Sensor device for identifying xylophagous insects and method for using same

Abstract

The present invention relates to a device for identifying xylophagous insects from a vibratory signal in a structure, for effective, remote, ex-situ identification that can be adapted to different species of xylophagous insects. The device comprises a first sensor with a first amplifier with a gain selector and a first filter with a frequency selector, a processor and a communication bus, all arranged in a first casing forming a first assembly (1). The processor memory contains reference values, the Wavelet transform or decision-tree-based artificial intelligence algorithms, and boosting-based artificial intelligence algorithms. The processor can identify the vibratory signal and select the gain level and the frequency ranges. The method comprises a first processing, a reconfiguration and a second processing, a communication of the result to an IoT device and a consultation via a user device.

Description

[0001] SENSOR DEVICE FOR THE IDENTIFICATION OF XYLOPHAGUS INSECTS AND

[0002] PROCEDURE FOR USING IT

[0003] TECHNICAL FIELD OF THE INVENTION

[0004] The present invention falls within the pest control sector, in the field of detection or identification of wood-eating insects for their subsequent elimination, specifically in the identification of the vibration waves produced by said insects and in the devices and procedures used for said identification.

[0005] BACKGROUND OF THE INVENTION

[0006] It is known that wood-eating insects, such as termites, attack both wood and any other occasional source of cellulose such as paper, fabrics, etc., acting inside the wood without the possibility of detecting their presence until the damage is visible and / or has a large scope.

[0007] For this reason, to reduce the damage caused by these insects, it is essential to detect and control their activity at the beginning of their life on wood or cellulose-based goods, before an infestation capable of destroying them develops. In this field, the terms "detection" and "identification" can be considered synonymous, implying the acquisition of insect signals and subsequent processing of the captured signal.

[0008] Current methods used for the detection of wood-eating insects, such as termites, can be classified into three categories:

[0009] -baits: require insects to approach the bait in order to be detected by visual inspection, requiring the "on-site" displacement of a specialist technician;

[0010] - portable detectors: these are based on the use of special devices whose results must be interpreted by a trained and experienced specialist technician, allowing detection only when the technician is using the equipment "in situ", such as the Audiotermes device (helpest21.com / catalogo / audiotermes) and the Inadec project (Inadec: Détection automatique des insectes dans le bois, Marie-Madeleine Sermet and Olivier Winkel, CTBA INFO, n s 59, May 1996) that detect sounds, Termatrac T3i (https: / / termatrac.com / t3i-all-sensor) that uses radar technology, or Termite Seeker (https: / / www.domyown.com / pest-barrier-termite-seeker-tcts3-p-8522.html, https: / / irt-tech.net / ) that is based on the detection of the accumulation of CO2 generated by termites, all of them being relatively expensive;

[0011] -devices with remote communication: such as the one set forth in patent publication number W02013 / 000028A1 , which analyses sound in a microphone-type detector. After detecting pest activity, a signal is sent wirelessly, which can be received on a mobile phone or at a remote base station and can take the form of a short message; this has the disadvantage of using a microphone, which is highly influenced by ambient noise and for this reason its normal use is with a bait isolated from ambient noise.

[0012] The patent with publication number FR2998970A1 is known, which discloses a system and its operating method for the identification of wood-eating insects, comprising an acoustic transducer, means for processing the collected acoustic signal in real time, detecting energy peaks representative of rupture of cellulosic fibers due to the activity of wood-eating insects, that is, means for comparing the signal with a reference, means for saving the signal with an amplitude that exceeds a certain threshold, the interval of interest being between 70 KHz and 1 MHz; having the disadvantage that the source code is within the microcontroller of the same "IoT" device, without the possibility of modifying its parameters without accessing the source code, it is focused on a standard "IoT" type device because it includes by default the communications network being a direct communication with the server / cloud without the option of being integrated into another local network of another developer;The frequency range is too high, which prevents real-time analysis and requires high energy expenditure.

[0013] Also known is the patent with publication number ES2390557T3, which sets forth a method for identifying wood-eating insects using a system with several devices that communicate with each other wirelessly to form a network, including acoustic sensors, a signal conditioning module with filtering, amplification and analog-to-digital conversion, where the digital processing of the signal for identifying the insects is performed in a central device, using an algorithm based on the Fourier transform. Such calculations are complex and consume a lot of the device's energy resources. Furthermore, the use of a central device for analyzing the audio from the other devices in the network means that the devices cannot be used independently of this same wireless network.

[0014] The Internet of Things (IoT) is currently defined as the field that allows everyday physical items to be connected to the Internet: such as household objects, medical devices, clothing, personal accessories, etc. IoT devices can be defined as devices in these everyday physical items that have the necessary components to collect data and send it over the Internet to a remote receiver, for example, a server or cloud. The interconnection between IoT devices and the Internet make up what are called IoT networks. An IoT network is made up of a network of IoT devices, where at least one of them, usually the coordinator, communicates over the Internet.

[0015] Artificial intelligence (AI) is a field of computer science that seeks to develop systems capable of learning, reasoning, adapting, and performing tasks that, until recently, could only be performed by humans. Artificial intelligence models are algorithms and computational approaches designed to simulate and replicate human cognitive and reasoning skills in machines and computer systems. Their goal is to enable machines or electronic devices to learn, adapt, and perform tasks that normally require human intelligence, such as recognizing the sound of a wood-eating insect. Therefore, when referring to artificial intelligence algorithms here, we should understand them as mathematical models or software functions, which are based on specific artificial intelligence models.

[0016] DESCRIPTION OF THE INVENTION

[0017] The invention is set forth and characterized in the independent claims, the dependent claims describe further features thereof.

[0018] The object of the invention is a device for the identification of wood-eating insects and the method of using it. The technical problem to be solved is to configure the components of the device and the steps of its method of use so that a remote identification is carried out, without the need for a bait or the "on-site" presence of the device operator. This identification is effective, adaptable to different species of wood-eating insects, without requiring as many resources as an algorithm based on the Fourier transform or other known algorithms that also require a lot of resources, without being affected by ambient noise, and integrable with "IoT" devices and / or in "IoT" networks created to measure or already existing "on-site". The captured signal is processed in real time, as it is acquired.

[0019] In view of the foregoing, the present invention relates to a sensor device for identifying wood-eating insects by means of the vibration signal generated by said wood-eating insects in a wooden or cellulose-based structure, comprising a first sensor connected to a first amplifier and a first filter, said first amplifier and first filter connected to a processor with memory, said processor connected to a communication bus, which can allow an external connection to an "IoT" device, as mentioned below in the detailed description, wherein an amplifier and a filter of those known have, by manufacture, fixed values, of gain in the case of the amplifier and of frequency in the case of the filter, preconfigured and, therefore, invariable, without the possibility of handling and changing them once installed.

[0020] The device is characterized by the fact that the first amplifier includes a gain selector; the first filter includes a frequency selector; the first sensor, the first amplifier, the first filter, the processor with memory, and the communication bus are arranged in a first housing, configuring a first assembly. The processor includes in memory reference values ​​for the vibration signal corresponding to wood-eating insects, the wavelet transform or decision tree-based artificial intelligence algorithms, and also boosting-based artificial intelligence algorithms.The processor is configured such that, using the wavelet transform or decision tree-based artificial intelligence algorithms, it can identify the vibration signal of the wood-eating insects captured by the first sensor and, using boosting-based artificial intelligence, it can select the gain level of the first amplifier and the frequency ranges of the first filter to improve identification and avoid saturation of the captured vibration signal, which occurs in state-of-the-art devices. It can also modify the configuration of the input parameters of the wavelet transform or the decision tree-based artificial intelligence algorithms to also improve the reliability of the identification.A percentage of the gain level of the first amplifier can be automatically adjusted according to the signal values ​​captured in real time by said first amplifier, normally up to a maximum value set by hardware during the manufacture of the device (for example, 10dB).

[0021] By "identification" in this invention we mean the sum of the detection of the insect signal plus the actual identification of the species to which the insect belongs, which uses the Wavelet transform or decision tree-based artificial intelligence algorithms to identify the signal, while "boosting"-based artificial intelligence algorithms are used to process the signal.

[0022] The amplification level depends on multiple factors, such as the noise level of the signal captured by the sensor, the ambient noise level (when a microphone is included), the type and larval state of the wood-eating insect to be identified, the characteristics of the vibratory signal generated by the insect, and the distance between this insect and the sensor device (when said distance is taken into account, as mentioned below in the detailed explanation, with the inclusion of a second sensor).

[0023] With the invention described here, the amplifier gain and frequency filtering are controlled by the device's software, which is somewhat complex, matching the complexity of the device's hardware. Wavelet transform or decision tree-based artificial intelligence algorithms and boosting-based artificial intelligence work together to make the identification of wood-boring insects more effective.

[0024] The exposed amplification is managed by two gain selectors, one automatic by hardware included in the amplifier itself, so as not to saturate the signal, and another managed by artificial intelligence algorithms based on "boosting", to generate a fixed amplification level (reference for the hardware selector) during the recording period or until the measurement conditions exceed a certain noise threshold. It cannot be fixed all the time since the automatic hardware amplification part will be adjusted according to the characteristics of the signal captured by the sensor device. Furthermore, the amplification level is fixed by software until the next measurement, when said software analyzes the noise level of the captured signal and the ambient noise level (through additional microphone-type sensors) and decides whether it needs to be adjusted again, being able to change the amplification level again.Signal filtering in the frequency range of interest, that of wood-eating insects, is also controlled by boosting-based artificial intelligence algorithms, using, for example, variable-capacitance diodes (also known as Varicap or Varactor diodes), adjusting the filtering to the specific frequency bands of each wood-eating insect species to be identified. In this way, a single device can identify a wide range of wood-eating insects; all that is required is to configure the device to adjust it according to the corresponding data of the wood-eating species to be identified.

[0025] The fact that the gain level of the first amplifier can be adjusted automatically by hardware means that it is done through automatic gain control (AGC), which is a feedback regulator circuit designed to maintain an adequate signal amplitude at its output, despite variations in the input signal amplitude. The average or peak output signal level is used to dynamically adjust the gain, allowing the circuit to operate satisfactorily with a wider range of input signal levels.

[0026] The Wavelet Transform has been known for some time, and its use is well-known. When used individually or in conjunction with an algorithm based on a decision tree, although it can also be based on a neural network, a first model is created for identifying wood-eating insects. A second model, for managing the gain amplification level of the frequency range, is based on boosting, which uses a set of decision trees to make the decision based on the majority, although it can also be based on a neural network. This second model can also be used to optimize the parameters of the first model.Both models use self-learning ("machine learning") algorithms that evolve and are based on an algorithm already trained using audio files collected in various environments and situations (environmental, structural, insect types); they are an alternative that allows for the creation of relatively low-complexity algorithms, using spectral and time-domain characteristics, but without calculating the Fourier transform. To achieve this, it is appropriate to rely on a database of sounds recorded with the same type of device for which the algorithm is generated for training the algorithm (for example, there may be more than 100 types of insects, from different environments and situations). The resulting algorithm is lighter, consumes fewer resources, and has the robustness and capacity to predict the presence of wood-boring insects with high accuracy.

[0027] The sensor device is thus configured as an encapsulated sensor that allows IoT device developers to integrate it with any type of IoT device and network. Since the intelligence for identifying wood-boring insects is encapsulated within the sensor device, the source code is not disclosed or shared with the sensor device when commercialized. The only communication with the sensor device is through the communication bus, using predefined commands according to the implemented communication protocol (I2C, UART, MODBUS, etc.).

[0028] Thus, the sensor device can be integrated with any IoT device through the protocols defined in the communication bus. The sensor device is just another sensor in the monitoring network, since once integrated with the IoT device, the measurement frequency and transmission mode to a server or cloud do not depend on the sensor device itself, but rather on how the developer or integrator implements the sensor device and the functionalities available for the IoT device used or the platform for receiving this data. Similarly, it can also be integrated with virtual assistants such as the well-known Sih, Alexa, or Google Assistant.

[0029] The use of this sensor device results in cost and infrastructure savings, as it is an additional sensor in a building's home automation network. This eliminates the need to acquire an entire system specialized solely for termite identification, as indicated in the aforementioned prior-art patents. It is also unnecessary to install an additional sensor network in the building dedicated exclusively to identifying wood-boring insects. The sensor device of the invention also allows for the remote collection and transmission of audio samples to facilitate the maintenance work of pest control companies. To do so, the user must request the corresponding audio file from the sensor device via the communications bus, using predefined commands and protocols recognized by the sensor device.This audio recording request can be pre-programmed during the on-site equipment installation phase or can be performed remotely at the user's request.

[0030] Thus, a device is configured that forms a single assembly by being integrated into a housing, like a module, with simple configuration, which also allows it to be reproduced in as many assemblies as required, that is, there can be two, three, or more assemblies operating simultaneously according to the needs of the specific application.

[0031] The operator does not need to be on-site to obtain the data from the identified signal. Instead, they can be located far from the measurement point and manage the data remotely, for example, through a web application. This is because the sensor device is a set with communication means that can be connected to an IoT device, either by cable or wirelessly, and can communicate with a server or the cloud.

[0032] The use of artificial intelligence helps adjust the input parameters used to identify the signal from wood-boring insects, while the device self-learns and discretizes problems or more or less exact values ​​over time. To achieve this, it uses a non-volatile or external memory where the configuration parameters, including reference values, are stored.

[0033] By controlling the gain amplification and filter values ​​of the sensor device, it is possible to achieve an amplification and frequency range capable of identifying various types of wood-eating insects and various larval stages at a distance of several meters.

[0034] The use of algorithms based on the Wavelet Transform or decision tree-based artificial intelligence and boosting-based artificial intelligence, instead of the Fourier Transform or other well-known algorithms that also require a lot of resources, improves the energy consumption of the sensor device and improves its autonomy, thus increasing the autonomy of the assembly formed with the associated IoT device.

[0035] Reference values, stored in the processor's non-volatile or external memory, can include details, allowing you to identify wood-boring insects by species type.

[0036] The invention is also the method of using a sensor device for the identification of wood-eating insects, with a device as described, comprising the following steps in sequence:

[0037] -capture by the first sensor of the vibration signal generated by the wood-eating insects; -first amplification of said signal by the first amplifier and first filtering by the first filter according to predefined values; as known in the prior art.

[0038] It is characterized in that the first amplifier includes a gain selector and the first filter includes a frequency selector, the procedure also comprises the following steps in sequence:

[0039] -first processing of said signal by the processor, which includes analog-to-digital conversion and the use of the Wavelet transform or decision tree-based artificial intelligence algorithms for the identification of wood-eating insects and artificial intelligence algorithms based on "boosting" to improve identification through a new configuration for the gain and frequency range;

[0040] -reconfiguration of the first amplifier with gain selector and the first filter with frequency selector, by the processor, for the gain and filtering parameters according to the results of the first processing;

[0041] -second processing of said signal to identify wood-eating insects through the use of the Wavelet transform or decision tree-based artificial intelligence algorithms;

[0042] -communication of the result of the second processing via the communication bus to an IoT device, which includes a communication module with a wired or wireless connection for sending the result directly to a cloud storage or server or via another IoT device; -consultation of the stored information via cable or wireless means via a user device.

[0043] An advantage of said procedure is that thanks to the amplification control, a dynamic identification is carried out at distances greater than 1 meter, according to the adjustment of the artificial intelligence algorithm based on "boosting", based on the noise levels detected, which allows to obtain a much clearer signal, that is, with a better signal-to-noise ratio, which in turn allows a much more precise identification than with the devices and according to the procedures known in portable devices, which, in the best of cases, have a fixed amplification and come close to, but not reaching, 50,000 times amplification, although it is normal for them to be well below, while with the invention disclosed herein, an amplification of up to 500,000 times can be reached.The invention allows a user, any person with access credentials to the platform, without being a pest control specialist, to receive the data and listen to the sounds, to evaluate them, being a basic evaluation and can be effective in the most favorable cases, when the sounds are clear and the configured identification limits are high, given that they are effectively processed and arranged digitally, so they can be viewed and treated with any common current device, normally with a screen: smartphone, tablet, computer, etc.

[0044] BRIEF DESCRIPTION OF THE FIGURES

[0045] This descriptive report is complemented with a set of figures, illustrative of the preferred example, and never limiting the invention.

[0046] Figure 1 shows a schematic of the device of the invention, with a first assembly containing a first sensor, a second assembly containing a second sensor, IoT devices, a server and cloud, and user devices such as a computer and a mobile phone. The path followed by the signals is represented by arrowheads.

[0047] Figure 2 shows a partial diagram of the invention, showing a signal generator, the first assembly and first sensor, the second assembly and second sensor. Figure 3 shows a diagram of a first assembly, showing the signal path.

[0048] DETAILED EXPLANATION OF THE INVENTION

[0049] Figure 1 shows a sensor device for the identification of wood-eating insects, by means of the vibration signal generated by said wood-eating insects in a structure (10) made of wood or cellulose-based, comprising a first sensor (1.1) connected to a first amplifier (1.2) and to a first filter (1.3), said first amplifier (1.2) and first filter (1.3) connected to a processor (1.4) - for example 32 bits - with memory, non-volatile or external, said processor (1.4) connected to a communication bus (1.5), which can in turn be connected to an "IoT" device (3), said "IoT" device (3) includes a communication module (3.1) with cable or wireless connection to another "IoT" device (3) or to a cloud storage or server (4). The first amplifier (1.2) includes a gain selector, the first filter (1.3) includes a frequency selector; the first sensor (1.1 ), the first amplifier (1 .2), the first filter (1 .3), the processor (1.4) and the communication bus (1.5) are arranged in a first housing (1.6) -normally with acoustic insulation and with weather protection characteristics, such as IP65 or higher-, configuring a first assembly (1 ). The processor (1.4) includes in the memory reference values ​​of the vibration signal corresponding to wood-eating insects, the Wavelet transform or artificial intelligence algorithms based on decision trees and artificial intelligence algorithms based on "boosting"; the processor (1 .4) is configured in such a way that by means of the Wavelet transform or the artificial intelligence algorithms based on decision trees it can identify the vibration signal of the wood-eating insects captured by the first sensor (1.1 ) and by means of artificial intelligence based on "boosting" it can select the gain level of the first amplifier (1 .2) and the frequency ranges of the first filter (1 .3) to improve identification and avoid saturation of the captured vibration signal, as well as modify the configuration of the input parameters of the Wavelet transform or the decision tree-based artificial intelligence algorithms also to improve the reliability of the identification; and in such a way that a percentage of the gain level of the first amplifier (1 .2) can be automatically adjusted according to the values ​​of the signal captured in real time by said first amplifier (1.2). As represented in the figure and with a dashed line, other "IoT" devices (3) can be included to configure wireless networks, such as ZigBee, LoRa or others, in various network topologies.

[0050] Although not shown, it is common for the assembly (1 ) to be electrical and electronic components arranged on PCB boards inside a container, formed in the invention set forth by the first housing (1 .6). The “IoT” device (3) can be of any brand or model, as long as it meets the communication requirements pre-established by the first assembly (1 ), such as, for example, the voltage (for example: 5 V) and the communication bus protocol (1 .5) (for example: I2C, UART, MODBUS, etc.) with the commands and parameters predefined by the first assembly (1 ). Likewise, other common elements can be included in the “IoT” device (3), such as, for example, a power supply -it can be any type of external power supply-, a headphone connector, a memory card slot -for example, SD type or similar-, etc.

[0051] One option is that the first filter (1.3) comprises variable capacitance diodes, also known as Varicap or Varactor diodes, so that it allows the configuration of a frequency range for filtering the vibration signal captured by the first sensor (1.1). One or more circuits of this type can be used in order to filter the signal in several frequency ranges or to subtract certain frequencies or frequency ranges that in certain scenarios can affect the quality of the captured signal and / or the identification of the wood-eating insects.

[0052] Another option is that the first housing (1.6) includes acoustic and electromagnetic field shielding. Said shielding is a conductive layer that prevents the coupling of noise and interference and prevents or reduces the penetration or radiation of electric and / or magnetic fields. To eliminate the noise generated during the operation of the artificial intelligence of the processor (1.4), a special shielding is used, for example, the first housing (1 .6) can be made of metallic material or one or two ground planes separated from the microcontroller are included. The solution of using ground planes is very effective and allows to obtain a sensor device of small dimensions. The first sensor (1.1 ) can be of any type, such as, for example, piezoelectric, capacitive, inductive, resistive, etc. An advantageous detail is that the first sensor (1 .1) is of the passive type, it only receives the vibratory signal, it does not emit any signal, being specially selected for the range of frequencies that are normally generated by the wood-eating insects that are to be identified.

[0053] One option, Figure 1, is that it also comprises a first interface (6) that is between the first sensor (1.1) and the structure (10) so that it provides a stable and uniform settlement of the first sensor (1.1) in the structure (10). For example, it can be a thin layer of an adhesive or a putty. The first interface (6) improves the transmission of vibration waves and filters a part of the possible noise transmitted from the environment to the structure (10).

[0054] Another option, figure 1, is that it also comprises a second sensor (2.1) connected to a second amplifier (2.2) with gain selector and to a second filter (2.3) with frequency selector, said second sensor (2.1), second amplifier (2.2) and second filter (2.3) are arranged in a second housing (2.4) configuring a second set (2), said second set (2) is connected to the processor (1.4), for example, by means of a 4-pin connector or by means of a 4-position 2.5 mm Jack type input, so that, with the data captured by the sensors of the sets (1,2), the Wavelet transform or the artificial intelligence algorithms based on the decision tree and the artificial intelligence based on "boosting" allow to detect with great precision the area of ​​the attack, from where the vibratory signal of the xylophagous insects captured by the first sensor (1.1) and the second sensor (2.1 ), which implies the direction and distance from where the vibration signal comes from. This gives the advantage of vectorizing the detected signal, identifying where the vibration signal comes from, through the displacement of the signal. Furthermore, in a manner analogous to the first sensor (1.1 ), a second interface (7) can be included that is between the second sensor (2.1 ) and the structure (10) so that it provides a stable and uniform settlement of the second sensor (2.1 ) in the structure (10), with the same advantages as those mentioned for the first interface (6).

[0055] Another option, figure 1 , is that it also comprises a microphone (8) connected to the processor (1.4), the microphone (8) being configured to capture external noise, so that the artificial intelligence based on "boosting" can adjust the gain level of the first amplifier (1.2) to counteract the influence of external noise on the signal captured by the first sensor (1.1 ). With this gain adjustment, a better signal-to-noise ratio and a better identification result are achieved, avoiding false positive results.

[0056] Another option, figure 1 , is that it also comprises a temperature and humidity sensor (9) connected to the processor (1 .4). This temperature and humidity sensor (9) is inserted into the structure (10), for example, wood or derivatives, or is placed in contact with it, so that the artificial intelligence based on "boosting" can adjust the parameters of the Wavelet transform or the artificial intelligence algorithms based on decision trees to adjust the level of interpretation of the vibration signal captured by the first sensor (1.1 ), that is, the input data of temperature and humidity allow the algorithm to make a fine compensation adjustment in the interpretation of the vibration signal.

[0057] One detail is that the gain selector of the first amplifier (1.2) that acts on it can amplify the signal more than 100,000 times. The amplification goes from 50,000 to 500,000 times, which corresponds to between 94 dB and 1.14 dB. In this way, there is a gain selector that can be configured both by hardware and software, so that the gain can be adapted to the possible noise coming from the environment and the distance with respect to the possible entry point of the insects (for example: 1-4 meters).

[0058] Another detail is that the range of the frequency range selector, that is, the frequency interval, of the first filter (1.3) is from 50 Hz to 150 kHz. It has been shown that, for example, termites emit sounds in the high frequency ranges 100 kHz - 150 kHz and low frequencies 50 Hz - 22 kHz, so that with the frequency selector it is possible to work in the most beneficial frequency range according to the case.

[0059] The method of using a sensor device for the identification of wood-eating insects, with a device as described, comprising the following steps in sequence:

[0060] -capture by the first sensor (1.1) of the vibration signal generated by the wood-eating insects; -first amplification of said signal by the first amplifier (1.2) with gain selector and first filtering by the first filter (1.3) with frequency selector according to predefined values;

[0061] -first processing, for example up to 5 seconds, of said signal by the processor

[0062] (1 .4), which includes analog-digital conversion and the use of the Wavelet transform or decision tree-based artificial intelligence algorithms for the identification of wood-eating insects and artificial intelligence algorithms based on “boosting” to improve the identification of wood-eating insects through a new configuration for the gain and frequency range;

[0063] -reconfiguration of the first amplifier (1 .2) with gain selector and the first filter (1 .3) with frequency selector, by the processor (1 .4), for the gain and filtering parameters according to the results of the first processing;

[0064] -second processing of said signal to identify wood-eating insects through the use of the Wavelet transform or decision tree-based artificial intelligence algorithms;

[0065] -communication of the result of the second processing through the communication bus

[0066] (1.5) to an “IoT” device (3) which includes a communication module (3.1) with a cable or wireless connection for sending the result directly to a cloud storage or server (4) or through another “IoT” device;

[0067] -consultation by cable or wirelessly of the information stored through a user device (5).

[0068] Specifically, the amplification is more than 100,000 times, and can reach 500,000 times, as mentioned above. This has the advantage of having a much higher amplification than the state of the art, which allows for better signal processing, identifying early larval stages of woodworms at greater distances more accurately and effectively.

[0069] Also, the selection of the frequency ranges for the first filter (1.3) is from 50 Hz to 150 kHz, as mentioned above. This provides the advantage of adjusting and using the sensor device in the most appropriate frequency range for each case.

[0070] One option is that the device also comprises a second sensor (2.1 ) connected to a second amplifier (2.2) with gain selector and to a second filter (2.3) with frequency selector, said second sensor (2.1 ), second amplifier (2.2) and second filter (2.3) are arranged in a second housing (2.4) configuring a second assembly (2), said second assembly (2) is connected to the processor (1 .4); prior to the identification stage, an "in situ" calibration and validation is carried out by means of a signal generator (1 1 ), figure 2, with signals similar to the vibratory signal produced by xylophagous insects in a structure (10), normally between 50 Hz and 10 kHz, the signal generator (1 1 ) being arranged at a predefined distance (D) with respect to the first sensor (1.1 ) and the second sensor (2.1 ), thus being a known parameter, said signal generator (1 1 ) transmitting said signals to the structure (10) by direct contact and carrying out the steps of the procedure in such a way that obtaining a positive result from the second processing, the capacity of the device to identify the exclusive vibratory signal of xylophagous insects is validated.

[0071] Another option is that when the device also comprises a second sensor (2.1) connected to a second amplifier (2.2) with gain selector and a second filter

[0072] (2.3) with frequency selector, said second sensor (2.1), second amplifier (2.2) and second filter (2.3) are arranged in a second housing (2.4) configuring a second assembly (2), said second assembly (2) is connected to the processor (1 .4), where in the capture stage said second sensor (2.1) captures the vibratory signal and in the second processing the identification allows the Wavelet algorithm or the artificial intelligence algorithms based on the decision tree and the artificial intelligence based on "boosting" integrated in the processor (1 .4) to detect the area from where the vibratory signal of the wood-eating insects comes from based on the difference in reception of the signals captured by the first sensor (1 .1) and the second sensor (2.1). Like the processor

[0073] (1 .4) communicates directly with the first amplifier (1 .2) and first filter (1 .3) to adjust the parameters and improve detection, the second amplifier (2.2) and the second filter (2.3) also have direct communication with the processor (1 .4) with the same objective. It is represented by the double arrow between the first set (1 ) and the second set (2) in figures 1 and 2 because the processor (1 .4) sends the signals to control the amplification and filtering parameters of the second set (2) and receives the vibration signal captured by the second sensor (2.1 ).

[0074] In the procedure other usual steps may take place, such as the generation of a warning or alarm from the server (4) when a preconfigured threshold is exceeded, said warning being sent to the user, especially to his mobile device, such as a smartphone (5), or the connection of the user to a web platform created for this purpose where data can be viewed and consulted in real time such as historical data, sounds, alarms, etc.

Claims

CLAIMS 1 - Sensor device for identifying wood-eating insects by means of the vibration signal generated by said wood-eating insects in a structure (10) made of wood or cellulose, comprising a first sensor (1 .1 ) connected to a first amplifier (1 .2) and to a first filter (1 .3), said first amplifier (1 .2) and first filter (1 .3) connected to a processor (1 .4) with memory, said processor (1 .4) connected to a communication bus (1.5), characterized in that the first amplifier (1.2) includes a gain selector, the first filter (1 .3) includes a frequency selector; the first sensor (1.1 ), the first amplifier (1 .2), the first filter (1 .3), the processor (1 .4) and the communication bus (1.5) are arranged in a first housing (1.6) configuring a first assembly (1 ); the processor (1.4) includes in the memory reference values ​​of the vibration signal corresponding to wood-eating insects, the Wavelet transform or decision tree-based artificial intelligence algorithms, and “boosting”-based artificial intelligence algorithms; the processor (1 .4) is configured in such a way that by means of the Wavelet transform or decision tree-based artificial intelligence algorithms it can identify the vibration signal of the wood-eating insects captured by the first sensor (1 .1) and by means of “boosting-based artificial intelligence” it can select the gain level of the first amplifier (1 .2) and the frequency ranges of the first filter (1.3) to improve identification and avoid saturation of the captured vibration signal, as well as modify the configuration of the input parameters of the Wavelet transform or the decision tree-based artificial intelligence algorithms also to improve the reliability of the identification; and in such a way that a percentage of the gain level of the first amplifier (1.2) can be automatically adjusted according to the values ​​of the signal captured in real time by said first amplifier (1.2).

2. - Sensor device for the identification of wood-eating insects according to claim 1, wherein the first filter (1.3) comprises diodes with variable capacity, such that it allows the configuration of a range of frequencies for filtering the vibration signal captured by the first sensor (1.1).

3. -Sensor device for the identification of xylophagous insects according to claim 1 in which the first housing (1 .6) includes acoustic and field shielding electromagnetic. 4.- Sensor device for the identification of xylophagous insects according to claim 1, which also comprises a first interface (6) that is between the first sensor (1.1) and the structure (10) so that it provides a stable and uniform settlement of the first sensor (1.1) in the structure (10).

5. - Sensor device for the identification of wood-eating insects according to claim 1, further comprising a second sensor (2.1) connected to a second amplifier (2.2) with gain selector and to a second filter (2.3) with frequency selector, said second sensor (2.1), second amplifier (2.2) and second filter (2.3) being arranged in a second housing (2.4) configuring a second assembly (2), said second assembly (2) being connected to the processor (1.4), such that the Wavelet transform or the artificial intelligence algorithms based on decision trees and the artificial intelligence based on "boosting" allow the area from which the vibratory signal of the wood-eating insects captured by the first sensor comes from to be detected. (1 .1 ) and the second sensor (2.1 ).

6. - Sensor device for the identification of wood-eating insects according to claim 5, which also comprises a second interface (7) that is between the second sensor (2.1 ) and the structure (10) in such a way that it provides a stable and uniform settlement of the second sensor (2.1 ) in the structure (10). 7.- Sensor device for the identification of wood-eating insects according to claim 1, which also comprises a microphone (8) connected to the processor (1.4), the microphone (8) being configured to capture external noise, such that the artificial intelligence based on "boosting" can adjust the gain level of the first amplifier. (1 .2) to counteract the influence of external noise on the signal captured by the first sensor (1.1 ).

8. - Sensor device for the identification of wood-eating insects according to claim 1, which also comprises a temperature and humidity sensor (9) connected to the processor (1.4) and inserted into the structure (10) or in contact with it, so that the artificial intelligence based on "boosting" can adjust the parameters. of the Wavelet transform or the artificial intelligence algorithms based on decision trees to adjust the level of interpretation of the vibration signal captured by the first sensor (1 .1 ).

9. -Sensor device for the identification of wood-eating insects according to claim 1, wherein the gain selector of the first amplifier (1.2) can amplify the signal more than 100,000 times.

10. -Sensor device for the identification of xylophagous insects according to claim I in which the range of the frequency range selector of the first filter (1 .3) is from 50 Hz to 150 kHz. II.-Process for using a sensor device for the identification of wood-eating insects, with a device according to claim 1, comprising the following steps in sequence: -capture by the first sensor (1.1) of the vibration signal generated by the wood-eating insects; -first amplification of said signal by the first amplifier (1 .2) and first filtering by the first filter (1 .3) according to predefined values; characterized in that the first amplifier (1 .2) includes a gain selector and the first filter (1 .3) includes a frequency selector; the method further comprises the following steps in sequence: -first processing of said signal by the processor (1.4), which includes analog-digital conversion and the use of the Wavelet transform or decision tree-based artificial intelligence algorithms for the identification of wood-eating insects and of artificial intelligence algorithms based on "boosting" to improve identification through a new configuration for the gain and frequency range; -reconfiguration of the first amplifier (1 .2) with gain selector and the first filter (1 .3) with frequency selector, by the processor (1 .4), for the gain and filtering parameters according to the results of the first processing; -second processing of said signal to identify wood-eating insects through the use of the Wavelet transform or decision tree-based artificial intelligence algorithms; -communication of the result of the second processing through the communication bus (1.5) to an “IoT” device (3), which includes a communication module (3.1) with a cable or wireless connection for sending the result directly to a cloud storage or server (4) or through another “IoT” device; -consultation by cable or wirelessly of the stored information, through a user device (5).

12. -Method of using a sensor device for the identification of wood-eating insects according to claim 11, wherein the amplification is more than 100,000 times.

13. -Process for using a sensor device for the identification of wood-eating insects according to claim 11, wherein the selection of the frequency ranges of the first filter (1.3) is from 50 Hz to 150 kHz.

14. -Method of using a sensor device for the identification of wood-eating insects according to claim 1 1, wherein the device further comprises a second sensor (2.1) connected to a second amplifier (2.2) with gain selector and to a second filter (2.3) with frequency selector, said second sensor (2.1), second amplifier (2.2) and second filter (2.3) being arranged in a second housing (2.4) configuring a second assembly (2), said second assembly (2) being connected to the processor (1 .4); prior to the identification stage, an "in situ" calibration and validation is carried out by means of a signal generator (1 1) with signals similar to the vibration signal produced by the wood-eating insects in a structure (10), the signal generator (1 1) being arranged at a predefined distance (D) with respect to the first sensor (1.1) and the second sensor (2.1 ), said signal generator (1 1 ) transmitting said signals to the structure (10) by direct contact, such that by obtaining a positive result from the second processing, the capacity of the device installed “in situ” to identify the exclusive vibratory signal of xylophagous insects is validated.

15. -Method of using a sensor device for the identification of wood-eating insects according to claim 11, in which the device further comprises a second sensor (2.1) connected to a second amplifier (2.2) with gain selector and to a second filter (2.3) with frequency selector, said second sensor (2.1 ), second amplifier (2.2) and second filter (2.3) are arranged in a second housing (2.4) configuring a second assembly (2), said second assembly (2) is connected to the processor (1 .4), where in the capture stage said second sensor (2.1 ) captures the vibration signal and in the second processing the identification allows the Wavelet algorithm or the artificial intelligence algorithms based on decision trees and artificial intelligence based on "boosting" integrated in the processor (1 .4) to detect the area from where the vibration signal of the wood-eating insects comes from based on the difference in reception of the signals captured by the first sensor (1 .1 ) and the second sensor (2.1 ).

Images