable detection system, the target area being comprised in the geographic area;

The insect distribution map created by various embodiments of methods described herein may be used as a basis for controlling precision spraying and/or other insect control measures such as the release of beneficial insects and/or biological agents for reducing or suppressing undesired insect activity. To this end, the insect distribution map may be used directly as an insect control prescription map representing the degree of local insect control measures. For example, a precision sprayer may be controlled to selectively spray insecticide only in subareas of the target area having an insect activity above a predetermined threshold, or the dosing of the insecticide may otherwise be controlled in dependence of the local insect activity as represented by the insect distribution map. When the insect distribution map identifies certain species, the type of insecticide or other insect control measure may automatically be selected so as to selectively target the identified insects. In other embodiments, general insect activity may be used as a proxy indicator for specific insect activity.

Accordingly, the computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area may be used in a method for controlling insect activity in a target area, which comprises:

- creating an insect control prescription map of the target area based on the insect distribution map, the insect control prescription map representing local insect control measures in respective portions of the target area, and

- controlling an insect activity control device to selectively perform local insect control measures in respective portions of the target area based on the created insect control prescription map.

Examples of insect control measures comprise releasing an insect control agent, in particular an insecticide and/or a biological control agent. Accordingly, the insect control prescription map may be a prescription release map indicative of local amounts and/or types of insect control agent to be sprayed or otherwise released in respective portions of the target area. Accordingly, the insect activity control device may be a sprayer configured to perform spraying responsive to the position of the sprayer within the target area. Alternatively, the insect activity control device may be another vehicle, such as framing vehicle, or other movable device for performing a suitable insect control measure.

According to another aspect, disclosed herein are embodiments of a data processing system configured to perform steps of the computer-implemented method described herein. In particular, the data processing system may have stored thereon program code adapted to cause, when executed by the data processing system, the data processing system to perform the computer-implemented steps of the method described herein. The data processing system may be embodied as a single computer or as a distributed system including multiple computers, e.g. a client-server system, a cloud based system, etc. The data processing system may include a data storage device for storing the computer program and data such as any or all of sensor data, locality information, target information, and historical data. The data processing system may directly or indirectly be communicatively coupled to the movable detection system and receive the acquired sensor data from the movable detection system. To this end, the data processing system may comprise a suitable wired or wireless communications interface.

According to another aspect, a computer program comprises program code adapted to cause, when executed by a data processing system, the data processing system to perform the computer-implemented steps of the method described herein. The computer program may be embodied as a computer-readable medium, such as a CD- ROM, DVD, optical disc, memory card, flash memory, magnetic storage device, floppy disk, hard disk, etc. having stored thereupon the computer program. According to one aspect, a computer-readable medium has stored thereupon a computer program as described herein.

According to one aspect, disclosed herein are embodiments of an apparatus for measuring insect activity in a geographic area, the apparatus comprising:

- a movable detection system comprising a first sensor and a second sensor configured to provide a first plurality of sensor data indicative of insect activity in a target area comprised in the geographic area, the first plurality of sensor data comprising a first set of sensor data from the first sensor and a second set of sensor data from the second sensor, the first set of sensor data comprising one or more digital images, the movable detection system being configured to traverse at least a portion of the target area; and

- a data processing system according to embodiments disclosed herein.

In some embodiments, the apparatus for measuring insect activity in a geographic area may be configured to autonomously or semi-autonomously control the movable detection system, or to provide instructions or guidance to an operator of a manually operated movable detection system. Some embodiments of the methods disclosed herein may comprise controlling - or providing instructions or guidance to an operator for controlling - the movement of a movable detection system, e.g. for controlling which parts of the target area to traverse by the movable detection system. In the aspects disclosed herein, terms and features relate to the terms and features having the same name in the other aspects and therefore the descriptions and explanations of terms and features given in one aspect apply, with appropriate changes, to the other aspects. Additional aspects, embodiments, features and advantages will be made apparent from the following detailed description of embodiments and with reference to the accompanying drawings.

EXAMPLES

Striped cucumber beetle (Acalymma vittatum)

The striped cucumber beetle will overwinter along the edges of the crop field, or in nearby fields, such as in nearby corn fields. During spring, adult striped cucumber beetles will fly from their wintering location into the crop field.

While within the crop field they will eat both the foliage and fruit of the crop, but they may also spread the bacterium which is the causative agent of bacterial wilt, resulting in plant death. Thus, the striped cucumber beetles can be highly damaging to crop fields and their numbers should be controlled. The flight behaviour of the adult striped cucumber beetles has been studied and one study has found that they fly once per hour per 50 insects per 4 plants, see "Lawrence, l/l/.S. and Bach, C.E., 1989. Chrysomelid beetle movements in relation to host-plant size and surrounding non-host vegetation. Ecology, 70(6), pp.1679-1690". When the striped cucumber beetles are not flying, often they are sitting on the leaves and fruits of the crop, due to their feeding and behaviour habits.

A detection system configured to record data indicative of insect activity while traversing the crop field may detect damage to leaves and/or fruit, and/or the wilting of plants, at least from the one or more digital images recorded by the first sensor. Further, adult striped cucumber beetles flying may be visible on the one or more digital images recorded by the first sensor or be detected using sensor data from the second sensor. For example, an optical insect sensor configured to be suitable for detecting flying insects may be used as the second sensor.

As the striped cucumber beetles overwinter outside of the crop field, they will have a higher presence on the field edges, where they will be noticed at first. Thus, the population numbers of the striped cucumber beetles is highly variable across a large crop field and a relevant parameter for an estimation of striped cucumber beetle activity based on sensor readings is relative to the location in which the sensor readings are made within the crop field.

The adults breed within the crop field during summer and may have 1-3 generations per growing season. Reducing the population of adults in the crop field and/or in nearby fields will lower the number of next generation striped cucumber beetles that develop, resulting in significantly reduced crop damage and increased crop production.

The striped cucumber beetles are primarily controlled using biological control agents, which target larvae growing on the roots of the crop, as well as by using foliar applied insecticides, which target adult striped cucumber beetles. Thus, one or more biological control agents and/or insecticides may be applied to the crop field inhibit the population growth or kill off adults based on a calculated measure of striped cucumber beetle activity. Having accurate knowledge of insect population and spatial dynamics can help farmers with precision targeting of the striped cucumber beetle, e.g. positioning insecticide or fungicide sprays in areas with high activity, and also potentially help identify problem areas, or areas which consistently show early signs of infestation before others.

An insect activity control device can be controlled to perform spraying responsive to the position of the sprayer within the target area substantially at the same time as insect activity measure is calculated or later. Additionally, spraying and/or one or more other suitable insect control measures may be planned for a later time based on the insect activity measure calculated based on sensor data recorded in the crop field. The insect activity control device may be the vehicle on which the detection system is mounted, or the insect activity control device may be another vehicle, such as framing vehicle, or other movable device for performing a suitable insect control measure.

Rice stink bug (Oebalus pugnax)

The rice stink bug is a pest that is known to attack cereal crops with small seeds, particularly wheat, sorghum, and rice. Feeding from both the nymphal and adult stages of the rice stink bug cause damage to the rice or wheat grains. However, the damage caused by the adult is more extensive, resulting in greater economic damage including total loss of the grains. Such damage caused to the grains is called "pecking" and may be visible as small dots/markings on the seeds.

The population rice stink bug is known to be higher on or near the edges of the crop field and the general recommendation is to sample at a distance of 50 m into the crop field from an edge in order to avoid population overestimation. However, using the methods and apparatus disclosed herein, the relative position of the sensors during recording of sensor data may be considered when calculating a resulting measure of insect activity.

For unknown reasons, the adult rice stink bugs, which are native to North America, fly often in July. Thus, a sensor suitable for detecting flying insects can be particularly useful for detecting the rice stink bug during the month of July.

There may be up to five generations annually, with two or three generations on the growing crop. Thus, there is ample time to detect the rice stink bug and spray during the earlier generations to decrease the growth of later generations. So, it is often possible to see the effect on later generations several times within the same year and thereby evaluate the effectiveness of the spraying early with the potential to adjust a spraying schedule.

The insecticides used to control the rice stink bug are relatively strong, have off-target effects, and may not be used within two weeks of harvest due to their potency. This provides a strong incentive for precision spraying, e.g. only spraying the areas with high activity, and for spraying as little as possible. Having relatively accurate knowledge of rice stink bug activity can help farmers with precision spraying and potentially help identify problem areas, or areas which consistently show early signs of infestation before others.

Southern Corn Rootworm (Diabrotica undecimpunctata)

The southern corn rootworm overwinters in the crop field and emerges in spring. It can eat the foliage of the plant as well as the silks, which are used for pollination. When silks are damaged, the corn does not produce kernels resulting in loss of yield. The leaves of the corn as well as the silks will show signs of damage if the southern corn rootworm is present. Such damage is indicative of the presence of adult southern corn rootworm.

The southern corn rootworm adult is known to fly relatively often: 5 times per hour per 50 beetles per 4 plants making it possible to detect by a sensor configured to detect flying insects.

There is a two-week window to treat the female rootworm adults before they return to the soil to lay eggs. In this period, they are often visible either on leaves or when flying around the silks.

After hatching, the southern corn rootworm larvae feed on the roots of the crop, which may cause the plant to "lodge", or tip slightly, and have bent stalks. Thus, this is an indication of the presence of southern corn rootworm larvae.

A detection system configured to record data indicative of insect activity while traversing the crop field may detect the damage to leaves and/or flying adults, and/or the tipped crop with bent stalks at least from the one or more digital images recorded by the first sensor. Further, southern corn rootworm flying may be visible on the one or more digital images recorded by the first sensor or be detected using sensor data from the second sensor. For example, an optical insect sensor configured to be suitable for detecting flying insects may be used as the second sensor.

An insect distribution map may be used to identify locations within a crop field in which it is likely that eggs are present, such that spraying with a suitable biological agent can be scheduled in these locations. Additionally, the insect distribution map may aid a farmer in the usual practice, whereby the farmer uproots a few plants to check for larvae as the insect distribution map may help narrow down locations wherein to do this.

Using historical data about previous crops in a field may facilitate in the identification and/or in the calculation of measure of insect activity of Southern Corn Rootworm as peanuts grown in rotation with corn can produce a relatively high second generation population in a corn field. 1

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments will be described in more detail in connection with the appended drawings, where

FIG. 1 shows a schematic view of an apparatus for measuring insect activity in a geographic area according to some embodiments,

FIG. 2 shows a schematic view of multiple detection systems for measuring insect activity mounted on a farming vehicle according to some embodiments,,

FIG. 3 schematically illustrates an embodiment of a data processing system according to some embodiments,

FIG. 4 shows a schematic view of an apparatus for measuring insect activity in a geographic area according to some embodiments,

FIG. 5 shows a schematic flow diagram of a computer-implemented method of determining a resulting measure of insect activity according to some embodiments,

FIG. 6 shows a schematic flow diagram of a computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area according to some embodiments,

FIG. 7 shows an illustrative graph of insect count versus time of day from the first and second sensor in the movable detection system according to some embodiments, and

FIG. 8 shows an illustrative graph of the effect of suppression of insect population numbers.

DETAILED DESCRIPTION

FIG. 1 shows a schematic view of an apparatus, generally designated by reference numeral 1, for measuring insect activity in a geographic area.

The apparatus comprises a movable detection system comprising a first sensor 7, a second sensor 9, and a data processing system 200. Each of the first and second sensors 7,9 are configured to provide sensor data indicative of insect activity. The sensors are mounted on a beam 5, such as a spray boom. The sensors provide a first plurality of sensor data, which includes a first set of sensor data from the first sensor and a second set of sensor data from the second sensor. The first sensor 7 is an image sensor, which records digital images.

The second sensor may be a sensor suitable for detecting flying insects, such as an optical insect sensor, an example of which will be described in greater detail with reference to FIG. 3 below.

The first sensor has a sensory field 11 within which it can record indications of insect activity. Similarly, the second sensor has a sensory field 13 within which it can record indications of insect activity. The sensory field associated with a sensor is preferably external to the sensor and located in the vicinity of the sensor. Insect activity indications that may be recorded include: one or more flying insects 15, one or more insects 25 sitting on the crop 23 or on the ground 27, and/or crop alteration. Examples of crop alteration include: change in physical appearance, such as of the stem 21 or one or more leaves 17 and/or damage to leaves 19. A change in physical appearance of the plant 23 could be e.g. damage to one or more leaves, and/or change in colouring of at least part of the crop, and/or change in leaf shape such as rolling of a leaf.

The sensory fields 11, 13 of the first and second sensors 7,9 are shown as extending from the sensor and towards the ground 27. However, one or both sensors may be configured such that the sensory field points in another direction. For example, the first sensor 7, which produces digital images as sensor data, may have its sensory field 11 pointing towards the ground 27 such that it is able to record insects sitting on leaves and/or alterations to the crop 23, while the second sensor 9 may be an optical insect sensor configured to have its sensory field 13, i.e. its probe volume, pointing substantially horizontally. The first and second sensory fields 11, 13 may overlap, wholly or partially, or not overlap at all.

The movable detection system 3 is intended to traverse a geographic target area in which insect activity is to be measured. To this end, the movable detection system 3 may be integrated into or mounted to a movable support, e.g. on a vehicle such as a tractor, a movable farming machine, or, as shown in fig. 1, a spray boom 5, etc. It will be appreciated that alternative embodiments may include multiple detection systems, e.g. as shown in fig. 2. It will be appreciated, that the number of detection systems may be chosen depending on factors such as the size and variability of the geographic area, a desired accuracy of a resulting measure of insect activity, a desired accuracy of spatial resolution of an insect distribution map, etc. It may also be that multiple movable detection systems traverse the geographic area or the target area, for example mounted on an autonomous vehicle or an operator-controlled vehicle, such as a remote controlled vehicle.

As the movable detection system traverses the target area, the first and second sensor 7, 9 each record data indicative of insect activity in the target area, where the target area may be an agricultural field for growing crops, an area of forest or another geographic area. The movable detection system is configured to determine a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data. Based on the first and second measure of insect activity, a resulting measure of insect activity is calculated, which is indicative of insect activity in at least part of the target area.

The movable detection system 1 is communicatively coupled to the data processing system 200 and communicates the collected sensor data, and other data, e.g. position data, to the data processing system 200. To this end, the movable detection system may include a suitable communications interface. The communications interface may be a wired or a wireless interface configured for direct or indirect communication of sensor data to the data processing system. In the example of FIG. 1, the collected sensor data is communicated via a cellular telecommunications network to the data processing system 200, e.g. via a GSM/GPRS network, USTM network, EDGE network, 4G network, 5G network or another suitable telecommunications network, such as e.g. a LoRa communication protocol. In some embodiments, the communications interface may be configured for communication via satellite. It will be appreciated that the communication may be a direct communication or via one or more intermediate nodes, e.g. via the movable support. Similarly, the communication may use alternative or additional communications technologies, e.g. other types of wireless communication and/or wired communication. For example, the communication may use a LoRa communication protocol. Yet further, the collected sensor data may be stored locally by the detection system or other for subsequent retrieval, e.g. after traversing the geographic area. To this end, the detection system or part of a vehicle to which it is mounted may include a local data storage device for logging the sensor data and for allowing the stored data to be retrievable via a data port or a removable data storage device.

It will be appreciated that the data acquisition is performed locally in the detection system. The remaining signal and data processing tasks may be distributed between the detection system 1 and the data processing system 200 in a variety of ways. For example, some or even all signal and/or data processing may be performed locally in the detection system. Similarly, some or even all signal and/or data processing tasks may be performed by the data processing system. For example, the determination of a first, second and resulting measure of insect activity from the sensor signals may be performed locally by the detection system while the creation of an insect distribution map from the resulting measure of insect activity and from sensor position information may be performed by the data processing system. Alternatively, the detection system may forward the sensor signals to the data processing system, which then performs the determination of a first, second and resulting measure of insect activity and the creation of the insect distribution map. Accordingly, depending on the distribution of processing tasks between the detection system and the data processing system, the sensor data communicated from the detection system to the data processing system may have different forms.

The detection system 1 or its movable support, e.g. the vehicle 30, may comprise a position sensor, e.g. a GPS sensor, for tracking the position of the detection system while traversing the target area. Accordingly, the detection system or the movable support may record its position a respective times, e.g. at regular time intervals, e.g. so as obtain a sequence of time-stamped position coordinates. The detection system or the movable support may further store time-stamped operational data, e.g. whether the detection system is acquiring sensor signals, one or more quality indicators of the acquired sensor signals, etc., so as to allow a determination of the actual time during which the detection system acquires usable sensor data in respective portions of the target area. The data processing system 200 may be configured, e.g. by a suitable computer program, to receive sensor data from the detection system 1 and position data from a position sensor as described above. The data processing system 200 may be configured to process the received sensor data and the received position data to create an insect distribution map as described herein. To this end, the data processing system 200 may be configured to execute a computer program for analysing the sensor data from the detection system and for creating an insect distribution map indicative of a spatial distribution of one or more desired quantities indicative of insect activity. The data processing system may output the created insect distribution map in a suitable form, e.g. on a display, on a storage device, via a data communications interface, and/or the like. The data processing system 200 may be a stand-alone computer or a system of multiple computers, e.g. a client-server system, a cloud-based system or the like. An example of a data processing system will be described in more detail below with reference to FIG. 3.

In some embodiments, the detection system 1 comprises or is communicatively coupled to one or more additional sensors, such as one or more environmental sensors for sensing environmental data, such as weather data. The one or more additional sensors may be deployed in the geographic area. Examples of environmental data include ambient temperature, humidity, amount of precipitation, wind speed, etc. The one or more additional sensors may be included in the detection system 1, in the movable support, e.g. in a vehicle, or they may be provided as a separate unit, e.g. a weather station, that may be communicatively coupled to a detection system and/or to the remote data processing system.

FIG. 2 shows a schematic view of multiple detection systems for measuring insect activity mounted on a farming vehicle 30. A plurality of detection systems mounted on the beam 5, such as a spray boom, of a tractor 30 may be used to detect indications of insect activity in an area, such as a crop field. A first sensor 7 and a second sensor 9 of a detection system may be mounted together, i.e. in relation to each other, at a predetermined distance to the other sets of sensors. For example, with a predetermined distance of 0.5 meters, or 1 meters, or 2 meters. The first and second sensor 7, 9 record data as they concurrently traverse the geographic area, while being coupled physically and having a constant distance to each other.

The detection systems have been configured such that the sensory field 11 of the first sensor 7 is shown as extending from the first sensor forward in the direction of primary movement 35 of the vehicle 30 and towards the ground and any crops in the target area that the beam 5 is adapted to be able to pass over. The sensory field 13 of the second sensor 9 is shown as extending from the second sensor 9 backwards and into the air behind the vehicle 30, as an alternative configuration to those shown in figs. 1 and 4. In this configuration, there will likely not be any overlap between the sensory fields of the first and second sensor at a given time.

The vehicle 30 may move across an entire crop field or in selected parts of a crop field. While moving in the field, the vehicle 30 may move at different speeds, may repeatedly re-visit some portions of the crop field or otherwise traverse the crop field in a non- uniform manner. The vehicle 30 may traverse the field in a stop-and-go pattern such that the sensors of the detection systems move between, and remains stationary at, different locations within the field. The sensors may detect indications of insect activity while being stationary and/or while moving between locations.

The beam may be configured as a spray boom such that spraying may be initiated, while the vehicle 30 is traversing the area. For example, spraying could be initiated shortly after a resulting measure of insect activity has been determined, for example due to the measure surpassing a predetermined threshold. Alternatively, or additionally, spraying may be initiated as part of a scheduled spray plan.

The vehicle 30 may have mounted on it a suitable communications interface through which the multiple detection systems are communicatively coupled to a data processing system 200, such as a data processing system as described elsewhere.

In some embodiments, the multiple detection systems mounted on a moving platform, such as a vehicle 30, comprise different types of sensors and/or a different number of sensors. For example, referring to fig. 2 the vehicle 30 illustrated may show six detection systems each containing a first sensor 7 and a second sensor 9 mounted on the visible part of the beam 5 (and possibly a number of detection systems may be mounted on a beam on the other, not visible, side of the vehicle). In other embodiments, a moving platform, such as a vehicle 30 as shown in fig. 2, may have one or more detection systems having more than two sensors. For example, the visible part of the beam 5 illustrated in fig. 2 may comprise a single detection system comprising a first sensor 7 and second sensor 9 as well as ten further sensors.

In some examples, one or more of the detection systems could comprise one or more additional sensors, such as an additional sensor, which records digital images.

For example, a 36 meter long spray boom could have mounted on it 36 detection systems each having a camera as first sensor and a sensor configured to detect insects in flight as second sensor, or the same spray boom could have 18 detection systems each having two cameras and a sensor configured to detect insects in flight, etc.

FIG. 3 shows a schematic view of an example of a data processing system. The data processing system 200 comprises a central processing unit 240 or other suitable processing unit. The data processing system further comprises a data storage device 230 for storing program code, received sensor data, and/or created insect distribution maps. Examples of suitable data storage devices include a hard disk, an EPROM, etc. The data processing system further comprises a data communications interface 270, e.g. a network adaptor, a GSM module or another suitable circuit for communicating via a cellular communications network or via another wireless communications technology. To this end, the data processing system may further comprise an antenna 271. It will be appreciated that the data processing system may include a wired data communications interface instead of or in addition to a wireless communication interface. The data processing system may receive sensor data from the insect sensor via one or more nodes of a communications network. The data processing system further comprises an output interface 220 e.g. a display, a data output port, or the like.

FIG. 4 shows a schematic view of a detection system for measuring insect activity in a geographic area. The detection system, generally designated by reference numeral 100, comprises an insect sensor 120, an image sensor 125, and a data processing system 200. The insect sensor 120 and the image sensor 125 are comprised within a sensor unit 128. The insect sensor may be an optical insect sensor. An optical insect sensor device may comprise an illumination module including a light source, such as one or more halogen lamps, one or more LEDs, one or more lasers, or the like, configured to illuminate a volume in a proximity of the insect sensor device. The insect sensor device may further comprise a detector module including one or more detectors and one or more optical elements configured to capture backscattered light from at least a portion of the illuminated volume and to guide the captured light onto the one or more detectors. The illuminated volume from which light is captured by the detector module for detecting insects is referred to as probe volume 150.

Generally, the probe volume 150 may be defined as the volume from which the detector module obtains light signals useful for detecting insects. The probe volume is typically defined by an overlap of the volume illuminated by the illumination module and by the field of view and depth of field of the detector module. In particular, the probe volume is not limited by any physical enclosure but is an open, unenclosed void or space that airborne, living insects may enter or exit in an unrestricted manner.

The probe volume is also the volume from which the insect sensor acquires measurements useful for detecting insects. Generally, the insect sensor 120 acquires sensor data from which insect detection events can be detected. An insect detection event refers to the detection of one or more insects being present in the probe volume 150. Detection of an insect detection event may be based on one or more criteria, e.g. based on a signal level of the detected sensor signal and/or on another property of the sensor signals sensed by the detector module of the insect sensor, e.g. in response to the received light from the probe volume.

The optical insect sensor uses reflected/backscattered light from insects in the probe volume 150 to detect insects and to measure optically detectable attributes of the detected insects, e.g. one or more of the following: one or more wing beat frequencies, a body-to-wing ratio, a melanisation ratio (colour), a detected trajectory of movement of an insect inside the probe volume, a detected speed of movement of an insect inside the probe volume, an insect glossiness, or the like. The image sensor 125 is arranged such that the field of view 122 of the image sensor overlaps with the probe volume 150.

The data processing system 200 is configured, e.g. by a suitable computer program, to receive sensor data from the insect sensor 120 and image data from the image sensor 125. The data processing system 200 may be a stand-alone computer or a system of multiple computers, e.g. a client-server system, a cloud-based system or the like. An example of a data processing system is described in more detail in the text accompanying FIG. 3.

The insect sensor 120 and/or the image sensor 125 is communicatively coupled to the data processing system 200 and can communicate acquired sensor data and/or image data to the data processing system 200. To this end, the sensor unit 128 may include a suitable communications interface. The communications interface may be a wired or a wireless interface configured for direct or indirect communication of data, such as sensor data and image data, to the data processing system. In the example of FIG. 4, the sensor unit 128 communicates the collected data via a cellular telecommunications network to the data processing system 200, e.g. via a GSM/GPRS network, USTM network, EDGE network, 4G network, 5G network or another suitable telecommunications network. In some embodiments, the communications interface may be configured for communication via satellite. It will be appreciated that the communication may be a direct communication or via one or more intermediate nodes, e.g. via a movable support. Similarly, the communication may use alternative or additional communications technologies, e.g. other types of wireless communication and/or wired communication. Yet further, the collected data may be stored locally by the sensor unit or by a movable support for subsequent retrieval from the sensor unit, e.g. after traversing a geographic area. To this end, the sensor unit or a movable support may include a local data storage device for logging the data and for allowing the stored data to be retrievable via a data port or a removable data storage device.

FIG. 5 shows a schematic flow diagram of a computer-implemented method of determining a resulting measure of insect activity according to some embodiments.

In step S51, a first plurality of sensor data from a movable detection system traversing at least a portion of a geographic target area is obtained. The movable detection system has a plurality of sensors. The first plurality of sensor data comprises a first set of sensor data from a first sensor and a second set of sensor data from a second sensor, where the first set of sensor data comprises one or more digital images.

Each set of sensor data may be associated with the current position of the detection system or the with the current position of the individual sensor, which recorded the set of sensor data. To this end, the process may further acquire sensor position data indicative of the position of the detection system and/or of the individual sensors within the target area at respective times.

The detection system communicates the sets of sensor data to a data processing system for further processing. The data processing system may be external to the detection system, e.g. as described in connection with figs. 1 or 4, or it may be integrated with the detection system.

In step S52, a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data is determined.

Based at least on sensor position data, the process may associate each of the first and second measure of insect activity with a corresponding measure position at which the sets of sensor data were recorded. Alternatively, the process may associate each of the first and second measure of insect activity with respective positions in a different manner.

Optionally, in step S63, the process of recording a plurality of sensor data is repeated by the movable detection system traversing at least a portion of the geographic target area at a later time to obtain a second plurality of sensor data. Thus, a second plurality of sensor data from a movable detection system traversing at least a portion of a geographic target area is obtained. The second plurality of sensor data comprises a third set of sensor data from the first sensor and a fourth set of sensor data from the second sensor, where the third set of sensor data comprises one or more digital images.

In step S54, a resulting measure of insect activity based at least on the first and second measure of insect activity is calculated. The resulting measure of insect activity is indicative of insect activity in at least a portion of the target area traversed by the movable detection system.

Based at least on sensor position data, the process may associate the resulting measure of insect activity with a corresponding measure position. Alternatively, the process may associate each of the resulting measure of insect activity with a position in a different manner.

Optionally, the calculation of the resulting measure of insect activity may be based on additional input. For example, on target information indicative of one or more anticipated insects in the target area, locality information relating to the locality conditions, and/or on historical data.

Optionally, in step S55, one or more local insect control measures in at least part of the target area are determined based on the resulting measure of insect activity. Insect control measures may be any or a plurality of known insect control measures, such as e.g. the release of beneficial insects and/or biological agents for reducing or suppressing undesired insect activity. For example, the process may determine that an insecticide should be sprayed due to the resulting insect activity measure being above a predetermined threshold. The spraying may be executed by the vehicle on which the detection system is mounted or by another vehicle, such as framing vehicle, or other movable device for performing a suitable insect control measure.

The type of insect control measure used may be based on an identification of one or more certain species of insect, and the process may be configured to select the type of insecticide or other insect control measure automatically so as to selectively target the identified insects.

Optionally, in step S56, an insect activity control device is controlled to perform one or more of the determined local insect control measures in the target area.

FIG. 6 shows a schematic flow diagram of a computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area according to some embodiments. In step 61, the process obtains the calculated resulting measure of insect activity based at least on a first and second, and optionally, on a third and fourth, measure of insect activity.

In step 62, an insect distribution map of the target area is created based on the resulting measure of insect activity, where the insect distribution map represents local insect activity in respective parts of the target area.

Optionally, in step 63, an insect control prescription map of the target area may be created based on the insect distribution map, where the insect control prescription map represents local insect control measures in respective portions of the target area.

The insect control prescription map may be a prescription release map indicative of local amounts and/or types of insect control agent to be sprayed or otherwise released in respective portions of the target area. The insect control prescription map may comprise a number of scheduled insect control measures to be performed in the target area.

The insect control prescription map, or prescription release map, may be created at least in part be based on a look-up table, a decision tree, a neural network, a support vector machine, and/or the like. An artificial intelligence, trained using machine learning to automatically learn from data on previous performed insect control measures, may be used to create the insect control prescription map, or prescription release map. In this manner, an Al-optimized control measure plan, such as an Al-optimized spray plan may be created by the process.

Optionally, in step 64, an insect activity control device is controlled to selectively perform local insect control measures in respective portions of the target area based on the created insect control prescription map.

The insect activity control device may be a sprayer configured to perform spraying responsive to the position of the sprayer within the target area or another vehicle suitable for performing local insect control measures.

FIG. 7 shows an illustrative graph of insect count versus time of day from the first and second sensor in the movable detection system according to some embodiments. The insect count on the vertical axis represents insect activity recorded by a sensor. Many insects fly only for short periods of time during the day and mostly during particular periods of the day such as morning, noon and/or evening. When not flying the insects are likely to sit on the crop or on the ground.

A sensor configured to obtain data indicative of an insect not in flight will primarily be able to detect such insects during the part of the day, where the insects do not fly or fly very little. An insect count from such a sensor over the course of a day is illustrated by the dash-dot-dash line 70 in fig. 7.

As the sun rises, the insects sit on leaves or the ground, while heating up before starting to take flight and they are mostly detectable by a sensor configured to obtain data indicative of an insect not in flight 70. However, during midday, where the insects are most likely to fly, that sensor will register a low insect count. Towards evening the insects stop flying as much and will again sit on the plants or ground cooling down, and the insect count from the sensor rises again. The sensor stops recording when it becomes too dark.

Likewise, a sensor configured to obtain data indicative of an insect in flight will primarily be able to detect such insects during the periods of the day, where the insect may fly. An insect count from such a sensor over the course of a day is illustrated by the dashed line 75 in fig. 7.

During morning no insect counts, or very few, are registered by the sensor configured to obtain data indicative of an insect in flight as the insects are mostly sitting on plants or the ground. As the time of day passes, the insects start taking flight and more and more insects are detected in flight until later in the day, where the insects again stop flying such that the number of insect counts again drops.

Additionally, or alternatively, one or both sensor may also be configured to detect crop alteration caused by one or more insects, which allows the sensor to detect indication of the insect activity throughout the day.

This is thus an example of how having a plurality of sensors can advantageously mitigate the limitations of a single sensor recording as basis for a calculation of a measure of insect activity in a target area. FIG. 8 shows an illustrative graph of the effect of suppression of insect population numbers on later generations. The insect activity on the vertical axis may e.g. be an actual number of insects detected, or a measure of insect activity taking into account one or more parameters.

Even when infested with one or more harmful insects, a crop field may still produce a high enough yield to be economically viable. However, at some level of harmful insect activity the yield from the crop field becomes low to sustain the crop production economically. Thus, there is a correlation between the insect activity, which may be calculated using the methods and apparatus disclosed herein and the economic viability of a crop field.

A graph of the insect activity indicated by one or more harmful insects across seasons is shown by the full line 80. In the first generation, the population numbers grows and then recedes at the end of the year. During this time, the insect activity is still low enough that its effect on the crop yield is below an economic threshold shown by a dashed line 82. The following years, the population numbers of the one or more harmful insects grow to a level that causes economic damage shown by the greyed out area 84. However, if a precision intervention according to one or more of the methods disclosed herein had been made early, e.g. at time 86, to e.g. reduce the adult population and/or kill larvae or eggs, a depressed population in the following years may be achieved. Ideally, the depressed population numbers in the following years would then remain low enough so as to be below the economic threshold 82, as represented by the dash-dot- dash line 88. To facilitate such a precision intervention accurate knowledge of insect activity is required.

Claims

1. A computer-implemented method of determining a resulting measure of insect activity in a geographic area, the method comprising:

- obtaining a first plurality of sensor data from a movable detection system comprising a plurality of sensors concurrently traversing the geographic area, each of the sensors in the plurality of sensors being coupled physically to each other, each sensor being configured to record data indicative of insect activity, the first plurality of sensor data comprising a first set of sensor data from a first sensor and a second set of sensor data from a second sensor, the first set of sensor data comprising one or more digital images, the second sensor being configured to obtain data indicative of an insect in flight;

- determining a first measure of insect activity based at least on the first set of sensor data, and a second measure of insect activity based on the second set of sensor data,

- calculating a resulting measure of insect activity based at least on the first and second measure of insect activity, the resulting measure of insect activity being indicative of insect activity in at least a portion of a target area traversed by the movable detection system, the target area being comprised in the geographic area.

2. A computer-implemented method according to claim 1, wherein each of the sensors in the plurality of sensors have a distance and/or spatial configuration relative to each other that at most varies within predetermined constraints.

3. A computer-implemented method according to any of the previous claims, wherein each of the sensors in the plurality of sensors have a constant distance and/or spatial configuration relative to each other.

4. A computer-implemented method according to any of the previous claims, wherein calculating the resulting measure of insect activity is responsive to input associated with expected insect behaviour.

5. A computer-implemented method according to any of the previous claims, wherein at least one sensor in the plurality of sensors is configured to obtain data indicative of an insect in flight, and at least one sensor in the plurality of sensors is configured to obtain data indicative of an insect not in flight and/or of crop alteration caused by one or more insects.

6. A computer-implemented method according to any of the previous claims, wherein the method further comprises obtaining a second plurality of sensor data from a movable detection system comprising a plurality of sensors, the second plurality of sensor data comprising a third set of sensor data from the first sensor and a fourth set of sensor data from the second sensor, the second plurality of sensor data being recorded at a later time during the same day or on a later day relative to the first plurality of sensor data, and the method further comprises determining a third measure of insect activity based on the third set of sensor data, and a fourth measure of insect activity based on the fourth set of sensor data, and wherein calculating a resulting measure of insect activity is further based on the third and fourth measure of insect activity.

7. A computer-implemented method according to any of the previous claims, wherein the method further comprises obtaining target information indicative of one or more anticipated insects in the target area, and wherein calculating a resulting measure of insect activity is responsive to the target information.

8. A computer-implemented method according to any of the previous claims, wherein the method further comprises obtaining locality information relating to the locality conditions such as e.g. degree days, time of day, date, GPS location, weather information, speed at which the movable detection system was moving during the recording of the sensor data, relative position, etc., and wherein calculating a resulting measure of insect activity is responsive to the local information.

9. A computer-implemented method according to any of the previous claims, wherein the method further comprises obtaining historical data, and wherein calculating a resulting measure of insect activity is responsive to the obtained historical data.

10. A computer-implemented method according to any of the previous claims, wherein the method further comprises determining one or more local insect control measures in at least part of the target area based on the resulting measure of insect activity.

11. A computer-implemented method according to claim 8, wherein the method further comprises controlling an insect activity control device to perform one or more of the determined local insect control measures in the target area.

12. A computer-implemented method for creating an insect distribution map indicative of insect activity in a geographic area, the method comprising:

13. A data processing system configured to perform the computer- implemented steps of the method according to any one of the previous claims.

14. A computer program product comprising computer program code configured, when executed by a data processing system, to cause the data processing system to perform the computer-implemented steps of the method according to any one of claims 1 through 12.

15. An apparatus for measuring insect activity in a geographic area, the apparatus comprising:

- a movable detection system comprising a first sensor and a second sensor configured to provide a first plurality of sensor data indicative of insect activity in a target area comprised in the geographic area, the first plurality of sensor data comprising a first set of sensor data from the first sensor and a second set of sensor data from the second sensor, the first set of sensor data comprising one or more digital images, the movable detection system being configured to traverse at least a portion of the target area;

- a data processing system according to claim 13.