Monitoring module

Abstract

Monitoring module (1) for pest control having at least one camera (2) and evaluation electronics (3), which are configured to determine a status of a pest trap (4) and / or of an environment (5) of the monitoring module (1), wherein the evaluation electronics (3) are configured to infer the status of the pest trap (4) and / or the status of the environment (5) from image data generated by the camera (2) by executing an algorithm.

Description

[0001] 3370841 / 32

[0002] 1

[0003] Monitoring module

[0004] The present disclosure relates to a monitoring module according to the features of the preamble of claim 1.

[0005] Snap traps for rodents and the like are known to use a magnetic field-based monitoring module to check whether the trap is still in the tensioned state or not.

[0006] Additionally, versions have become known in which an additional magnetic field sensor can be used to determine whether the trap is empty or not.

[0007] It is impossible to determine whether a captured animal is still alive or not.

[0008] Solutions exist where images of the trap are taken and these images are then transmitted to a higher-level monitoring system (also referred to as a secondary communication system). The images can be captured cyclically or triggered by passive infrared sensors. A disadvantage of this approach is the relatively high bandwidth required for image transmission. While this isn't inherently problematic, communication modules capable of such high bandwidths also consume a significant amount of energy, resulting in either short transmission times or a correspondingly large power supply.

[0009] The object of the present disclosure is to implement a monitoring system for pest control which is characterized, on the one hand, by very long operating times, preferably at least one year, and, on the other hand, by the reliable determination of the status of the pest trap. This object is achieved by the features of claim 1 in that a monitoring module for pest control is provided with at least one camera and evaluation electronics which are configured to determine the status of a pest trap and / or the environment of the monitoring module, wherein the evaluation electronics are configured to infer the status of the pest trap and / or the status of the environment from image data generated by the camera by processing an algorithm.

[0010] For the purposes of this document, pests are defined as animal and other organisms whose presence is undesirable because they are harmful or disruptive to humans. Examples include rodents such as mice and / or rats, certain insects, and / or mold.

[0011] The monitoring module according to the present disclosure makes it possible to eliminate the need for manual checks of the trap's status and / or the monitoring module's surroundings. This remote monitoring could therefore save the user from having to make regular checks, which in turn could have a positive effect on CO₂ emissions.

[0012] Image data within the meaning of this disclosure can be data produced by the camera. It can consist of single images or sequences of images using a wide variety of known encoding methods.

[0013] Preferably, the pest trap should not use any toxic substances. It is conceivable that the pest trap could be suitable for various locations. Examples include use in pipes; outdoors; in buildings; and in buildings to monitor hard-to-reach areas.

[0014] The monitoring module can have a length in the range of 4 cm to 7.5 cm, particularly preferably of about 6 cm.

[0015] The monitoring module can have a width in the range of 4 cm to 5.4 cm, preferably of about 5.4 cm.

[0016] The monitoring module can have a first height (in the area without a camera) in a range of 1.5 cm to 2.5 cm, particularly preferably in a range of 2 cm and 2.5 cm.

[0017] The monitoring module can have a second height (in the area of ​​the camera) in a range of 2.5 cm to 3.5 cm, particularly preferably in a range of 2.5 and 3 cm.

[0018] The length can be understood as the longest dimension of any side of the monitoring module.

[0019] The first height can be the shortest dimension of the monitoring module.

[0020] Preferably, the algorithm includes at least one object detection algorithm and / or at least one object classification algorithm and distinguishes at least one, preferably all, of the following states of the pest trap: trap operational; trap triggered but empty; trap triggered, pest dead; trap triggered, pest caught alive. Thus, the algorithm can detect whether the trap is operational; the trap is triggered but empty; the trap is triggered and the pest is dead; or the trap is triggered and the pest is caught alive.

[0021] In other preferred embodiments, the status of the snap trap may not be determined within the monitoring module, or only the triggered or untriggered status may be determined, and subsequently an image (the image data) in segmented form is sent to a server, and the status is determined by an operator based on the transmitted image or by the server. In these embodiments, the server assumes at least part of the role of the evaluation electronics.

[0022] The techniques presented here, especially those using artificial intelligence for evaluating image data, can all (and preferably) be implemented using an algorithm running on the server.

[0023] Preferably, at least one object detection algorithm and / or at least one object classification algorithm should be trained using the Edge Impulse platform.

[0024] It's possible that the image data used to train the algorithms originates from a development phase where real-world data from customer environments was collected using dummies. It's conceivable that the images for training could be processed in a standardized RGB format with a resolution of 96x96 pixels.

[0025] Preferably, the object classification algorithm should be able to distinguish between at least one, preferably all, of the following states of the pest trap: trap operational; trap triggered but empty; trap triggered, pest dead; trap triggered, pest caught alive. Preferably, the object classification algorithm is based on a convolutional neural network that operates with multiple convolution and pooling layers.

[0026] The input layer is designed to process features, followed by two convolutional layers with different filter sizes to extract relevant features. A subsequent flattening layer can convert the features into a dense representation. This is necessary to make the extracted features usable for the following fully connected dense layers. The classification module can consist of two dense layers with 64 and 128 neurons respectively, as well as a dropout layer to reduce overfitting. The final output layer can determine the trap status based on the learned features.

[0027] The model can preferably be trained over 35 training cycles with a learning rate of 0.0003. 20% of the data could be used for validation, while the batch size can be 32. To maximize efficiency on embedded systems, the model can be quantized to int8. After training is complete, it could be exported as a TensorFlow Lite model and provided as an Arduino library for easy integration into the target hardware.

[0028] In addition to classification, the object detection algorithm (technical term: "object detection model") can extend its functionality by identifying and locating objects within the field of view. This model can also be trained to recognize the three possible states of a mousetrap. The architecture could be based on FOMO (Faster Objects, More Objects) MobileNetV2 0.35, which enables efficient object detection with minimal computational load. This is essential for resource-constrained devices like the ESP32S3.

[0029] Training can preferably be performed over 55 cycles with a learning rate of 0.001. Enabling data augmentation could make the model more resilient to variations. 20% of the data could serve as a validation set, and the batch size could also be 32. Again, the model could be quantized to int8 and exported as a TensorFlow Lite model after training is complete. Deployment could be done as an Arduino library to allow seamless integration into existing embedded systems.

[0030] The two algorithms—at least one for object recognition and / or at least one for object classification—are preferably not used simultaneously. Only one of the two algorithms is ever in use per AI chip.

[0031] The algorithms can be specifically optimized for energy-efficient IoT devices and could be run with high efficiency on edge devices such as the ESP32-S3. Detection could occur autonomously without a permanent cloud connection, creating a reliable solution for industrial and consumer applications.

[0032] Preferably, the algorithm should be an artificial intelligence. It is possible that the artificial intelligence comprises at least one algorithm. For example, at least one machine learning algorithm and / or at least one neural network could be used.

[0033] Preferably, the algorithm, and especially all algorithms, can be a supervised learning algorithm. It is equally conceivable that it could be an unsupervised learning algorithm.

[0034] Furthermore, it is possible that the algorithm, preferably all algorithms, is trained using the same and / or different image data.

[0035] It is therefore possible that the algorithm, preferably all algorithms, is trained using the image data generated by the camera. More specifically, it is conceivable that the algorithm, preferably all algorithms, can, through continuous training, detect at least one, preferably all, of the following states of the pest trap: trap operational; trap triggered but empty; trap triggered, pest dead; trap triggered, pest caught alive.

[0036] In particularly preferred embodiments, the monitoring module is designed to distinguish between the status "trap triggered, pest dead" and the status "trap triggered, pest caught alive." When a pest is caught alive, a very rapid response is often necessary. This is generally not required when a pest is dead. By differentiating between these situations, the monitoring module ensures that monitoring personnel can react quickly to a live-captured pest without having to constantly check triggered traps with short reaction times.

[0037] Training data for the algorithm can preferably be obtained by generating image data of real traps or model traps. This data can then be used in the usual way to train an image recognition intelligence.

[0038] This means that the training data can consist of a database of image data, where a known status is stored for each image or sequence of images, which is used for training.

[0039] As a result, the algorithm will then output the desired correct status of the pest trap or the environment of the pest trap when newly generated image data is used as input.

[0040] It is also conceivable that the algorithm, or preferably all algorithms, is not trained continuously, but rather that training is stopped at a certain point. Subsequently, further "retraining" (technical term: "fine-tuning") can be performed.

[0041] In a narrower sense, it is therefore conceivable that the algorithm, preferably all algorithms, could, through continuous training, detect at least one state of the pest trap more efficiently and accurately. It is equally conceivable that the algorithm, preferably all algorithms, is not trained continuously, but rather is set to a predetermined training level upon reaching that level.

[0042] It is possible that at least one sensor is provided which is signal-connected to the evaluation electronics, the evaluation electronics being designed to activate the camera if sensor data from the at least one sensor meets at least one trigger criterion.

[0043] In the description of the present disclosure, a "trigger criterion" can be a feature that the sensor detects. Preferably, a "trigger criterion" is the feature that causes the sensor to activate the camera.

[0044] For example, the trigger criterion can be implemented as at least one threshold value, whereby triggering occurs as soon as the signal detected by at least one sensor exceeds the threshold value.

[0045] The evaluation electronics may include at least two processors, wherein a first processor is configured to put a second processor into a wake-up state if the sensor data of the at least one sensor fulfills at least one trigger criterion, and wherein the second processor in the wake-up state is configured to control the camera to generate image data and to evaluate the generated image data by processing the algorithm, preferably to infer the status of the pest trap.

[0046] The term "awake state" in the present disclosure can be understood to mean that the second processor is switched on, or in other words activated, from a standby mode and / or deep sleep mode. Of course, it is also possible for the second processor to be switched from an activated state to a standby or deep sleep mode.

[0047] In other words, this could mean that a first processor can activate a second processor from a sleep or deep sleep mode. Once activated, the second processor can control the camera, which then generates image data. This image data can be analyzed by processing the algorithm to determine the status of the pest trap.

[0048] Embodiments in which the first processor is configured to put the second processor into the wake-up state if the sensor data of the at least one sensor fulfills the at least one trigger criterion, and in which the second processor, in the wake-up state, is configured to control the camera to generate image data and to evaluate the generated image data by processing the algorithm, preferably to infer the status of the pest trap, can be particularly preferred because they enable particularly energy-efficient monitoring of the pest trap. Since the second processor, which generally has higher energy consumption, is not in the wake-up state most of the time, its energy consumption is minimized. During the short periods when the trigger criterion is fulfilled, for example, because a pest has been detected in the trap, the computing capacity of the second processor is then available.This allows for exceptionally long operating or battery life and reliable monitoring simultaneously. Human intervention or control during the period in which the pest trap is in use is thus eliminated. It is particularly preferred that the evaluation of the image data by processing the algorithm, preferably all algorithms, can be carried out directly in the monitoring module using at least one microcontroller and / or processing core. More detailed explanations of the procedure are described in the disclosure.

[0049] One advantage of this is that no cloud connection or similar is necessary, which promotes data privacy and helps avoid hacker attacks.

[0050] The microcontroller architecture can include at least one microcontroller development board.

[0051] The microcontroller development board can include an ESP32-S3 microcontroller.

[0052] In the present description, the microcontroller can also be understood as an AI chip (“Artificial Intelligence” chip).

[0053] For the sake of completeness, it should be mentioned that the term AI chip is to be understood broadly in the present disclosure.

[0054] The microcontroller could, for example, be the ESP32-S2 model. Preferably, the processing core is part of the microcontroller.

[0055] An ESP32-S3 is generally understood to be a microcontroller suitable for IoT applications, for example. It can be used in applications requiring high computing power, connectivity, and / or the integration of at least one artificial intelligence. The first communication system within the monitoring module comprises at least one sensor, preferably at least one accelerometer and / or at least one passive infrared sensor and / or at least one temperature sensor and / or at least one humidity sensor, and / or at least one deep sleep chip and / or one camera and / or at least one on / off switch and / or at least one reset button and / or at least one maintenance button and / or at least one AI chip, wherein the at least one AI chip and / or the at least one processing core is configured to:

[0056] - to receive and / or evaluate the sensor data detected by the at least one sensor, preferably the at least one accelerometer and / or the at least one passive infrared sensor and / or the at least one temperature sensor and / or the at least one humidity sensor; and / or

[0057] - to receive and / or evaluate the image data generated by the camera; and / or

[0058] to receive and / or evaluate at least one signal from the on / off button to switch the monitoring module on and / or off; and / or

[0059] to receive at least one signal from the deep sleep chip; and / or

[0060] to receive at least a signal from the reset button; and / or

[0061] to receive at least a signal from the maintenance button. It is possible that, in addition to the disclosed algorithm, the pest trap includes a magnetic tape sensor, as is already known in the prior art.

[0062] Preferably, the second processor should have at least one processing core optimized for evaluating neural networks.

[0063] Accordingly, the term AI chip in this disclosure is to be understood broadly. It can also refer to a chip with a processing core for machine learning and / or deep learning and / or neural networks.

[0064] The term "processing kernel" can refer to a part of a system responsible for calculations. A processing kernel can process data, execute algorithms, and solve tasks such as simulations, data analyses, and / or mathematical models.

[0065] The processing core can be part of a microcontroller architecture.

[0066] Preferably, the AI ​​chip and / or microcontroller etc. can include the ESP32-S2 module.

[0067] A computing core specifically optimized for evaluating neural networks can also be called a Tensor Processing Unit (TPU), enabling the efficient solution of computationally intensive tasks of neural networks.

[0068] It is therefore conceivable that the second processor includes a neural network with which the second processor can evaluate the image data generated by the camera and / or infer the status of the pest trap.

[0069] In preferred embodiments, the monitoring module, in particular the evaluation electronics, is designed to segment the image data generated by the camera into a multitude of individual image sections, preferably according to a predetermined pattern.

[0070] The communication module is designed to transmit the image sections to a server located separately from the monitoring module.

[0071] It is particularly preferred that at least an initial message be transmitted to the server, the initial message comprising:

[0072] at least one image section number,

[0073] - a number of image sections to be transmitted, at least a data overview of the trigger criterion from the at least one sensor, preferably the motion sensor, the data overview particularly preferably shows the trigger criterion, preferably in this case the movement behavior of the pest, before and / or at the time of triggering the sensor.

[0074] Particularly preferred is an analysis of at least one trigger criterion.

[0075] Transmitting image sections in chunks can also be referred to as image transmission in "chunks." A specialist may also call this process "chunked transfer encoding." Each image section preferably contains an image identifier and a segment identifier (also called image section identifier or chunk ID), which defines the position of the respective image section within the entire image dataset.

[0076] It is particularly preferred that the image data generated by the camera is not transmitted to the server as a single image file, but in a multitude of individual image sections.

[0077] Particularly preferred is the segment identification number designed in such a way as to enable server-side reconstruction of the image data generated by the camera by appropriate merging, preferably according to a predefined pattern, of the received individual image sections.

[0078] The individual image segments are preferably transmitted individually and, more preferably, sequentially to the server-side receiving unit. In particular, the transmission takes place in a pattern defined by the segment identification number, whereby the image segments can, more preferably, be transmitted independently of one another.

[0079] It is possible that the server, which is arranged separately from the monitoring module, is designed to reconstruct the image data generated by the camera at least partially, preferably completely, using the image sections.

[0080] It may be possible to use a further algorithm on the server to analyze the image data generated by the camera, which can be reconstructed at least partially, preferably completely, using the image sections, particularly preferably to perform a validation and / or re-evaluation of the image data generated by the camera.

[0081] After transmitting the individual, preferably all, image excerpts, a final message is preferably sent to the server.

[0082] The final message preferably includes sensor data from at least one sensor, particularly preferably the accelerometer, acquired during the transmission phase.

[0083] This sensor data represents, in particular, the movement behavior of the pest according to the trigger criterion and during the transmission of the image sections, and enables further server-side analysis of the individual states and / or the movement state of the pest in connection with the image data generated by the camera.

[0084] The sensor data transmitted in the initial and final messages, especially data from the motion sensor, are particularly favored for comparison in order to detect a temporally correlated movement pattern.

[0085] The term “transmission phase” can be defined in the present disclosure as a phase in which the transmission of the individual image sections takes place.

[0086] The transmission phase is particularly preferred – a phase in which the camera does not generate image data.

[0087] For the sake of completeness, it should be mentioned again that the transmission of the individual image sections to the server, which is located separately from the monitoring module, is preferably carried out using the "Long Range Wide Area Network" (LoRaWAN) standard.

[0088] It is possible that at least one sensor includes an accelerometer and / or a passive infrared sensor and / or a temperature sensor and / or a humidity sensor.

[0089] An accelerometer can be understood as a device that measures the acceleration an object experiences. It can mean that the accelerometer measures the acceleration experienced by the pest, preferably the rodent and / or insect.

[0090] The accelerometer can transmit the recorded sensor data to the first processor.

[0091] In a particularly preferred embodiment, it can be provided that acceleration data is acquired by means of an accelerometer, preferably while the segmented image data is being transmitted.

[0092] The acceleration data can then be used by the monitoring module itself to determine the status of the snap trap.

[0093] Alternatively or additionally, the acceleration data can be sent to the server, preferably after the transmission of the segmented image data, and the acceleration data can then be used by the server itself to determine the status of the snap trap. A passive infrared sensor can be understood as a device that can measure infrared radiation, i.e., thermal radiation, in its environment. The sensor is described as passive because it does not emit radiation but can only detect the thermal radiation emitted by objects, in a preferred embodiment, rodents.

[0094] It may be a cumulatively triggered passive infrared sensor that is only activated when the accelerometer meets a trigger criterion.

[0095] It is also conceivable that the passive infrared sensor and the accelerometer are permanently activated.

[0096] A temperature sensor can be understood as a device that can measure the temperature of its surroundings. This can be particularly important when dealing with pests, especially mold and similar organisms, which reproduce better and / or faster within a specific temperature range.

[0097] However, it is also conceivable that the temperature of the environment is important information for other pests, such as rodents, as it affects the behavior, for example, the reproduction, of the pests.

[0098] A humidity sensor can be understood as a device that can measure the humidity in its environment.

[0099] It is conceivable that the user will receive a notification when a threshold value related to temperature and / or humidity is reached. The evaluation electronics may be designed to send readings from the temperature and / or humidity sensor to a higher-level monitoring system for the purpose of mold monitoring.

[0100] The term “higher-level monitoring system” in this disclosure can also be understood as a “second communication system”. For example, it could be a system belonging to the operator of the pest trap or the monitoring module.

[0101] The overarching monitoring system is particularly well-suited for monitoring a large number of pest traps and / or monitoring modules.

[0102] It is also conceivable that the user receives a notification as soon as the temperature and / or humidity deviates from the reference value for a period of time, preferably at least one day and / or at least one week and / or at least one month.

[0103] Preferably, the battery status should be transmitted to the user daily and / or weekly and / or monthly.

[0104] The monitoring module may be battery-powered.

[0105] Preferably, the rechargeable battery may include a lithium-ion or nickel-metal hydride (NiMH) battery.

[0106] The battery may be removable from the monitoring module at any time and reusable, preferably without damage. The monitoring module may also include a port through which the batteries can be recharged without removing them from the module.

[0107] Preferably, it is intended that by using a deep sleep mode and optimized energy consumption, the monitoring module can be operated for several months, preferably up to twelve months, using the same battery.

[0108] Since the second processor is only activated when the sensor data of at least one sensor meets at least one trigger criterion, it is conceivable, in a narrower sense, that the second processor is only activated when needed, namely only when at least one trigger criterion is met, which leads to a reduction in power consumption.

[0109] Preferably, a communication module is provided which is designed to transmit the specific statuses of the pest trap to a higher-level monitoring system, preferably to trigger a notification to an operator of the pest trap.

[0110] In other words, it can mean using a communication module to provide the user of the pest trap with the specific statuses of the pest trap.

[0111] It is possible that the communication module is designed to communicate using the "Long Range Wide Area Network" (LoraWAN) standard.

[0112] The term "Long Range Wide Area Network" (LoraWAN) can be understood as a wireless communication protocol and network protocol for IoT (Internet of Things) devices, based on LoRa (Long Range) radio technology.

[0113] It can be used specifically for deployment in wide areas with low energy consumption and can enable IoT devices to send and receive data over long distances.

[0114] It is therefore conceivable that the LoRaWAN standard enables communication between devices over a long range, preferably 0.001 to 10 kilometers, and particularly preferably 0.5 to 1 kilometer. Of course, it is also possible that the range could be more than 10 kilometers and / or less than 10 kilometers with appropriate adjustments.

[0115] Even a single gateway can cover many pest traps over a wide radius, which can keep costs down.

[0116] It is conceivable to use at least one LoRaWAN gateway. Using more than one LoRaWAN gateway can have the advantage of covering a large area and resulting in less data loss.

[0117] It is possible that an additional gateway is present, which is connected to the monitoring module on the one hand and to the internet on the other.

[0118] LoRaWAN can be used with or without a gateway.

[0119] It can be a private or public LoRaWAN. Other standards can also be used instead of LoRaWAN. Centralized control via platforms such as Azure IoT Hub may enable continuous monitoring of the malware trap's status.

[0120] Preferably, the second processor is configured to transmit the status of the pest trap to the first processor, and the first processor is configured to send the status to the higher-level monitoring system via the communication module.

[0121] In addition to the monitoring module, a pest infestation is also protected with a monitoring module as disclosed.

[0122] The pest trap may contain at least one snap trap and / or at least one sticky trap.

[0123] Preferably, rodents should be caught using at least one snap trap. Likewise, it is preferred to catch insects, such as moths, using at least one sticky trap.

[0124] A sticky trap can contain typical substances to attract insects, such as moths. These insects can then stick to the sticky trap.

[0125] It is conceivable that the camera system includes a high-resolution camera and detects the sticky trap, in particular the at least one insect stuck to it, and generates corresponding image data. It is also conceivable that the system notifies the user when a certain threshold of insects is present on the sticky trap. This threshold can be individually set by the user. Alternatively, the accelerometer could detect the insects and transmit the sensor data to the evaluation electronics, preferably to the first processor.

[0126] In addition to the disclosure of the monitoring module itself, the disclosure also relates to a monitoring system with one or more pest traps according to the invention, preferably wherein a gateway is additionally provided which is signal-connected to the monitoring module on the one hand and to the Internet on the other.

[0127] The disclosure preferably relates to a surveillance system with a server arranged separately from the surveillance module, which is configured to reconstruct the image data generated by the camera at least partially, preferably completely, using the image sections.

[0128] The disclosure relates primarily to a surveillance system, wherein the server and / or at least one other server is trained to infer the status of the trap from the reconstructed image data.

[0129] The status of the snap trap may preferably include at least one, preferably all, of the following: trap ready for operation; trap triggered but empty; trap triggered, pest dead; trap triggered, pest caught alive.

[0130] In addition to the disclosure of the monitoring module itself, the disclosure also relates to a monitoring method for monitoring a pest trap, wherein image data generated in the area of ​​a pest trap and / or its surroundings are processed by an algorithm, thereby drawing conclusions about the status of the pest trap and / or its surroundings. Protection is also sought for the use of a monitoring module according to the present disclosure in a method according to the present disclosure.

[0131] Protection is also sought for a pest trap with a monitoring module.

[0132] It is also possible, of course, that the monitoring module is arranged separately from the pest trap.

[0133] Protection is also sought for the use of a pest trap in a process according to the present disclosure.

[0134] Protection is also sought for a computer program product, comprising commands which, when the program is executed by a computer, cause it to process image data generated in the area of ​​a pest trap and / or an environment of the pest trap using an algorithm, and thereby to infer the status of the pest trap and / or the environment of the pest trap.

[0135] Protection is also sought for a transitory or non-transient computer-readable storage medium on which the computer program product is stored according to the present disclosure.

[0136] Further features and details of preferred embodiments of the disclosure are explained by way of example in the following figure description. Figures 1 and 1a show: a representation of a monitoring module according to the disclosure and a pest trap; Figure 2 shows a representation of the monitoring module according to Figure 3.

[0137] 1 with the camera;

[0138] Fig. 3-4: a side view of the monitoring module according to

[0139] Fig. 1;

[0140] Fig. 5-6: an exploded view of individual components of the monitoring module according to Fig. 1;

[0141] Fig. 7-8: a bottom view of a component according to Fig. 5-6; Fig. 9: a side view of components according to Fig. 5-6; Fig. 10: a top view of a component according to Fig. 5-6; Fig. 11: a schematic representation of a monitoring method with the monitoring module according to the disclosure.

[0142] Fig. 1 shows a monitoring module 1 with a pest trap 4, wherein the pest trap 4 is at least partially enclosed by a pest trap housing upper part 4a and a pest trap housing lower part 4b.

[0143] The pest trap 4 in Fig. 1 is a snap trap. It is possible that rodents running into the pest trap 4 from the surrounding area 5 will be killed by this pest trap 4. It is therefore conceivable that rodents could run into the pest trap 4 through a first opening 4c and / or a second opening 4d and thus be caught and / or killed.

[0144] The pest trap 4 can also be a glue trap for insects. The monitoring module 1 is attached to the upper part 4a of the pest trap housing in Fig. 1. It is conceivable that the monitoring module 1 can be removed from the upper part 4a of the pest trap housing at any time.

[0145] The monitoring module 1 may be attached to the upper part of the pest trap housing 4a via a plug connection. It is also conceivable that the monitoring module 1 is screwed or glued to the upper part of the pest trap housing 4a.

[0146] It is therefore conceivable that the upper part of the pest trap housing 4a has a recess (not shown) through which a camera 5 is inserted and can generate image data about the contents of the pest trap 4, which can be processed by means of at least one algorithm in order to infer at least one status of the pest trap 4.

[0147] As disclosed, it is intended to use image data generated by camera 2 to infer the status of the pest trap 4 and / or environment 5 by processing an algorithm.

[0148] For the sake of completeness, it should be mentioned that the pest trap housing upper part 4a in Fig. aa is a partially cutaway perspective.

[0149] It is also possible, of course, that the monitoring module 1 is arranged separately from the pest trap 4.

[0150] Fig. 2 shows the monitoring module 1 according to the disclosure with the camera 2. It is conceivable that the evaluation electronics 3 are designed to determine a status of the pest trap 4 and / or the environment 5 from Fig. 1.

[0151] Furthermore, the evaluation electronics 3 may be designed to use image data generated by the camera 2 to infer the status of the pest trap 4 and / or the environment 5 by processing an algorithm.

[0152] Fig. 2 shows that the monitoring module 1 has a monitoring module housing 9, wherein the monitoring module housing 9 comprises a monitoring module housing upper part 10 and a monitoring module housing lower part 11.

[0153] Fig. 2 shows that a second housing 24 is arranged on the monitoring module 1 along a position on the underside 11a of the lower part of the monitoring module housing 11. It can be seen that the second housing 24 is closed with a housing cover 25.

[0154] The housing cover 25 includes a recess through which the camera 2 protrudes and can generate image data of the snap trap and / or environment 5.

[0155] It is therefore conceivable that the upper part of the pest trap housing 4a has a recess corresponding to the circumference of the camera module 2a. The camera module 2a can include, in addition to the camera 2 itself, a plate 2b on which the camera 2 is mounted.

[0156] Figures 3 and 4 each show a side view of the monitoring module 1. Figure 3 shows the on / off switch 19. It is therefore conceivable that the monitoring module 1 is switched on and off using the on / off switch 19. The lower part 11 of the monitoring module housing can have a recess through which the on / off switch 19 protrudes. It is therefore conceivable that the on / off switch 19 is arranged at a position along the circuit board 12.

[0157] Fig. 3 shows the width 20 of the monitoring module 1. The width 20 can have a value in the range of 5 cm to 5.4 cm, preferably 5.4 cm.

[0158] Fig. 4 shows the length 21 of the monitoring module 1. The length 21 can have a value in the range of 4 cm to 10 cm, preferably 9 cm. Fig. 4 shows a first height 22 and a second height 23 of the monitoring module 1.

[0159] The first height 22 includes the monitoring module 1. The first height 22 can have a value in the range of 1.5 cm to 3 cm, preferably 2.5 cm.

[0160] The second height 23, which results from the monitoring module housing 9, the second housing 24, and the camera 2. The second height 23 can have a value in the range of 2.5 cm to 4 cm, preferably 4 cm.

[0161] Fig. 5 shows an exploded view of the monitoring module 1. It shows that the monitoring module housing 9 includes the circuit board 12 and the camera module 2a with the plate 2b and the camera 2.

[0162] It can be seen that the plate 2b is connected to the circuit board 12 via a second connector 15. The second connector 15 can include a cable connection 16.

[0163] Preferably, it comprises a connecting element 15 configured to transmit information and / or electrical current. For simplicity, the printed circuit board 12 is shown without electrical components in Figures 5 and 6. The individual electrical components of the printed circuit board are shown and described in Figures 7 to 10.

[0164] It is conceivable that the monitoring housing upper part 11 and the monitoring housing lower part are connected to each other by means of a first connecting means 13 (not shown), preferably a screw connection 14 or a tool-free quick connection.

[0165] Fig. 6 shows that the camera 2 is arranged on the plate 2b. The plate 2b is in turn connected to the circuit board 12. The camera 2 and the circuit board 12 can therefore be indirectly connected to each other.

[0166] For the sake of completeness, it should be noted that the underside of the circuit board 12 is the side opposite the top side of the circuit board 12. The underside of the circuit board 12 is preferably the side opposite the pest trap 4.

[0167] In other words, it can be said that the underside of the circuit board 12 includes the camera 2, whereby image data generated by the camera 2 is used to infer the status of the pest trap 4 and / or the environment 5 by processing an algorithm.

[0168] Figures 7 and 8 show that at least one sensor 6 is arranged along one position of the circuit board 12. Figure 7 shows that at least one LED module 8 is arranged at one position along the plate 2b.

[0169] Figs. 7 and 8 show underside views of the circuit board 12 of the monitoring module 1 according to the disclosure from different perspectives.

[0170] Fig. 7 shows a top view from a bird's-eye perspective of the printed circuit board 12. The individual components of the printed circuit board 12 can also be seen.

[0171] Fig. 7 shows an LED module 8. It may be that this can provide sufficient photo quality in poor lighting conditions.

[0172] For example, it could be 250mA per LED module 8.

[0173] It is possible that the LED module 8 will automatically switch on shortly before the image data is generated and switch off automatically again after the image data has been generated.

[0174] It can be seen that an accelerometer 6a is arranged.

[0175] Fig. 10 shows a top view of the top surface of the printed circuit board 12. It can be seen that the top surface of the printed circuit board 12 includes at least one sensor 6. The passive infrared sensor 6b and / or the temperature sensor 6c and the humidity sensor 6d may be arranged along one position on the top surface of the printed circuit board 12. The printed circuit board 12 shows an antenna 18. This may be a high-frequency antenna, as it can enable devices to communicate over long distances.

[0176] Preferably, the second processor is configured to transmit the status of the pest trap 4 to the first processor, and the first processor is configured to send the status to the higher-level monitoring system via the communication module 7.

[0177] Fig. 11 illustrates the procedure or the individual communication steps between the user and the monitoring module 1.

[0178] The IoT Portal (English technical term: “Internet of Things”) generally describes a network of interconnected objects and / or, preferably internet-enabled, devices.

[0179] In the course of the present disclosure, the term “IoT Portal” may refer to networked devices which have at least one sensor 6 and / or at least one software and / or at least one other technology to transmit and / or receive data to and from other devices.

[0180] It is conceivable that this transmission (which in the context of this disclosure can also be referred to as communication or connectivity) could be enabled via standard Wi-Fi and / or LoRaWAN. Therefore, it is possible that at least one LoRaWAN gateway and / or at least one Wi-Fi router would be present.

[0181] The present disclosure distinguishes between a first communication system 26 within the monitoring module 1 and a second communication system 27 outside the monitoring module 1.

[0182] In other words, the monitoring module 1 does not include the second communication system 27, but it does not include a communication module 7, which is designed to communicate using the LoRaWAN and / or WLAN standard.

[0183] The first communication system 26 within the monitoring module 1 comprises at least one sensor 6, preferably at least one accelerometer 6a and / or at least one passive infrared sensor 6b and / or at least one temperature sensor 6c and / or at least one humidity sensor 6d, and / or at least one deep sleep chip and / or one camera 2 and / or at least one on / off button 19 and / or at least one reset button and / or at least one maintenance button and / or at least one AI chip, wherein the at least one AI chip 17 and / or the at least one processing core is configured as follows:

[0184] - to receive and / or evaluate the sensor data detected by the at least one sensor 6, preferably the at least one accelerometer 6a and / or the at least one passive infrared sensor 6b and / or the at least one temperature sensor 6c and / or the at least one humidity sensor 6d; and / or

[0185] - to receive and / or evaluate the image data generated by camera 2; and / or

[0186] to receive and / or evaluate at least one signal from the on / off switch 19, for switching the monitoring module 1 on and / or off; and / or to receive at least one signal from the deep sleep chip; and / or

[0187] to receive at least a signal from the reset button; and / or

[0188] to receive at least a signal from the maintenance button.

[0189] Furthermore, it is conceivable that the AI ​​chip 17 is designed to receive detected sensor data and / or to process the image data generated by the camera 2 and / or to process this data with at least one algorithm and / or to transmit it to the second communication system 27 outside the monitoring module 1 by means of a communication module 7.

[0190] The first communication system 26 can be implemented as a software module running on the first processor. Of course, the communication system 26 can also be implemented as a hardware module. Hybrid forms are also conceivable, where some functions are handled by a software module and others by a hardware module.

[0191] It is also possible that the first communication system 26 operates autonomously if it sends directly to the IoT portal as a GSM module and e-SIM without the second communication system 27.

[0192] The second communication system 27, which is located outside the monitoring module 1, comprises at least one LoRaWAN gateway and / or at least one WLAN router and / or at least one mobile device and / or at least one mobile network and / or at least one internet connection and / or an IoT portal. It can therefore be said that the communication module 7, located between the first communication system 26 and the second communication system 27, can use the LoRaWAN and / or WLAN standard for communication and can forward the signals to a LoRaWAN gateway or a WLAN router, respectively.

[0193] Figure 11 shows that signals from the AI ​​chip 17 are forwarded via the communication module 7. Figure 11 further shows that the signals from the LoRaWAN gateway are forwarded to the IoT portal via mobile network and then via the internet.

[0194] Another possibility is that signals received by the AI ​​chip 17 are forwarded via the communication module 7 to at least one WLAN router using the WLAN standard. Fig. 11 shows that the signals from the at least one WLAN router are forwarded directly to the IoT portal via the internet.

[0195] It is also possible that a mobile device receives the signals from the AI ​​chip via Wi-Fi and / or that the signals are transmitted from the mobile device to the AI ​​chip 17 via standard Wi-Fi.

[0196] The communication paths described above naturally work in both directions, i.e. from the first communication system 26 to the second communication system 27 and / or from the second communication system 27 to the first communication system 26.

[0197] It is particularly preferred that the second communication system 27 can forward commands, such as setting changes requested by the user, to the second communication system 26 and that the AI ​​chip 17 can process and / or execute and / or have these commands executed.

[0198] The monitoring module according to the present disclosure can also be used in other variants, for example in rat boxes, live traps, glue traps and the like. Reference numeral:

[0199] 1 monitoring module

[0200] 2 cameras

[0201] 2a Camera module

[0202] 2b plate

[0203] 3 Evaluation electronics

[0204] 4 Pest trap

[0205] 4a Pest trap housing upper part 4b Pest trap housing lower part 4c First pest trap housing opening 4d Second pest trap housing opening 5 Surroundings

[0206] 6 Sensor

[0207] 6a Accelerometer

[0208] 6b Passive infrared sensor

[0209] 6c temperature sensor

[0210] 6d humidity sensor

[0211] 7 Communication module

[0212] 8 LED modules

[0213] 9 monitoring module housings

[0214] 10 Monitoring module housing top

[0215] 11 Monitoring module housing base

[0216] 12 circuit boards

[0217] 13 First connecting element

[0218] 14 screw connection

[0219] 15 Second connecting element

[0220] 16 cable connection

[0221] 17 AI chips

[0222] 18 Antenna

[0223] 19 On / Off switch

[0224] 20 width

[0225] 21 Length 22 First Height

[0226] 23 Second Height

[0227] 24 Second case

[0228] 25 Housing covers

[0229] 26 First communication system

[0230] 27 Second communication system (higher-level communication system)

Claims

3370841 / 32 38 Patent claims 1. Monitoring module (1) for pest control with at least one camera (2) and evaluation electronics (3) which is configured to determine the status of a pest trap (4) and / or an environment (5) of the monitoring module (1), characterized in that the evaluation electronics (3) is configured to infer the status of the pest trap (4) and / or the status of the environment (5) from image data generated by the camera (2) by processing an algorithm.

2. Monitoring module ( 1 ) according to claim 1, characterized in that the algorithm distinguishes at least one, preferably all, of the following states of the pest trap (4 ): trap ready for operation; trap triggered but empty; trap triggered, pest dead; trap triggered, pest caught alive.

3. Monitoring module ( 1 ) according to any of the preceding claims, wherein the algorithm is an artificial intelligence.

4. Monitoring module ( 1 ) according to one of the preceding claims, wherein at least one sensor ( 6 ) is provided which is signal-connected to the evaluation electronics (3 ), wherein the evaluation electronics (3 ) is configured to activate the camera (2 ) if sensor data from the at least one sensor ( 6 ) meet at least one trigger criterion.

5. Monitoring module (1) according to at least one of the preceding claims, wherein the evaluation electronics (3) includes at least two processors, wherein a first processor is configured to put a second processor into a wake-up state if the sensor data of the at least one sensor (6) fulfills the at least one trigger criterion, and 3370841 / 32 39 wherein the second processor in the waking state is configured to control the camera (2 ) to generate image data and to evaluate the image data by processing the algorithm, preferably to infer the status of the pest trap (4 ).

6. Monitoring module ( 1 ) according to one of claims 4 or 5, wherein the at least one sensor ( 6) comprises an accelerometer ( 6a ) and / or a passive infrared sensor ( 6b ) and / or a temperature sensor ( 6c ) and / or a humidity sensor ( 6d ).

7. Monitoring module ( 1 ) according to claim 6, wherein the evaluation electronics (3) is configured to send measured values ​​from the temperature sensor ( 6c) and / or the humidity sensor ( 6d) to a higher-level monitoring system (27 ) for the purpose of mold monitoring.

8. Monitoring module ( 1 ) according to one of the preceding claims, wherein a communication module (7 ) is provided which is configured to transmit the specified states of the pest trap (4 ) to a higher-level monitoring system (27 ), preferably to trigger a notification to an operator of the pest trap (4 ).

9. Monitoring module ( 1 ) according to claim 5, wherein the second processor has at least one processing core which is optimized for evaluating neural networks.

10. Monitoring module (1) according to one of the claims, wherein the monitoring module (1), in particular the evaluation electronics (3), is configured to segment the image data generated by the camera (2) into a plurality of individual image sections, preferably according to a predetermined pattern. 3370841 / 32 40 11. Monitoring module ( 1 ) according to claim 8, wherein the communication module (7 ) is configured to transmit the image sections to a server arranged separately from the monitoring module ( 1 ).

12. Monitoring module ( 1 ) according to claim 11, wherein the server arranged separately from the monitoring module ( 1 ) is configured to at least partially, preferably completely, reconstruct the image data generated by the camera (2 ) using the image sections.

13. Monitoring module ( 1 ) according to any of the preceding claims, wherein the monitoring module is battery-operated.

14. Monitoring module ( 1 ) according to claim 8, wherein the communication module (7 ) is designed to communicate using the “Long Range Wide Area Network” (LoRaWAN) standard.

15. Monitoring module ( 1 ) according to claim 5 and claim 8 or 11, wherein the second processor is configured to transmit the status of the pest trap (4 ) to the first processor, and the first processor is configured to send the status via the communication module (7 ) to the higher-level monitoring system (27 ).

16. Pest trap (4) with a monitoring module (1) according to one of the preceding claims.

17. Pest trap (4) according to claim 16, wherein the pest trap (4) comprises at least one snap trap and / or at least one sticky trap.

18. Monitoring system with one or more pest traps (4) according to claim 16 or 17, preferably wherein additionally a 3370841 / 32 41 A gateway is present which is connected to the monitoring module ( 1 ) on the one hand and to the Internet on the other.

19. Monitoring system according to claim 18 with a server arranged separately from the monitoring module ( 1 ), which is configured to reconstruct the image data generated by the camera (2 ) at least partially, preferably completely, using the image sections.

20. Monitoring system according to claim 19, wherein the server and / or at least one further server is configured to infer the status of the snap trap from the reconstructed image data.

21. Monitoring method for monitoring a pest trap (4), in particular according to one of claims 16 or 17, wherein image data generated in the area of ​​a pest trap (4) and / or an environment of the pest trap (4) are processed by an algorithm and thereby the status of the pest trap (4) and / or the environment (5) of the pest trap (4) is inferred.

22. Use of a monitoring module ( 1 ) according to at least one of claims 1 to 15 in a method according to claim 21.

23. Computer program product comprising instructions which, when the program is executed by a computer, cause it to process image data generated in the area of ​​a pest trap (4 ) and / or an environment of the pest trap (4 ) by means of an algorithm and thereby infer the status of the pest trap (4 ) and / or the environment (5) of the pest trap (4 ).

24. Transitory or non-transient computer-readable storage medium on which the computer program product according to claim 23 is stored.

Images