Mosquito attraction trend prediction method, device and system based on dynamic threshold

Abstract

The application discloses a kind of based on dynamic threshold's mosquito attraction tendency prediction method, device and system, it is related to data processing technical field, method includes: obtaining the relevant data of target area, according to mosquito calculation period generation environment parameter effective state and mosquito environment state set;Stable window qualification examination is carried out to historical data, and dynamic mosquito boundary and natural window state are generated;Real window state is obtained by identifying and peeling off artificial response caused by mosquito execution;Tendency prediction state is generated according to real window state and determines mosquito execution state, and then subsequent prediction logic is corrected according to execution response stage matching result, effective trapping sample and wind direction propagation response.The application solves the problem that mosquito device actively releases light source, odor and carbon dioxide pollute its own environment input after, lead to natural trapping window misjudgment, improve mosquito attraction tendency prediction accuracy and multi-node collaborative control reliability.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to a method, device, and system for predicting mosquito attraction trends based on dynamic thresholds. Background Technology

[0002] Mosquito traps are typically deployed in residential greenbelts, garden landscapes, the periphery of rural residential areas, the periphery of farms, wetland edges, outdoor monitoring points, and other semi-open target areas. They are used to release attractants such as light, odor, and carbon dioxide during periods of high mosquito activity and to trap mosquitoes through trapping chambers, negative pressure structures, or trapping channels.

[0003] Existing mosquito control methods often employ fixed-time activation, fixed-intensity output, or on / off judgment based on a single environmental parameter. For example, they may determine whether the mosquito-attracting device activates based solely on whether ambient temperature, humidity, light intensity, or carbon dioxide concentration reaches a fixed threshold. While these methods can achieve basic control in scenarios with relatively slow environmental changes and where the mosquito-attracting device does not actively alter its local environmental input, in outdoor or semi-open target areas, ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and light intensity typically have different sampling frequencies and rates of change. Directly splicing or comparing data from different frequencies within the same judgment period may result in data not corresponding to the same environmental state, easily leading to biased judgments about mosquito attraction trends.

[0004] Mosquito traps actively release light, odors, or carbon dioxide during operation. Activation of the mosquito-attracting light source may alter the ambient light intensity, odor release may change the activity state of targets near the capture channel, and carbon dioxide release may change the carbon dioxide concentration. If the control method does not distinguish between changes in the natural environment and human responses caused by the mosquito-attracting process, the system may mistake changes in light intensity, carbon dioxide concentration, or disturbances in the capture channel caused by the trap's own release behavior as the formation of a natural window, leading to misjudgment, ineffective preheating, or ineffective release.

[0005] In areas where multiple mosquito-attracting nodes are deployed in a coordinated manner, the mosquito-attracting status of upwind nodes may have a propagation impact on downwind nodes. However, if downwind nodes enter the preparatory state simply because they receive a propagation prompt, without considering the local actual window status, effective wind direction status, and effective stable window, they are also prone to false activation when the local environment is not close to the natural window. Summary of the Invention

[0006] To address the issue of how actively activating light sources, releasing odors, or releasing carbon dioxide by mosquito-attracting devices can impact local environmental data collection, this invention provides a method, device, and system for predicting mosquito attraction trends based on dynamic thresholds. This system distinguishes between changes in the natural environment and artificial responses generated during mosquito attraction, and reduces misjudgments of natural window states, ineffective preheating, ineffective release, and false triggering of multi-node collaboration.

[0007] A mosquito-attracting trend prediction method based on dynamic thresholds includes: acquiring ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within a target area, and acquiring mosquito-attracting execution records, readback states, and capture feedback states; time-aligning the heterogeneous sampling data of each environmental parameter according to the mosquito-attracting calculation cycle, removing missing samples and isolated abnormal samples, and converting them into a mosquito-attracting environmental state set according to the same historical stable range; dividing the historical mosquito-attracting environmental state set, historical mosquito-attracting execution records, historical capture feedback states, and historical readback states into candidate historical windows, excluding candidate historical windows with continuous missing data, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback, or unprocessed carbon dioxide release responses, to obtain effective stable windows; generating dynamic state boundaries and natural window state records based on the effective stable windows and the current mosquito-attracting environmental state set; marking and masking artificial light source responses, artificial release responses, and channel interference in the natural window state records based on the mosquito-attracting execution records, readback states, wind direction, and capture feedback states, to obtain the true window state; and determining the trend prediction state and mosquito-attracting execution state based on the true window state and the dynamic state boundary.

[0008] A mosquito-attracting trend prediction device based on dynamic thresholds includes a data alignment module, a stable boundary module, a response stripping module, an execution determination module, and a feedback correction module. The data alignment module acquires ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within a target area, and obtains mosquito-attracting execution records, readback states, and capture feedback states. It also performs time alignment on the heterogeneous frequency sampling data of each environmental parameter according to the mosquito-attracting calculation cycle, removes missing samples and isolated abnormal samples, and converts them into a mosquito-attracting environmental state set according to the same historical stable range. The stable boundary module divides the historical mosquito-attracting environmental state set, historical mosquito-attracting execution records, historical capture feedback states, and historical readback states into candidate historical windows, excluding candidate historical windows with continuous missing data, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback, or unprocessed carbon dioxide release responses, to obtain an effective stable window. It also generates dynamic state boundaries and natural window state records based on the effective stable window and the current mosquito-attracting environmental state set. The response stripping module is used to mark and mask artificial light source responses, artificial release responses, and channel interference in the natural window state record based on the mosquito-attracting execution record, readback status, wind direction, and capture feedback status, thereby obtaining the true window state. The execution determination module is used to determine the trend prediction state and the mosquito-attracting execution state based on the true window state and the dynamic state boundary. The feedback correction module is used to generate a theoretical response stage sequence and an actual response stage sequence after the mosquito-attracting execution state changes, and to update the dynamic state boundary based on the order matching result of the two and the effective trapping samples.

[0009] A mosquito-attracting trend prediction system based on dynamic thresholds includes multiple mosquito-attracting nodes, a data storage unit, and a display terminal. Each mosquito-attracting node includes an environmental data acquisition unit, a mosquito-attracting execution unit, a capture feedback unit, a node processing unit, and a node communication unit. The environmental data acquisition unit is used to acquire ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within the target area. The mosquito-attracting execution unit is used to activate the mosquito-attracting light source, release odors, and release carbon dioxide, and generate mosquito-attracting execution records and readback status. The capture feedback unit is used to generate capture feedback status, which includes the capture channel activity status and the actual number of mosquitoes captured. The node processing unit is used to time-align the heterogeneous sampling data of each environmental parameter according to the mosquito-attracting calculation cycle, remove missing samples and isolated abnormal samples, and convert them into a mosquito-attracting environmental state set according to the same historical stable range; it is also used to exclude candidate historical windows that do not meet the stable window qualification review, obtain effective stable windows, and generate dynamic state boundaries and natural window state records based on effective stable windows and the current mosquito-attracting environmental state set; it is also used to mark and block artificial light source responses, artificial release responses, and channel interference in the natural window state records based on mosquito-attracting execution records, readback status, wind direction, and capture feedback status, to obtain the true window state; it is also used to determine the trend prediction state and mosquito-attracting execution state based on the true window state and dynamic state boundary; it is also used to update the dynamic state boundary based on the response stage sequence matching results and effective trapping samples. The node communication unit is used to transmit the true window state, effective wind direction state, positional relationship, and propagation cues between adjacent mosquito-attracting nodes; the mosquito-attracting node receiving the propagation cues only enters the ready state when the local true window state, effective wind direction state, and effective stable window support it. The data storage unit is used to store mosquito-attracting environment state sets, effective stable windows, dynamic state boundaries, natural window state records, actual window states, artificial response shielding results, response phase sequence matching results, effective trapped samples, and propagation prompt records. The display terminal is used to display the mosquito-attracting execution status, trend prediction status, actual number of captured samples, equipment energy consumption, and abnormal prompts; it is not used to directly generate actual window states or directly switch mosquito-attracting execution states.

[0010] The beneficial effects of this invention are as follows: In the entire dynamic threshold-based mosquito attraction trend prediction method, firstly, the ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and ambient light intensity are time-aligned according to the mosquito attraction calculation cycle to reduce judgment bias caused by direct splicing of data from different frequencies. Furthermore, historical windows with missing data, abnormal sampling, inconsistent readbacks, abnormal capture feedback, or unprocessed carbon dioxide release responses are excluded through stable window qualification review, ensuring that the dynamic state boundary is established on available historical data. Further, the marking and shielding of artificial light source responses, artificial release responses, and channel interference prevent the real window state from directly using environmental changes caused by mosquito attraction itself. Further, the dynamic state boundary is updated through response phase sequence matching and effective captured samples, allowing subsequent mosquito attraction trend prediction to reference the actual capture quantity, equipment energy consumption, and readback status. Furthermore, in multi-node collaborative scenarios, mosquito attraction nodes receiving propagation cues also need to combine the local real window state, effective wind direction status, and effective stable window to determine whether to enter the preparatory state, reducing false activation caused by relying solely on propagation cues. Attached Figure Description

[0011] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0012] Figure 1 This is a schematic diagram illustrating the steps of the mosquito attraction trend prediction method based on dynamic threshold of the present invention; Figure 2 This is a schematic diagram of a portion of steps S1 in the mosquito attraction trend prediction method based on dynamic threshold of the present invention; Figure 3 This is a schematic diagram of part of step S2 in the mosquito attraction trend prediction method based on dynamic threshold of the present invention; Figure 4 This is a schematic diagram of part of step S3 in the mosquito attraction trend prediction method based on dynamic threshold of the present invention; Figure 5 This is a schematic diagram of a portion of step S4 in the mosquito attraction trend prediction method based on dynamic threshold of the present invention; Figure 6 This is a schematic diagram of the module structure and data flow of the mosquito attraction trend prediction device based on dynamic threshold of the present invention; Figure 7 This is a schematic diagram of the module structure and data flow of the mosquito attraction trend prediction system based on dynamic threshold of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0014] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0015] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0016] This invention provides a method for predicting mosquito attraction trends based on dynamic thresholds, such as... Figure 1 As shown, in one specific embodiment, the method includes: S1. Obtain the ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity and sampling time within the target area, and obtain the mosquito attraction execution record, readback status and capture feedback status; perform time alignment on the heterogeneous sampling data of each environmental parameter according to the mosquito attraction calculation cycle, remove missing samples and isolated abnormal samples, and convert them into a mosquito attraction environmental state set according to the historical stable range of the same type; S2. Divide the historical mosquito-attracting environment state set, historical mosquito-attracting execution record, historical capture feedback state and historical readback state into candidate historical windows, exclude candidate historical windows with continuous missing, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback or unprocessed carbon dioxide release response, obtain effective stable windows, and generate dynamic state boundary and natural window state records based on the effective stable windows and the current mosquito-attracting environment state set. S3. Based on the mosquito-attracting execution record, readback status, wind direction, and capture feedback status, mark and block artificial light source response, artificial release response, and channel interference in the natural window status record to obtain the true window status; determine the trend prediction status and mosquito-attracting execution status based on the true window status and the boundary between the dynamic status. S4. After the mosquito-attracting execution state changes, generate the theoretical response stage sequence and the actual response stage sequence, perform response stage sequence matching, and update the dynamic state boundary based on the response stage sequence matching results and the effective trapping samples.

[0017] In this specific embodiment, the target area can be a residential greenbelt, a garden landscape area, the periphery of a rural residential area, the periphery of a farm, the edge of a wetland, an outdoor monitoring point, or other semi-open mosquito-attracting scenarios. The method is used to determine the mosquito-attracting trend and the mosquito-attracting execution status, but not to determine the risk of mosquito bites to humans, nor to control air conditioners, smart mosquito nets, windows, fans, or sunshades to alter the human comfort environment. The dynamic state boundary is a specific state transition representation of the dynamic threshold in this application, used to determine whether there is a switch between the natural window state, the real window state, and the trend prediction state, and is not used as a fixed numerical threshold for a single environmental parameter.

[0018] In this embodiment, it should be noted that in S1, the system first collects environmental temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within the target area. Simultaneously, it reads mosquito-attracting execution records such as the activation of the mosquito-attracting light source, odor release, carbon dioxide release, execution intensity, execution time, and readback time. Because the sampling frequency and rate of change of different environmental parameters vary, the system does not directly stitch together data obtained at different times. Instead, it uses the mosquito-attracting calculation cycle as the basic processing unit, converting each environmental parameter into a periodic state within the same mosquito-attracting calculation cycle. For example, the ambient light intensity may change every [period]. A single sampling may result in a carbon dioxide concentration of [number] times per [number]. Each sampling may change the wind direction. Sampling is performed once, and the mosquito-attracting calculation period is taken as follows: The system will then be the same The valid samples of each parameter are processed into valid states of the corresponding environmental parameters. The system employs gradual variation processing for ambient temperature and humidity, background variation processing for atmospheric pressure, anti-abrupt processing for wind speed and carbon dioxide concentration, extreme sampling rejection processing for ambient light intensity, and processing for wind direction based on the dominant diffusion direction. The resulting mosquito-attracting environmental state set is no longer a direct combination of original values ​​of different dimensions, but rather a collection of relative positions, directions of change, and stability within the same type of parameter. This serves as the data basis for subsequent natural window state recording and real window state judgment.

[0019] In S2, the system does not directly use all historical data as the basis for dynamic state boundaries. Instead, it first divides the historical mosquito-attracting environmental state set, historical mosquito-attracting execution records, historical capture feedback states, historical carbon dioxide release records, and historical readback states into candidate historical windows in chronological order. Then, it performs a stable window eligibility review on these candidate historical windows. The purpose of the stable window eligibility review is to exclude data segments that do not reflect the natural trapping background of the target area. For example, if the environmental humidity in a candidate historical window is continuous... The fact that no effective state was formed in one mosquito-attracting calculation cycle indicates that there are continuous gaps in the window; the ambient light intensity in a certain candidate historical window was discarded due to reflection from vehicle lights. Secondly, the carbon dioxide concentration was eliminated due to people passing by. If an abnormal sampling cluster is found in a candidate historical window, it indicates that the window contains such a cluster. If a carbon dioxide supply module in a candidate historical window has performed a release but failed to complete the corresponding manual release response, then that window is also unsuitable as a reference for the dynamic state boundary. Candidate historical windows that pass the review are determined as valid stable windows. The system generates a background window trend based on the continuous change path of the mosquito-attracting environmental state set within the valid stable window, and generates a neighboring window trend based on the mosquito-attracting environmental state set within the current mosquito-attracting calculation cycle and its neighboring calculation cycles. Subsequently, the system marks the direction of change of the same environmental parameter in consecutive mosquito-attracting calculation cycles as rising, falling, stable, oscillating, or abrupt, and merges consecutive changes with the same direction to form trend segments, thereby generating natural window state records and dynamic state boundaries.

[0020] In S3, the system uses manual response marking and masking to address issues related to mosquito-attracting actions altering environmental data collection. Mosquito-attracting execution records include mosquito-attracting light source activation, odor release, carbon dioxide release, execution intensity, execution time, and readback time. The readback status confirms whether the action actually occurred, wind direction determines whether the carbon dioxide or odor release airflow passed the collection location, and the capture feedback status determines whether changes in capture channel activity correspond to an increase in the actual capture quantity. When a change in ambient light intensity occurs after the mosquito-attracting light source is activated and confirmed by readback, and the ambient light intensity collection location is within the influence range of the mosquito-attracting light source, the system marks the corresponding ambient light intensity change as an artificial light source response and masks its contribution to the light intensity trend segment. When a change in carbon dioxide concentration occurs after carbon dioxide release and confirmed by readback, and the wind direction shows that the release airflow passed the carbon dioxide concentration collection location, the system marks the corresponding carbon dioxide concentration change as an artificial release response and masks its contribution to the carbon dioxide trend segment. When the capture channel activity changes after odor release, but diffusion is not yet complete or the actual capture quantity does not increase, the system marks the corresponding change as channel interference and prohibits it from being used as feedback on the mosquito-attracting effect. The real window state obtained after the above shielding is used to determine the trend prediction state and the mosquito-attracting execution state, so that the mosquito-attracting device's own execution action does not directly affect the natural window judgment.

[0021] In S4, the system performs response phase sequence matching after the mosquito-attracting execution state changes, and uses the response phase sequence matching result and the effective trapped samples for dynamic state boundary updates in the next mosquito-attracting calculation cycle. The system does not perform post-event control on completed mosquito-attracting execution phases, but instead records the sequential relationship between mosquito-attracting actions, readback confirmation, environmental response, diffusion waiting, changes in the active state of the capture channel, and changes in the actual captured quantity. If the mosquito-attracting action includes odor release, the theoretical response phase sequence should include odor release readback confirmation, odor diffusion waiting, identification of changes in the active state of the capture channel, and identification of changes in the actual captured quantity. If the phase sequence in the actual sampling sequence is consistent with the theoretical response phase sequence, the corresponding state transition rules are retained; if the actual response phase sequence has missing phases, incorrect sequence, or inconsistent readbacks, the corresponding processing results are marked as delayed triggering, suspension of environmental parameter contribution, or execution link anomaly. The system saves the real window state, trend prediction state, mosquito-attracting execution state, manual response shielding result, actual captured quantity, equipment energy consumption, and readback state according to the statistical sample window. After removing samples with incomplete data, abnormal capture feedback, or unshielded manual responses, the effective trapped samples are obtained. Effectively captured samples are used in the next mosquito-attracting calculation cycle or the next update of the same state migration rule.

[0022] In summary, this implementation method forms a data processing chain centered on the same objective technical problem through mosquito-attracting calculation cycle alignment, effective stable window screening, manual response stripping, real window state judgment, response stage sequence matching, and effective trapping sample updating. Each process in this chain revolves around the environmental input changes caused by the mosquito-attracting device itself, leading to trend misjudgments: cycle alignment ensures that the input data has the same time basis; stable window qualification review excludes unreliable historical references; manual response stripping distinguishes between natural environmental changes and mosquito-attracting execution responses; execution state determination controls the mosquito-attracting actions in the current or next mosquito-attracting calculation cycle; and feedback correction ensures that subsequent dynamic state boundaries reference the actual capture effect and execution chain state.

[0023] like Figure 2 As shown, in one specific embodiment, S1 includes: S11. Obtain the ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity and sampling time in the target area, and obtain the mosquito trapping execution record, readback status and capture feedback status. S12. Time-align the heterogeneous sampling data of each environmental parameter according to the mosquito attraction calculation cycle, and generate the effective status of the environmental parameters; S13. Perform anomaly removal, calibration correction, missing data processing, and scale unification on the effective states of each environmental parameter to generate a mosquito-attracting environmental state set.

[0024] In this embodiment, it should be noted that in S11, the system records the sampling time for each sampled data point. The sampling time is used to determine whether the environmental change occurred before or after the mosquito-attracting execution. Ambient temperature and humidity reflect the basic environment for mosquito activity, atmospheric pressure reflects changes in the weather background, wind speed is used to determine the conditions for mosquito activity and odor diffusion, wind direction is used to determine which adjacent area the odor release and carbon dioxide release may spread to, carbon dioxide concentration is used to determine the natural chemical attraction background and whether the carbon dioxide supply module causes self-interference, and ambient light intensity is used to distinguish between changes in the natural light background and the mosquito-attracting light source execution response. The mosquito-attracting execution record includes the mosquito-attracting light source activation record, odor release record, carbon dioxide release record, execution intensity, execution time, and readback time. The readback status includes the mosquito-attracting light source output readback, odor release readback, carbon dioxide release readback, and execution status readback. The capture feedback status includes the capture channel activity status and the actual number of captured items. If the carbon dioxide concentration increases after the carbon dioxide release, and the wind direction shows that the release airflow passes through the carbon dioxide concentration collection location, then this increase in carbon dioxide concentration cannot be directly identified as an enhancement of the natural chemical attraction background. Instead, it should be determined in subsequent S3 whether it belongs to an artificial release response.

[0025] In S12, the mosquito-attracting calculation cycle refers to the basic time unit for the mosquito-attracting node to perform a cycle alignment, state generation, and trend judgment on environmental data. The input to the mosquito-attracting calculation cycle is multi-frequency sampled data of ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and ambient light intensity within the target area. The output is the valid state of environmental parameters and the mosquito-attracting environmental state set within the same cycle. The usage phases include data alignment, stable window qualification review, real window state generation, and execution state determination. Its technical role is to ensure that data from different sampling frequencies are incorporated into the same judgment time basis. The value of the mosquito-attracting calculation cycle is jointly determined by the historical sampling interval, execution readback latency, and the controller's processing capacity. The system reads continuously... Daily or no less than Historical environmental data from each valid sampling point were used to calculate the shortest stable change time for ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and ambient light intensity. Then, the longest time from the issuance of control commands to the confirmation readback from the mosquito-attracting light source, odor release module, and carbon dioxide supply module was read. A period not shorter than the longest readback time and not longer than the shortest stable change time was selected as the mosquito-attracting calculation period. For example, the shortest stable change time for ambient light intensity was... The longest readback time for carbon dioxide release is The controller processing time is less than Then it is acceptable. As a mosquito-attracting calculation cycle.

[0026] In S13, the effective state of an environmental parameter refers to the periodic state of the same environmental parameter obtained within a mosquito-attracting calculation cycle after sampling time alignment, anomaly removal, calibration correction, and missing value handling. The mosquito-attracting environmental state set refers to the set of states obtained by transforming the effective states of each environmental parameter within the same mosquito-attracting calculation cycle. The system does not directly add the original values ​​of different physical quantities, nor does it directly compare data from different units. Instead, it converts each environmental parameter into a relative state within the same type of parameter. The system reads the historical stability range of the same type of parameter from the effective stability window, marks the current effective state of the environmental parameter as one or more of the following states: low-level near, moderately stable, high-level near, rising, falling, oscillating, or abrupt change, and then writes the relative states of each environmental parameter into the mosquito-attracting environmental state set. For example, the common range of environmental humidity within the effective stability window is... to The current ambient humidity is If the ambient light intensity increases for two consecutive mosquito attraction calculation cycles and then remains stable, it can be marked as moderately stable and close to the historically suitable range. The ambient light intensity within one mosquito attraction calculation cycle is as follows: , , , , ,in Neither the preceding nor following sampled values ​​showed a sustained increase, and the sampled point was outside the most recent effective stable window. to The system excludes samples from extreme areas caused by vehicle lights or reflections.

[0027] In one specific implementation, calibration correction parameters are obtained from equipment installation and commissioning records and periodic calibration records. During the installation phase, the system uses standard testing equipment or manual calibration records within the same target area to compare the ambient temperature, ambient humidity, atmospheric pressure, wind speed, carbon dioxide concentration, and ambient light intensity data acquisition components. The deviation range between the acquisition component readings and the standard test values ​​is recorded, and corrections are made according to the internal rules for similar parameters during subsequent operation. For example, during commissioning, the carbon dioxide concentration acquisition component continuously... All readings were higher than those of standard testing equipment. to The system uses the aforementioned deviation range as the basis for carbon dioxide concentration calibration correction; in subsequent operation, if the reading of the carbon dioxide concentration acquisition component is... If the system corrects the deviation range corresponding to the debugging record, it will generate an effective carbon dioxide concentration state after correction.

[0028] In one specific implementation, S1 further includes a sampling data reliability record. This sampling data reliability record is not used as an independent control result, but rather as an auxiliary basis for subsequent stability window qualification review. If the ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and ambient light intensity all form a valid environmental parameter state within a certain mosquito-attracting calculation period, and no extreme sampling aggregation occurs, then the data integrity of the mosquito-attracting calculation period is relatively good. If multiple missing markers or multiple extreme sampling rejection markers exist within the same period, then even if the mosquito-attracting calculation period can form a mosquito-attracting environmental state set, it will be subject to stability window qualification review restrictions in S2.

[0029] like Figure 3 As shown, in one specific embodiment, S2 includes: S21. Divide the historical mosquito-attracting environmental status set, historical mosquito-attracting execution record, historical capture feedback status, historical carbon dioxide release record and historical readback status into candidate historical windows, and conduct a stable window qualification review on the candidate historical windows. S22. The candidate historical windows that have passed the eligibility review for stable windows are identified as valid stable windows, and a path for historical status changes is formed based on the valid stable windows. S23. Generate dynamic state boundaries and natural window state records based on the effective stable window and the mosquito-attracting environment state set near the current mosquito-attracting calculation cycle.

[0030] In this embodiment, it should be noted that in S21, the candidate historical window refers to a data window that is divided from historical data in chronological order and awaits the qualification review of a stable window. A valid stable window is a candidate historical window that does not contain consecutive missing data, abnormal sampling clusters, failed mosquito attraction, unprocessed carbon dioxide release self-interference, abnormal capture feedback, or inconsistent readback. The system divides historical data into multiple candidate historical windows in chronological order and reviews the qualification of each one for stability. The length of the candidate historical window is determined based on the mosquito attraction calculation cycle and the daily environmental change rate in the target area. The system uses a length of not less than... No more than A series of consecutive mosquito-attracting calculation cycles are used as candidate historical window lengths, with priority given to time periods that cover a single instance of light intensity decrease, humidity change, wind speed stability, and natural background carbon dioxide variation. For example, the mosquito-attracting calculation cycle is... At that time, the system can Each consecutive mosquito-attracting calculation cycle is used as a candidate historical window, with a window length of [value missing]. ;like Historical data includes Each mosquito-attracting calculation cycle can be divided into several time periods. A candidate history window.

[0031] In S22, candidate historical windows that pass the stable window eligibility review are determined as valid stable windows. When the system performs the stable window eligibility review, it first checks for consecutive missing values, then for anomalous sampling clusters, and finally checks for consistency between execution records, readback states, and capture feedback. The boundary for determining consecutive missing values ​​is determined by the missing value distribution in recent candidate historical windows. The system statistically analyzes recent... The number of periods within a candidate historical window where each environmental parameter has not formed a valid environmental parameter state, if the same environmental parameter is continuously within the same candidate historical window. A mosquito-attracting calculation cycle is missing, or there are mosquitoes simultaneously within the same mosquito-attracting calculation cycle. If a certain environmental parameter is missing, the candidate historical window is marked as a continuously missing window. For example, if the environmental humidity in a candidate historical window is missing on the 1st... The and the first If no valid state is established within any of the mosquito-attracting calculation cycles, this window is not considered a valid stable window even if other parameters are complete. The boundary for determining abnormal sampling clusters is determined by the historical stable range of similar environmental parameters and the number of extreme sampling rejections. If the same environmental parameter is rejected within a candidate historical window... More than once, or at least If all environmental parameters show extreme sampling rejection flags, it is considered an abnormal sampling cluster.

[0032] In S23, the dynamic state boundary is the specific state transition manifestation of the dynamic threshold in this application. It is used to determine whether there is a switch between the natural window state, the real window state, and the trend prediction state. The input to the dynamic state boundary is the effective stable window, the background window trend, the neighboring window trend, the artificial response shielding result, the response stage sequence matching result, and the effective trapping sample. The generation method is to adjust the conditions for entering the preparatory state or the working state based on the state transition record and its capture feedback result. The output result is the state transition condition. The usage stages are the generation of the natural window state record, the generation of the real window state, and the determination of the mosquito-trapping execution state. The technical role is to replace the fixed switching threshold. The natural window state record refers to the trapping window state record formed by the environmental change trend before deducting the artificial response. The natural window state record includes a state label, the effective state of the environmental parameters involved in forming the state label, the trend segment, the sampling time, and the artificial response candidate label. The state label includes natural window forming, natural window not yet formed, instantaneous impact, oscillation disturbance, and continuous distance.

[0033] In one specific implementation, the system generates a background window trend based on the continuous change path of the mosquito-attracting environmental state set within an effective stable window, and generates a neighboring window trend based on the mosquito-attracting environmental state set within the current mosquito-attracting calculation cycle and its neighboring calculation cycles. The background window trend reflects the dominant changes in the target area within a historical range, while the neighboring window trend reflects the immediate changes when approaching the trapping window. The system converts the direction of change in the mosquito-attracting environmental state within consecutive mosquito-attracting calculation cycles into trend symbols, and merges consecutive identical trend symbols into trend segments. Trend symbols include rising, falling, stable, oscillating, and abrupt changes. Rising indicates that the relative state of the same environmental parameter continuously increases in adjacent mosquito-attracting calculation cycles; falling indicates that the relative state of the same environmental parameter continuously decreases in adjacent mosquito-attracting calculation cycles; stable indicates that the relative state of the same environmental parameter remains within the same state range across multiple mosquito-attracting calculation cycles; oscillating indicates that the direction of change of the same environmental parameter repeatedly switches within a short period of time; abrupt changes indicate that the same environmental parameter exceeds the historical stable range of the same type within a single or a few mosquito-attracting calculation cycles.

[0034] In one specific implementation, the dynamic state boundary is not based on the original value of a single environmental parameter as a fixed threshold, but rather on state transition records as the adjustment object. The system reads the state transition records before the formation of the natural window within the effective stable window, and then... In a number of valid trapping samples, statistically analyze the changes in the number of traps and the changes in equipment energy consumption corresponding to the same state migration records; if not less than If the actual number of samples increases and there is no readback inconsistency, then the state transition record is added to the state transition condition for entering the preparatory state; if it is not less than If a sample corresponds to an increase in device energy consumption but no increase in the number of samples captured, then that state transition record is added to the state transition conditions for delayed execution. For example, if a state transition record shows a continuous decrease in ambient light intensity, followed by an increase and then a stable ambient humidity, and a stable wind speed with a slowly approaching natural background carbon dioxide, and multiple instances of an increase in the number of samples captured, then subsequent identical state transition records can enter the preparatory state.

[0035] like Figure 4 As shown, in one specific embodiment, S3 includes: S31. Based on the mosquito attraction execution record, readback status, wind direction and capture feedback status, mark the artificial light source response, artificial release response, channel interference and artificial attraction feedback in the natural window status record; S32. Obtain the true window state by masking the trend contributions corresponding to artificial light source response, artificial release response and channel interference from the natural window state record; S33. Generate trend prediction state based on the real window state and dynamic state boundary, and determine the mosquito-attracting execution state.

[0036] In this embodiment, it should be noted that in S31, the artificial response includes artificial interference response and artificial attraction feedback. Artificial interference response includes artificial light source response, artificial release response, and channel interference. Artificial attraction feedback includes capture feedback formed after effective mosquito attraction execution. The real window state masks the trend contribution corresponding to the artificial interference response, but does not mask the artificial attraction feedback; the artificial attraction feedback, as the capture effect after mosquito attraction execution, enters the record of the sample to be screened, and forms an effective sample after being screened by the statistical sample window in S4. The mosquito attraction execution record includes the mosquito attraction light source activation record, odor release record, carbon dioxide release record, execution intensity, execution time, and readback time. The effective wind direction state is used to determine whether the released airflow will affect the carbon dioxide concentration sampling location, and also to determine the odor diffusion direction. The capture feedback state includes the capture channel activity state and the actual number of captured samples. When the ambient light intensity change occurs after the mosquito attraction light source is activated and confirmed by readback, and the ambient light intensity sampling location is within the influence range of the mosquito attraction light source, the system marks the corresponding ambient light intensity change as an artificial light source response.

[0037] In S32, the true window state refers to the trapping window state obtained by filtering out the trend contributions of artificial light source response, artificial release response, and channel interference from the natural window state record. The inputs to the true window state are the natural window state record, mosquito-attracting execution record, readback state, wind direction, and capture feedback state. The output is the window state after eliminating the influence of artificial interference response. Its usage phases are trend prediction state generation and mosquito-attracting execution state determination. The technical function is to prevent the mosquito-attracting device's own actions from affecting trend judgment. The system obtains the true window state by filtering out the trend contributions of artificial light source response, artificial release response, and channel interference from the natural window state record. Artificial light source response does not participate in the judgment of natural light background changes; artificial release response does not participate in the judgment of natural carbon dioxide background changes; channel interference does not participate in the judgment of actual capture effect. Artificial attraction feedback is saved as the mosquito-attracting execution result but does not negatively affect the natural window state record.

[0038] In S33, the trend prediction state is the mosquito-attracting trend judgment result generated based on the actual window state, dynamic state boundary, and continuous state transition requirements. It includes a suitable mosquito-attracting prediction state, a near-suitable mosquito-attracting prediction state, and an unsuitable mosquito-attracting prediction state. The mosquito-attracting execution state is the execution result determined based on the trend prediction state, including a working state, a ready state, and an energy-saving dormant state. The trend prediction state and the mosquito-attracting execution state belong to different levels. The trend prediction state is used for pre-adjustment in the current execution phase or for determining the execution state in the next mosquito-attracting calculation cycle. The mosquito-attracting execution state is used for the execution phase controlling the mosquito-attracting light source, odor release, and carbon dioxide release. When the actual window state indicates that a natural window is forming, and the artificial light source response, artificial release response, and channel interference have been shielded, the system generates a suitable mosquito-attracting prediction state and determines the mosquito-attracting execution state as the working state in the current execution phase or the next mosquito-attracting calculation cycle. When the actual window state indicates that a natural window has not yet formed, but the trend of neighboring windows is continuously approaching the dynamic state boundary, and the background window trend has not reversed, the system generates a near-suitable mosquito-attracting prediction state and determines the mosquito-attracting execution state as the ready state in the next mosquito-attracting calculation cycle. When the actual window state indicates instantaneous impact, oscillation disturbance, continuous distance from the dynamic state boundary, or artificial light source response, artificial release response, or incomplete shielding of channel interference, the system generates an unsuitable mosquito-attracting prediction state and determines the mosquito-attracting execution state as either an energy-saving sleep state or maintains the previous effective mosquito-attracting execution state.

[0039] In one specific implementation, the effective wind direction status is determined jointly by wind direction continuity, wind speed range, and relative node positions. The system converts wind directions that appear frequently within the same mosquito-attracting calculation cycle into dominant diffusion directions and checks whether these dominant diffusion directions are continuous. More than one mosquito-attracting calculation cycle points to the carbon dioxide concentration collection location or the direction of adjacent mosquito-attracting nodes; simultaneously, the system reads the available wind speed range for the corresponding target area within the effective stable window. If the current wind speed is... , in to If the wind speed is within the historical available wind speed range and the wind direction continuously points downwind, it is marked as effective propagation. The capture channel activity status is state-type data in the capture feedback status. Its input includes at least one of the following: capture channel occlusion event, occlusion duration, adjacent occlusion interval, capture chamber image count result, and capture chamber weight change result. The system first excludes false triggers caused by excessively short occlusion duration, abnormally dense adjacent occlusion intervals, no new capture targets appearing in the image, or weight changes inconsistent with the common mosquito weight range. Then, based on the number and duration of remaining valid trigger events within the statistical sample window, the capture channel activity status is marked as inactive, low-activity, continuously active, or abnormally active.

[0040] In one specific implementation, the diffusion waiting state is determined based on the release type, the distance from the release source to the capture channel or collection location, the current wind speed, and the feedback lag time in historical valid samples. The system first reads the execution time of odor or carbon dioxide release, then reads the wind speed range and the distance from the release source to the target location within the most recent valid stabilization window, and uses the earliest stabilization time in historical valid samples where the capture channel's active state began to change as the shortest waiting period. For example, if the wind speed is... to Odor release source distance from capture channel Historical valid samples show that the active state of the capture channel is usually after release. to If a change occurs, the odor diffusion will be acceptable. Changes in channel activity that occur before completion are flagged as channel interference if no new captures are added to support them.

[0041] In one specific implementation, the continuous state transition requirement is determined by the state duration within the most recent effective stable window. The system counts the number of duration periods for each combination of environmental states within the effective stable window before the natural window is formed, and selects the shorter but recurring duration period as the switching requirement. If the most recent There are 1 effective stable window Each window needs at least [a certain number of windows] before a natural window can be formed. If, during a continuous mosquito-attracting calculation cycle, ambient light intensity decreases, ambient humidity remains constant, wind speed is stable, and artificial response is blocked, then the current operation must also meet the continuous requirements. The system switches from the standby state to the working state only after one mosquito-attracting calculation cycle. Brief changes in a single cycle are only recorded and do not directly trigger the working state.

[0042] In one specific implementation, the output intensity of the low-output preheating state is determined jointly by the operating intensity range of the mosquito-attracting light source, historical effective trapping samples, and artificial light source response records. The system reads the commonly used output levels of the mosquito-attracting light source under its operating state and compares whether a continuous artificial light source response is formed at the ambient light intensity sampling location under different low output levels. If the commonly used output in the operating state is... Historical samples show The output only causes a short-term change in light intensity. If the output for two consecutive mosquito-attracting calculation cycles is marked as an artificial light source response, then the low output preheating state in the preparatory state is taken. The low-frequency parameters of odor release are determined by the release interval range supported by the odor release module, the diffusion waiting state, and the artificial induction feedback records. If the odor release module supports each... , , Released once, historical samples show each There is a lot of channel interference during release, each Release capture feedback lag, every If both the active state of the capture channel and the actual number of captures can be recorded within the statistical sample window during release, then the preparatory state uses a per A single low-frequency release.

[0043] like Figure 5 As shown, in one specific embodiment, S4 includes: S41. Perform response phase sequence matching after the mosquito-attracting execution state changes; S42. Update the dynamic state boundary based on the effective trapped samples; S43. Correct the preparatory state triggering rules of adjacent mosquito-attracting nodes based on the wind direction propagation response.

[0044] In this embodiment, it should be noted that in S41, the execution response stage sequence matching result refers to the sequence matching result between the theoretical response stage sequence and the actual response stage sequence. The inputs to the execution response stage sequence matching result are the mosquito-attracting action, execution readback, environmental response, diffusion wait, capture channel activity state, and actual capture quantity. The generation method is to compare the theoretical and actual stages in chronological order. The output results include sequence consistency, stage missing, sequence error, and readback inconsistency. The stage used is dynamic state boundary update, and its technical function is to identify execution link anomalies and response lags. If the mosquito-attracting action includes enhancing the mosquito-attracting light source output, the theoretical response stage sequence includes mosquito-attracting light source readback confirmation, ambient light intensity response identification, capture channel activity state change identification, and actual capture quantity change identification. If the mosquito-attracting action includes odor release, the theoretical response stage sequence includes odor release readback confirmation, odor diffusion wait, capture channel activity state change identification, and actual capture quantity change identification. If the mosquito-attracting action includes carbon dioxide release, the theoretical response stage sequence includes carbon dioxide release readback confirmation, artificial release response processing, carbon dioxide diffusion wait, capture channel activity state change identification, and actual capture quantity change identification.

[0045] In S42, effective trapping samples are data records saved through a statistical sample window and filtered for use in the next mosquito-attracting calculation cycle or the next update of the same type of state migration rules. The system records the actual window state, trend prediction state, mosquito-attracting execution state, manual response shielding results, actual capture quantity, equipment energy consumption, and readback status according to the statistical sample window. After removing samples with incomplete data, abnormal capture feedback, or unshielded manual responses, the effective trapping samples are obtained. Equipment energy consumption is the electrical energy consumed by the mosquito-attracting node during the mosquito-attracting action within the statistical sample window. Its inputs are the operating time and corresponding power records of the mosquito-attracting light source, odor release module, carbon dioxide supply module, negative pressure capture structure, communication unit, and control unit. The output unit is... or The system updates the dynamic state boundary based on changes in the number of captures and changes in device energy consumption corresponding to the same state migration records in the effective trapping samples. For example, in the recent... In a valid trapped sample, a certain state transition record has If the actual number of samples increases and there is no readback inconsistency, then the state transition record is added to the state transition condition for entering the preparatory state.

[0046] In S43, wind direction propagation response is used for coordinated preparation among multiple mosquito-attracting nodes. The distance range between adjacent mosquito-attracting nodes is determined based on the target area area, node installation spacing, communication accessibility, and effective wind direction propagation response records. During the installation phase, the system records the positional relationships of each mosquito-attracting node. During the operation phase, it statistically analyzes whether downwind nodes show a close proximity to their local real-window status and an increase in the actual number of mosquitoes captured after propagation alerts. If the node spacing in a residential or garden area is... to Historical samples show that downwind nodes within this distance range can exhibit an increase in the number of captures within the statistical sample window, therefore... to The propagation prompt serves as the basis for determining adjacent mosquito-attracting nodes. If two nodes fail to respond to multiple capture feedbacks during propagation prompts, they will no longer be considered adjacent mosquito-attracting nodes for propagation prompts. The validity period of the propagation prompt is determined by the duration of stable wind direction, the distance between nodes, and historical wind direction propagation response records. When sending a propagation prompt, the system records the sending time, the actual window status of the source node, the current effective wind direction status, and the node positional relationship, and statistically analyzes the time range within which downwind nodes show changes in capture channel activity in historical effective propagation samples. For example, the distance between upwind and downwind nodes... Historical effective propagation samples show that downwind nodes are usually in to If a change in the activity status of the capture channel occurs, the propagation prompt's validity period is valid. After this time, the notification will only be saved as a historical record.

[0047] In one specific implementation, the response waiting period is adjusted based on the actual lag time of the most recent valid response sample. The system reads the most recent... The sequence matching results for each response stage are considered valid samples with consistent order. The waiting time between the execution of an action and the corresponding environmental response or capture feedback change is statistically analyzed, and the recurring and relatively stable waiting times are used as the response waiting period for the next stage. For example, the most recent valid samples show that the capture channel activity level after odor release is typically... It appears later, and the current waiting period is... The system will then diffuse the odor and wait for it to be released by... Adjusted to .

[0048] In one specific implementation, the determination of a valid increase in the actual capture quantity is performed on a statistical sample window basis. The system compares the number of new captures in the current statistical sample window with that in the previous statistical sample window, and checks for source conflicts between the capture chamber image count, infrared channel count, and estimated capture chamber weight change. If the number of new captures in the current window is higher than that in the previous window, and at least one automatic capture feedback source is valid, and there are no capture feedback anomaly flags, then the increase in the actual capture quantity is considered valid. For example, if the number of new captures in the previous statistical sample window was... Only, the current statistical sample window image count has been increased. If the number of valid events in the infrared channel also increases, and there are no abnormalities in weight changes due to rain or debris, then the current window can be considered as having an effective increase in the number of captures. The number entered through manual inspection is only used for statistical sample window correction and is not used for trend prediction status switching for a single mosquito-attracting calculation cycle.

[0049] In one specific implementation, the dynamic state boundary update in this implementation does not employ a training model or model inference. The system uses a limited number of effective trapping samples as its object, and forms state transition records based on the actual window state, trend prediction state, mosquito-trapping execution state, manual response shielding results, response phase sequence matching results, actual capture quantity changes, and equipment energy consumption changes. If a certain type of state transition record corresponds to an increase in the capture quantity and consistent readback state across multiple effective trapping samples, the priority for this state transition record to enter the preparatory or working state is increased. If a certain type of state transition record corresponds to multiple increases in energy consumption but no increase in the capture quantity, this state transition record is marked as having its environmental parameter contribution delayed or suspended. Therefore, subsequent dynamic state boundary updates have clearly defined inputs, processing rules, and output results.

[0050] like Figure 6 As shown, the present invention also provides a mosquito-attracting trend prediction device based on dynamic thresholds, including a data alignment module, a stable boundary module, a response stripping module, an execution determination module, and a feedback correction module. The data alignment module is used to perform the data acquisition, periodic alignment, anomaly removal, and mosquito-attracting environment state set generation corresponding to S1 above; the stable boundary module is used to perform the candidate historical window division, stable window qualification review, effective stable window generation, dynamic state boundary generation, and natural window state record generation corresponding to S2 above; the response stripping module is used to perform the artificial light source response, artificial release response, and channel interference marking and shielding corresponding to S3 above; the execution determination module is used to determine the trend prediction state and mosquito-attracting execution state based on the real window state and the dynamic state boundary; the feedback correction module is used to perform the response stage sequence matching, effective trapping sample generation, and dynamic state boundary update corresponding to S4 above.

[0051] like Figure 7 As shown, this invention also provides a mosquito-attracting trend prediction system based on dynamic thresholds, including multiple mosquito-attracting nodes, a data storage unit, and a display terminal. Each mosquito-attracting node includes an environmental data acquisition unit, a mosquito-attracting execution unit, a capture feedback unit, a node processing unit, and a node communication unit. The environmental data acquisition unit is used to acquire ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within the target area. The mosquito-attracting execution unit is used to execute the activation of the mosquito-attracting light source, odor release, and carbon dioxide release, and generate mosquito-attracting execution records and readback status. The capture feedback unit is used to generate capture feedback status. The node processing unit is used to execute the above method steps. The node communication unit is used to transmit the real window status, effective wind direction status, positional relationship, and propagation cues between adjacent mosquito-attracting nodes. The data storage unit is used to save the mosquito-attracting environmental state set, effective stable window, dynamic state boundary, natural window status record, real window status, artificial response shielding result, response stage sequence matching result, effective trapping sample, and propagation cue record. The display terminal is used to show the mosquito trapping execution status, trend prediction status, actual number of mosquitoes captured, device energy consumption, and abnormal prompts. It does not directly generate a real window status or directly switch the mosquito trapping execution status.

[0052] To further illustrate the operation of the technical solution of the present invention, the following describes the processing logic of the mosquito attraction trend prediction method, device and system based on dynamic threshold, using a scenario containing specific parameters and data.

[0053] Deployment in a residential garden target area There are [number] mosquito-attracting nodes, with a spacing of [distance] between each node. to The target area is Each mosquito-attracting node includes components for collecting ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, and ambient light intensity; a mosquito-attracting light source; an odor release module; a carbon dioxide supply module; a capture channel; and an image counting component for the capture chamber. The mosquito-attracting calculation cycle is set to... The statistical sample window is set to The system retrieves the past. Historical data for the day, including Each mosquito-attracting calculation cycle. Because each Each consecutive mosquito-attracting calculation cycle constitutes a candidate historical window, from which the system obtains... A candidate history window. The system records it in one of them. Some candidate historical windows did not undergo self-interference processing after carbon dioxide release. There are abnormalities in the capture feedback of one candidate history window. Several candidate historical windows have consecutive missing values, therefore these candidate historical windows are excluded, leaving... Several candidate historical windows were used as effective stable windows. The data values ​​in this scenario are within the common range for nighttime mosquito attraction monitoring in residential areas and gardens, and are used to illustrate the executable process of the method steps.

[0054] In S1, the system is to Environmental data for mosquito-attracting nodes upwind is obtained from the statistical sample window. The ambient temperature is determined by... Descending to The ambient humidity is from Rise to Atmospheric pressure remains to The wind speed remained at to The wind direction was northeast, and the natural background carbon dioxide concentration was... Approach slowly Ambient light intensity from Descending to Because the mosquito attraction calculation cycle is The system divides the statistical sample window into Each mosquito-attracting cycle is calculated, and the sampling data of different frequencies within each cycle are aligned. If the ambient light intensity appears in a certain cycle... The isolated sampling points, while the sampling points before and after are respectively and The system will Sampling points were eliminated to avoid misjudging light intensity trends due to vehicle lights or reflections. After processing, the system obtained a set of mosquito-attracting environmental states, including decreasing ambient temperature, increasing and then maintaining ambient humidity, stable atmospheric pressure, stable wind speed, wind direction allowing for propagation, slowly approaching carbon dioxide concentration, and decreasing ambient light intensity.

[0055] In S2, the system sets the current mosquito-attracting environment state together with... The state change paths within each effective stable window were compared. The common mosquito-attracting window formation process recorded within the effective stable window is as follows: ambient light intensity at... The number of consecutive mosquito-attracting cycles decreased, and the ambient humidity was at... to The range remains constant, and the wind speed is within the range. to Within the range, there was no sudden increase in natural background carbon dioxide after release, and no records of unmasked artificial responses were found. In the current statistical sample window, the ambient light intensity decreased from [previous range]. Down to Humidity from Rise to The wind speed remained at to The carbon dioxide concentration is from near This is similar to the state transition records within an effective stable window. Based on this, the system generates natural window state records and marks the state as a natural window in formation. Simultaneously, the system records the effective states, trend segments, sampling times, and candidate markers for human responses of the environmental parameters involved in forming the state markers, providing locatable data objects for human response stripping in S3.

[0056] In S3, the upwind mosquito-attracting node is located at Entering the ready state, the mosquito-attracting light source is in use Low output preheating, odor release module per A low-frequency release occurs; the carbon dioxide supply module will not release carbon dioxide temporarily. Ambient light intensity is... Later by Rise to The system checked the mosquito-attracting light source activation record and found that the change occurred after the mosquito-attracting light source was activated and confirmed, and the ambient light intensity sampling location was at a distance from the mosquito-attracting light source. Therefore, this change is labeled as an artificial light source response, and its contribution to the light intensity trend segment is masked. The odor release module in... Perform low-frequency release to capture channel activity status. The number started to increase, and the actual number captured was in to The statistical results are from the previous window Only increased to Only when the change in the active state of the capture channel occurs after the diffusion wait is complete and the actual number of captured samples increases, will the system mark the corresponding change as artificial induction feedback and write it into the record of samples to be screened for trapping. At this time, the artificial light source response has been masked from the natural window state record, and the capture feedback after odor release is saved as the execution effect, without contaminating the natural window state.

[0057] In the subsequent execution of S3, the carbon dioxide supply module... Carbon dioxide release begins. The effective wind direction indicates that the released airflow may pass through the carbon dioxide concentration sampling location; the carbon dioxide concentration will decrease after release. Rise to The system checks the carbon dioxide release records and readback status, confirming that the concentration increase occurred after the carbon dioxide release and readback confirmation, and that the wind direction shows the released airflow passed through the carbon dioxide concentration collection location. Therefore, this carbon dioxide concentration change is marked as an artificial release response, and its contribution to the carbon dioxide trend segment is masked. After masking, the true window status is still determined based on the decrease in light intensity, maintained humidity, stable wind speed, and natural background changes in carbon dioxide before the release in the natural environment, without including... The concentration of the released carbon dioxide directly enhances the natural chemical attraction background. Therefore, the system maintains the trend prediction state as a suitable mosquito-attracting prediction state and determines the mosquito-attracting execution state as the working state, rather than repeatedly triggering the preparatory state or erroneously increasing the release intensity due to a sudden increase in carbon dioxide after release.

[0058] In S4, the system... to The execution process involves matching the response phase sequence. The theoretical response phase sequence corresponding to odor release is: odor release readback confirmation, odor diffusion waiting, identification of changes in the active state of the capture channel, and identification of changes in the actual captured quantity. In the actual response phase sequence, odor release readback confirmation occurs... The odor is spreading and waiting to be collected. The change in the active state of the capture channel begins at The actual number captured was The statistical results are from Only increased to Therefore, the actual response phase sequence is consistent with the theoretical response phase sequence, and the system retains this state transition rule. The theoretical response phase sequence corresponding to carbon dioxide release is: carbon dioxide release readback confirmation, artificial release response processing, carbon dioxide diffusion waiting, identification of changes in the active state of the capture channel, and identification of changes in the actual captured quantity. In the actual response phase, the carbon dioxide concentration changes from... Rise to The changes have been marked as artificial release responses and masked by the system; they are not used as a basis for the formation of natural windows. After the above records enter the statistical sample window, they form valid decoy samples because the data is complete, the readback is consistent, the capture feedback is effective, and the artificial response has been masked.

[0059] In a multi-node collaborative scenario, the downwind mosquito-attracting node is located approximately downwind of the upwind mosquito-attracting node. When the upwind mosquito-attracting node is When the mosquito-attracting prediction system is in a suitable state and the wind direction indicates that the odor diffusion direction is from the upwind area to the downwind area, the upwind mosquito-attracting node sends a propagation alert to the downwind mosquito-attracting node. The validity period of the propagation alert is based on historical valid propagation samples. Downwind mosquito-attracting nodes read local data within the validity period of the propagation indication, and the local ambient light intensity decreased. Down to The ambient humidity is from Rise to Wind speed is Furthermore, since manual responses have been disabled, the system determines that the local real-world window state supports natural window proximity, and therefore enters the preparation state. In subsequent statistical sample windows, the actual number of mosquitoes captured by downwind mosquito-attracting nodes increased from... Only increased to If the system confirms that the wind direction propagation response is valid, and if subsequent propagation prompts do not correspond to an increase in the number of local captures, the propagation prompt rule between nodes will be marked as delayed triggering or suspended.

[0060] The computational logic in this implementation mainly manifests in periodic division, window filtering, manual response marking, state transition condition updating, and response stage sequence matching. Taking candidate historical window division as an example, the system retrieves... Historical data and As a mosquito attraction calculation cycle, daily includes One mosquito-attracting calculation cycle, The Japanese Communist Party includes Each mosquito-attracting calculation cycle; the system will calculate each Each mosquito-attracting calculation cycle is used as a candidate historical window, thus obtaining Each candidate historical window has several windows. This windowing approach doesn't treat all data from a single day as a single entity, nor does it use a single period as the unit for historical judgment. Instead, it ensures that each candidate historical window covers... The continuous changing process can include a decrease in light intensity, changes in humidity, stable wind speed, and changes in the natural background carbon dioxide. Subsequently, the system excludes... One unprocessed window of carbon dioxide release response, A capture feedback exception window and A series of missing windows, retain A valid stable window is established. This process uses data integrity, execution reliability, and manual response status as admission criteria for historical reference data to avoid abnormal historical windows affecting dynamic state boundaries.

[0061] During the manual response marking and masking process, the system does not simply delete all executed data. Instead, it determines whether data should participate in the natural window judgment based on execution records, readback status, sampling time, wind direction, and capture feedback status. Taking ambient light intensity as an example, ambient light intensity is determined by… Rise to The change occurred in the mosquito-attracting light source After low-output preheating, and with the light intensity collection position at a distance from the mosquito-attracting light source... Therefore, this change satisfies the response conditions of an artificial light source, and its contribution to the light intensity trend range should be masked. Taking carbon dioxide concentration as an example, the carbon dioxide concentration changes from... Rise to The change occurred after the carbon dioxide supply module released and confirmed the data, and the wind direction indicated that the released airflow passed the sampling location. Therefore, this change meets the artificial release response conditions and is not included in the natural carbon dioxide background judgment. In contrast, the activity level of the capture channel increases after the odor release and diffusion wait is complete, and the actual capture quantity increases from... Only increased to However, this change is a result of artificially induced feedback and should be included in the record of samples to be screened as an execution effect, rather than being deleted as channel interference. This process enables the system to distinguish between data collection contamination caused by the execution action and effective trapping feedback generated by the execution action.

[0062] During dynamic state boundary updates, the system uses state transition records as objects, rather than the original values ​​of single environmental parameters. If recently... Of the effective trap samples, there are If a sample corresponds to a state transition record of continuously decreasing ambient light intensity, rising and then maintaining ambient humidity, stable wind speed, and shielded artificial response, and these samples all correspond to an increase in the actual number of captures and consistent readback states, then the system adds this state transition record to the state transition conditions for entering the preparatory state. If another state transition record is in... Among the effective trapping samples If a sample corresponds to an increase in device power consumption but no increase in the number of captures, the system adds it to the state transition condition for delayed execution. This update method is based on changes in the number of captures, changes in device power consumption, and readback status in a finite number of samples, and can be implemented through ordinary data tables, state records, and rule tables, without requiring an uninterpretable training process. The response phase sequence matching also adopts a deterministic approach; the system only compares the time order of the theoretical phase with the actual phase. When a phase is missing, the sequence is incorrect, or the readback is inconsistent, the sample is excluded or marked as an execution link anomaly.

[0063] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various simple modifications can be made to the technical solution of the present invention, and these simple modifications all fall within the protection scope of the present invention.

[0064] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present invention will not describe the various possible combinations separately.

[0065] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed by the present invention.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for predicting mosquito attraction trends based on dynamic thresholds, characterized in that the method... include: The system acquires ambient temperature, humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within the target area, and also acquires mosquito trapping execution records, readback status, and capture feedback status. The sampling data of each environmental parameter were time-aligned according to the mosquito attraction calculation cycle, missing samples and isolated abnormal samples were removed, and the data were converted into a mosquito attraction environmental state set according to the historical stable range of the same type. The historical mosquito-attracting environmental state set, historical mosquito-attracting execution record, historical capture feedback state and historical readback state are divided into candidate historical windows. Candidate historical windows with continuous missing data, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback or unprocessed carbon dioxide release response are excluded to obtain effective stable windows. Dynamic state boundaries and natural window state records are generated based on the effective stable window and the current mosquito-attracting environment state set. Based on the mosquito trapping execution record, readback status, wind direction, and capture feedback status, artificial light source response, artificial release response, and channel interference in the natural window status record are marked and masked to obtain the true window status. The trend prediction state and mosquito attraction execution state are determined based on the actual window state and the dynamic state boundary.

2. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, The sampling data of each environmental parameter were time-aligned according to the mosquito attraction calculation cycle, missing samples and isolated outliers were removed, and the data were converted into a mosquito attraction environmental state set according to the historical stable range of the same type, including: For ambient temperature and humidity, within the same mosquito-attracting calculation cycle, missing samples and samples that exceed the historical stable range of the same type and do not show the same direction of change between sampling points are removed to generate a slowly changing representative state. For atmospheric pressure, a background representative state is generated based on the continuous change direction of adjacent mosquito-attracting calculation cycles; For wind speed and carbon dioxide concentration, isolated abnormal samples with sudden increases or decreases at a single sampling point and no synchronous changes at adjacent sampling points are removed to generate a representative state resistant to mutations. For ambient light intensity, extreme samples caused by obstruction, reflection or sudden light sources are removed to generate a representative light intensity state; Regarding wind direction, the dominant diffusion direction is generated based on the wind direction that appears most frequently and continuously within the same mosquito-attracting calculation cycle. The slowly changing representative state, background representative state, anti-mutation representative state, light intensity representative state, and dominant diffusion direction state are converted into mosquito-attracting environmental states according to their relative positions, change directions, and stability within the same type of parameters, and a mosquito-attracting environmental state set is formed.

3. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, Historical mosquito-attracting environmental state sets, historical mosquito-attracting execution records, historical capture feedback states, and historical readback states were divided into candidate historical windows. Candidate historical windows with consecutive missing values, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback, or unprocessed carbon dioxide release responses were excluded to obtain effective stable windows, including: The historical mosquito-attracting environmental state set, historical mosquito-attracting execution record, historical capture feedback state, historical carbon dioxide release record, and historical readback state are divided into candidate historical windows according to time order. When the same environmental parameter in the candidate history window fails to form a valid state for multiple consecutive mosquito-attracting calculation cycles, or when multiple environmental parameters are missing simultaneously in the same mosquito-attracting calculation cycle, it is marked as consecutively missing. When the extreme sampling rejection flags of the same type of environmental parameter in the candidate history window reach a predetermined number, or when multiple environmental parameters show extreme sampling rejection flags, it is marked as an abnormal sampling cluster. When the mosquito-attracting execution command is inconsistent with the readback status, the capture feedback source conflicts, the artificial release response processing is not completed after the release of carbon dioxide, or the theoretical response stage after the execution action is missing, the corresponding candidate historical window is excluded. Candidate historical windows that were not excluded were identified as valid stable windows.

4. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, Based on the effective stable window and the current mosquito-attracting environment state set, dynamic state boundaries and natural window state records are generated, including: The background window trend is generated based on the continuous change path of the mosquito-attracting environmental state set in the effective stable window; A neighboring window trend is generated based on the current mosquito attraction calculation cycle and the mosquito attraction environment state set within its neighboring mosquito attraction calculation cycles. The direction of change of the same environmental parameter in a continuous mosquito-attracting calculation period is marked as rising, falling, stable, oscillating, or abrupt, and continuous changes in the same direction are merged into trend segments; A natural window state record is generated based on the background window trend, the trend of adjacent windows, the number of duration periods of the trend segment, the switching order of the trend segment, and the sequential relationship between trend segments of different environmental parameters. The natural window state record includes state labels, valid states of environmental parameters involved in forming the state labels, trend segments, sampling time, and candidate markers for human responses. The dynamic state boundary is used to determine whether there is a switch between the natural window state, the real window state, and the trend prediction state, and is not used as a fixed numerical threshold for a single environmental parameter.

5. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, Based on mosquito trapping execution records, readback status, wind direction, and capture feedback status, artificial light source responses, artificial release responses, and channel interference in the natural window status records are marked and masked to obtain the true window status, including: When the change in ambient light intensity occurs after the mosquito-attracting light source is activated and confirmed by readback, and the ambient light intensity acquisition location is within the influence range of the mosquito-attracting light source, the corresponding change in ambient light intensity will be marked as an artificial light source response, and its contribution to the light intensity trend segment in the natural window state record will be masked. When a change in carbon dioxide concentration occurs after a carbon dioxide release and readback confirmation, and the wind direction shows that the released airflow passes through the carbon dioxide concentration collection location, the corresponding change in carbon dioxide concentration is marked as an artificial release response, and its contribution to the carbon dioxide trend segment in the natural window state record is masked. When the active state of the capture channel changes after the odor is released, and the diffusion wait is not completed or the actual capture number does not increase, the corresponding change will be marked as channel interference and prohibited from being used as feedback on the mosquito attraction effect. When the capture channel activity increases after the diffusion wait is complete and the actual capture quantity increases, the corresponding change is marked as artificial induction feedback and written into the record of the samples to be screened for capture. The true window state is obtained by masking the trend contributions marked as artificial light source response, artificial release response and channel interference in the natural window state record.

6. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, The trend prediction state and mosquito attraction execution state are determined based on the actual window state and the dynamic state boundary, including: When the real window state indicates that a natural window is forming and artificial light source response, artificial release response and channel interference have been shielded, a suitable mosquito-attracting prediction state is generated, and the mosquito-attracting execution state is determined as the working state in the current execution phase or the next mosquito-attracting calculation cycle. When the true window state indicates that the natural window has not yet been formed, but the trend of the neighboring windows is continuously approaching the dynamic state boundary and the trend of the background window has not reversed, a state close to the mosquito-attracting prediction is generated, and the mosquito-attracting execution state is determined as the preparatory state in the next mosquito-attracting calculation cycle. When the actual window state indicates instantaneous impact, oscillation disturbance, continuous distance from the dynamic state boundary, or artificial light source response, artificial release response, or incomplete shielding of channel interference, an unsuitable mosquito-attracting prediction state is generated, and the mosquito-attracting execution state is determined as either an energy-saving dormant state or maintaining the previous effective mosquito-attracting execution state. The trend prediction status is used for pre-adjustment of the current execution phase or for determining the mosquito-attracting execution status of the next mosquito-attracting calculation cycle, and is not used to modify mosquito-attracting execution phases that have already ended.

7. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, After determining the mosquito-attracting implementation status, the following is also included: After the mosquito-attracting execution state switches from energy-saving dormancy state to standby state or working state, a theoretical response stage sequence is generated based on the activation of the mosquito-attracting light source, odor release or carbon dioxide release. Identify readback confirmation, environmental response, diffusion wait, changes in the active state of the capture channel, and changes in the actual capture quantity from the actual sampling sequence to generate an actual response phase sequence; When the actual response phase sequence is consistent with the theoretical response phase sequence, the corresponding state transition rule is retained. When there are missing stages, incorrect order, or inconsistent readbacks in the actual response phase sequence, the corresponding processing results will be marked as delayed triggering, suspension of environmental parameter contribution, or execution link anomaly. The effective trapping samples are obtained by recording the actual window status, trend prediction status, mosquito trapping execution status, manual response shielding results, actual number of traps, equipment energy consumption, and readback status of the statistical sample window. After removing samples with incomplete data, abnormal capture feedback, or no manual response shielding, the effective trapping samples are obtained. Based on the changes in the number of mosquitoes captured and the changes in equipment energy consumption corresponding to the same state migration records in the effective mosquito trapping samples, the dynamic state boundary is updated in the next mosquito trapping calculation cycle.

8. The mosquito attraction trend prediction method based on dynamic threshold according to claim 1, characterized in that, The processing of the captured feedback status and wind direction includes: The capture feedback status includes the capture channel activity status and the actual number of captures. The capture channel activity status is generated based on at least one of the capture channel occlusion events, occlusion duration, adjacent occlusion intervals, capture bin image count results, and capture bin weight change results. The actual number of mosquitoes captured is primarily determined by the image count of the capture chamber within the statistical sample window. If the image count of the capture chamber is unavailable, the effective occlusion event count of the infrared channel is used. If both the image count of the capture chamber and the effective occlusion event count of the infrared channel are unavailable, the estimated result of the weight change of the capture chamber is used. The number of mosquitoes captured by manual inspection is only used for the correction of the statistical sample window and is not used for the trend prediction status switching of a single mosquito-attracting calculation cycle. The effective wind direction status is generated based on the dominant diffusion direction status, wind speed stability, mosquito attraction node location relationship, and the relative direction from the release source to the collection location. When a neighboring mosquito-attracting node sends a propagation alert, the mosquito-attracting node receiving the propagation alert will determine whether to enter the preparation state only within the validity period of the propagation alert, based on the local real window status, the effective wind direction status, and the effective stable window. It will also confirm the effectiveness of the wind direction propagation response based on whether the actual number of mosquitoes captured in the subsequent statistical sample window increases.

9. A mosquito-attracting trend prediction device based on dynamic thresholds, characterized in that, It includes a data alignment module, a stable boundary module, a response stripping module, an execution determination module, and a feedback correction module; The data alignment module is used to acquire ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity and sampling time within the target area, and to acquire mosquito-attracting execution records, readback status and capture feedback status; it is also used to perform time alignment on the heterogeneous sampling data of each environmental parameter according to the mosquito-attracting calculation cycle, remove missing samples and isolated abnormal samples, and convert them into a mosquito-attracting environmental state set according to the same historical stable range. The stable boundary module is used to divide the historical mosquito-attracting environment state set, historical mosquito-attracting execution record, historical capture feedback state and historical readback state into candidate historical windows, and exclude candidate historical windows with continuous missing data, abnormal sampling clusters, inconsistent readbacks, abnormal capture feedback or unprocessed carbon dioxide release response to obtain an effective stable window; it is also used to generate dynamic state boundary and natural window state record based on the effective stable window and the current mosquito-attracting environment state set. The response stripping module is used to mark and block artificial light source responses, artificial release responses and channel interference in the natural window state record based on mosquito attraction execution records, readback status, wind direction and capture feedback status, so as to obtain the true window state. The execution determination module is used to determine the trend prediction state and the mosquito attraction execution state based on the real window state and the dynamic state boundary. The feedback correction module is used to generate a theoretical response phase sequence and an actual response phase sequence after the mosquito-attracting execution state changes, and to update the dynamic state boundary based on the order matching result of the two and the effective trapping samples.

10. A mosquito attraction trend prediction system based on dynamic thresholds, characterized in that, It includes multiple mosquito-attracting nodes, data storage units, and display terminals; Each mosquito-attracting node includes an environmental data acquisition unit, a mosquito-attracting execution unit, a capture feedback unit, a node processing unit, and a node communication unit; The environmental data acquisition unit is used to acquire ambient temperature, ambient humidity, atmospheric pressure, wind speed, wind direction, carbon dioxide concentration, ambient light intensity, and sampling time within the target area. The mosquito-attracting execution unit is used to activate the mosquito-attracting light source, release odors and carbon dioxide, and generate mosquito-attracting execution records and readback status; The capture feedback unit is used to generate a capture feedback status, which includes the capture channel activity status and the actual number of captures. The node processing unit is used to time-align the heterogeneous sampling data of each environmental parameter according to the mosquito-attracting calculation cycle, remove missing samples and isolated abnormal samples, and convert them into a mosquito-attracting environmental state set according to the same historical stable range; it is also used to exclude candidate historical windows that do not meet the qualification review for stable windows, obtain effective stable windows, and generate dynamic state boundaries and natural window state records based on effective stable windows and the current mosquito-attracting environmental state set; it is also used to mark and mask artificial light source responses, artificial release responses, and channel interference in the natural window state records based on mosquito-attracting execution records, readback status, wind direction, and capture feedback status, to obtain the true window state; it is also used to determine the trend prediction state and mosquito-attracting execution state based on the true window state and the dynamic state boundary; and it is also used to update the dynamic state boundary based on the response stage sequence matching results and effective trapping samples. The node communication unit is used to transmit the real window status, effective wind direction status, location relationship and propagation prompts between adjacent mosquito-attracting nodes; Mosquito-attracting nodes that receive propagation cues only enter the ready state when the local real window state, wind direction effective state, and effective stable window support are present. The data storage unit is used to store mosquito-attracting environment state set, effective stable window, dynamic state boundary, natural window state record, real window state, artificial response shielding result, response stage sequence matching result, effective trapping sample and propagation prompt record; The display terminal is used to display the mosquito-attracting execution status, trend prediction status, actual number of mosquitoes captured, device energy consumption, and abnormal prompts. It is not used to directly generate the actual window status or directly switch the mosquito-attracting execution status.

Images