A pest control system and method based on sound and light

Abstract

This invention discloses an acoustic and optical pest control system and method. The invention relates to the field of pest identification technology. By setting up a pest identification image generation unit, based on the growth characteristics of crops in image data and the collection time, the invention determines the corresponding local or cloud collection nodes, obtains node numbers, and generates corresponding pest identification images. According to the node number, whether it is a local or cloud collection node, the corresponding model is selected and activated for pest identification. For local collection nodes, the corresponding pest-specific identification model is activated to identify the corresponding pest type; for cloud collection nodes, the cloud pest control server performs pest type identification. This intelligent selection and operation of the corresponding identification model avoids excessive and unnecessary pest type identification and accelerates the identification rate of the corresponding pest type.

Description

Technical Field

[0001] This invention relates to the field of pest identification technology, specifically to an acoustic and optical pest control system and method. Background Technology

[0002] With the development of smart agriculture, traditional pest control methods (such as chemical pesticide spraying and manual inspection) are gradually being replaced by intelligent solutions. Currently, the field of agricultural pest control mainly relies on intelligent identification and control technologies. Traditional methods provide pest identification services to various agricultural areas through cloud servers. The specific process is as follows: monitoring equipment in the agricultural area collects image data and transmits it to the cloud. A pre-trained general pest identification model is used to analyze the images, and finally the identification results are returned to the local area to trigger control measures. However, this method requires the agricultural area to upload a large amount of image data to the cloud in real time. Due to limitations in network bandwidth and transmission stability, the feedback of identification results is delayed, making it difficult to meet the timeliness requirements of pest control. Moreover, the general pest identification model needs to cover multiple crops and pest types, resulting in high model complexity and high computational resource consumption.

[0003] To address the aforementioned issues, existing technologies propose to offload pest identification tasks locally by deploying specialized identification models for specific crops or pest types in corresponding areas. However, the same agricultural area may be planted with multiple crops, and the peak pest seasons for different crops vary periodically. To cover all possible pest types, multiple specialized models need to be run for a long time, resulting in redundant consumption of computing resources and power energy.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] The purpose of this invention is to provide an acoustic and optical pest control system and method to solve the problems mentioned in the background art.

[0006] This invention provides an acoustic and visual pest control system, comprising:

[0007] The image atlas generation unit is used to identify the growth characteristics of crops in the image data after receiving real-time image data of the planting area, and to determine the growth stage of the crop corresponding to the image data.

[0008] The image atlas generation unit is also used to determine the node number of the image data by determining the corresponding growth stage of the crop and the acquisition time of the image data.

[0009] The identification atlas generation unit is also used to generate a pest identification atlas for the node number when the sum of the acquisition times of all image data of the planting area stored in it is consistent with the acquisition time of the local acquisition node or cloud acquisition node corresponding to the node number of any image data in all the image data.

[0010] The local branch identification unit is used to identify pests in the pest identification map set according to the preset identification steps after receiving the pest identification map set with the node number to obtain the pest identification result data of the node number.

[0011] The pest identification result data includes an identification signal, which is selected from the numbers 1 and 0. When the identification signal is the number 1, the pest identification result data also includes several pest types.

[0012] The acoustic and light control response module is used to obtain the corresponding acoustic and light control strategy based on several pest types contained in the pest identification result data after receiving pest identification result data with an identification signal of 1, and to adjust the acoustic and light control of the planting area according to the acoustic and light control strategy.

[0013] The cloud-based pest control service is used to provide identification services for pest control in planting areas. The cloud-based pest control service includes a cloud-based pest identification unit and a cloud-based pest analysis unit.

[0014] Furthermore, the various growth stages of crops are defined by agricultural technicians based on the crop's growth cycle.

[0015] Furthermore, the cloud pest identification unit pre-stores crop pest and disease identification models.

[0016] Furthermore, the cloud pest analysis unit stores several pest identification record data, which includes identification signals, planting areas, the acquisition time of the starting image, the acquisition time of the ending image, several pest types and their type feature sets.

[0017] Furthermore, the cloud pest analysis unit analyzes all the stored pest identification record data to obtain pest identification branch data of crops in the planting area at several growth stages. The pest identification branch data includes several independent pest types, merged pest types and their associated type sets.

[0018] Furthermore, agricultural technicians, for crops in the planting area, at any growth stage, based on the pest identification and decomposition data of the crops at the growth stage, divide the interval of the growth stage into several local collection nodes and cloud collection nodes, and pre-set a collection duration for each local collection node and cloud collection node, wherein one local collection node corresponds to one pest-specific identification model.

[0019] After the division is completed, the local collection nodes and cloud collection nodes divided from this growth stage are numbered sequentially starting from 1.

[0020] A method for controlling pests through sound and light includes the following steps:

[0021] Step 1: After receiving real-time image data of the planting area, the image generation unit identifies the growth characteristics of the crops in the image data to determine the growth stage of the crops corresponding to the image data. Based on the determined growth stage of the crops and the acquisition time of the image data, the node number of the image data is determined.

[0022] Step 2: When the sum of the acquisition times of all image data of the planting area stored in the identification image generation unit is consistent with the acquisition time of the local acquisition node or cloud acquisition node corresponding to the node number of any image data in all the image data, the identification image generation unit generates the pest identification image set of the node number based on all the image data of the planting area stored in the unit, and transmits it to the local branch identification unit.

[0023] Step 3: After receiving the pest identification map set of the node number, the local branch identification unit performs pest identification on the pest identification map set according to the preset identification steps to obtain the pest identification result data of the node number. The pest identification result data contains an identification signal, which is selected from the numbers 1 and 0. When the identification signal is the number 1, the pest identification result data also contains several pest types.

[0024] When the identification signal is 1, the pest identification result data is transmitted to the sound and light control response module;

[0025] Step 4: After receiving the pest identification result data with an identification signal of 1, the sound and light control response module obtains the corresponding sound and light control strategy based on the several pest types contained in the pest identification result data, and adjusts the sound and light control measures for the planting area according to the sound and light control strategy.

[0026] Compared with existing technologies, it has the following advantages:

[0027] This invention sets up an image atlas generation unit to determine the corresponding local or cloud collection nodes based on the crop growth characteristics and collection time in the image data, thereby obtaining node numbers and generating corresponding pest identification atlases. Based on the node number, it selects and activates the corresponding model for pest identification, either as a local or cloud collection node. For local collection nodes, it further activates the pest-specific identification model corresponding to the local collection node to identify the corresponding pest type. For cloud collection nodes, it selects to transmit the data to the cloud pest control server to use the crop disease and pest identification model for pest type identification. This intelligently selects the corresponding identification model for operation and performs specialized identification, avoiding unnecessary pest type identification and resource waste caused by running multiple identification models locally, while also accelerating the identification rate of the corresponding pest type.

[0028] This invention analyzes pest identification records from different planting areas using a cloud-based pest analysis unit to determine pest identification branches at several growth stages of crops in the corresponding planting areas. A cloud-based pest decomposition unit extracts pest identification data at several growth stages from the crop pest identification model across all planting areas. Agricultural technicians then divide the corresponding crop growth stages into local and cloud collection nodes based on the identified pest type. This approach allows pest types that occur less frequently and require significant resources to deploy identification models locally to be identified in the cloud. This not only optimizes the use of pest type identification models and simplifies their complexity but also meets the timeliness requirements for pest control. Attached Figure Description

[0029] Figure 1 This is a system block diagram of the present invention;

[0030] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figure 1 , Figure 2 This application provides a pest control system and method using sound and light, including a pest monitoring terminal, a pest branch identification module, a sound and light control response module, and a cloud pest control server.

[0033] The pest monitoring terminal is used to monitor all planting areas within the target site. The pest monitoring terminal includes several pest monitoring modules, and one pest monitoring module corresponds to one planting area within the target site.

[0034] In this application, the planting area is a greenhouse, which has a controllable environment throughout the year. That is, the temperature, humidity, light and other control factors inside the greenhouse are controlled by humans, so the pests inside are not limited by the season.

[0035] In this application, the delineation of the planting area is carried out by agricultural technicians based on the types of crops, growth and distribution characteristics, historical occurrence patterns of pests and diseases, and ecological environment conditions of the target site, combined with GIS spatial analysis technology.

[0036] In this application, only one type of crop is planted in a planting area at any given time. In this application, the type of crop planted in any planting area within the target site is selected by agricultural technicians.

[0037] In this application, only one type of crop is planted in any planting area within the target site;

[0038] The pest monitoring module collects image data of the corresponding planting area in real time and transmits it to the pest branch recognition module. The pest branch recognition module includes a recognition image generation unit and a local branch recognition unit.

[0039] After receiving the image data, the pest branch identification module transmits it to the identification atlas generation unit. After receiving the image data, the identification atlas generation unit identifies the growth characteristics of the crop in the image data and determines that the image data corresponds to the growth stage of the crop.

[0040] Specifically: extract several growth features of crops from the image data, calculate the similarity between the several growth features and several standard growth features of the crops at each growth stage that are pre-stored, and determine the standard growth feature with the highest similarity as the growth stage of the crops corresponding to the image data.

[0041] In this application, growth characteristics include, but are not limited to, the morphological characteristics of crops such as roots, stems, leaves, and flowers;

[0042] In this application, the growth stages of any one of the crops are divided by agricultural technicians according to the growth cycle of the crops;

[0043] The node number of the image data is determined by the corresponding growth stage of the crop and the acquisition time of the image data. The determination steps are as follows:

[0044] The pest branch identification module obtains the collection duration and numerical ID of several local collection nodes and several cloud collection nodes of the crop during the growth stage.

[0045] Then, all image data corresponding to the growth stage currently stored in the pest branch identification module are obtained, and all the obtained image data are sorted from left to right according to the order of the acquisition time of the image data. If the sum Z1 of the image data before the image data after sorting satisfies Y1+Y2+...+Yy1≤Z1≤Y1+Y2+...+Yy1+1, then the node number of the image data is determined to be y1+1, where Y1, Y2, ..., Yy1 are the acquisition duration of the local acquisition node or cloud acquisition node corresponding to the numerical numbers 1, 2, ..., y1, Yy1+1 respectively.

[0046] The image data and its node numbers are temporarily stored;

[0047] For a planting area, when the total collection time of all image data of the planting area stored in the identification atlas generation unit reaches the collection time of the local collection node or cloud collection node corresponding to its node number, a pest identification atlas of the node number is generated based on all image data of the planting area stored in the unit.

[0048] It should be noted that all the image data contained in the generated pest identification atlas have the same node number;

[0049] After receiving the pest identification map set with the node number transmitted, the local branch identification unit performs pest identification on the pest identification map set according to the preset identification steps. The identification steps are as follows:

[0050] S11: If the node number corresponds to a local collection node, then obtain the pest-specific identification model corresponding to the local collection node corresponding to the node number, use the pest identification atlas as the input of the pest-specific identification model, and use the pest-specific identification model to identify the pest identification result data of the node number, wherein the pest identification result data contains the identification signal.

[0051] In this application, the identification signal is selected from the numbers 1 and 0. When the identification signal is the number 1, it indicates that a pest has been identified at the local acquisition node. At this time, the pest identification result data also includes the planting area corresponding to the crop, the growth stage, the local acquisition node, the acquisition time of the starting image, the acquisition time of the ending image, several pest types and their type feature map sets. Each type feature map set contains image data that identifies the corresponding pest type, and all the image data contained in a type feature map set are ordered and arranged from left to right according to the order of acquisition. The starting image and the ending image are the image data located at the leftmost and rightmost ends of the pest identification map set, respectively.

[0052] If the identification signal is 0, it means that no pests were identified at the local acquisition node.

[0053] If the identification signal in the pest identification result data is 1, then the pest identification result data is transmitted to the sound and light control response module.

[0054] S12: If the node number corresponds to a cloud collection node, then the pest identification atlas is transmitted to the cloud pest control service terminal.

[0055] The cloud pest control service terminal is used to provide identification services for pest control in several planting areas within a target site. The cloud pest control service terminal includes a cloud pest identification unit, a cloud pest analysis unit, and a cloud pest disassembly unit.

[0056] After receiving the transmitted pest identification image set, the cloud pest control server transmits it to the cloud pest identification unit.

[0057] The cloud pest identification unit has a pre-stored crop pest identification model, which is used to identify various types of crop pests from image data.

[0058] After receiving the transmitted pest identification atlas, the cloud pest identification unit inputs it into the crop pest identification model to obtain the pest identification result data of the node number.

[0059] If the identification signal in the pest identification result data is 1, then the pest identification result data is transmitted to the sound and light control response module.

[0060] The sound and light control response module is used to carry out sound and light control on the corresponding planting area according to the pest types identified in several planting areas within the target area. The sound and light control response model pre-stores several sound and light control strategies for crops based on several pest types. Each sound and light control strategy includes light adjustment parameters and sound wave adjustment parameters for controlling the corresponding pest type.

[0061] After receiving the transmitted pest identification result data, the sound and light control response module obtains the corresponding sound and light control strategy according to the pest type contained in the pest identification result data, and adjusts the sound and light control for the corresponding planting area according to the sound and light control strategy.

[0062] The cloud pest analysis unit stores pest identification record data for all planting areas within the target site at several growth stages. The pest identification record data includes identification signals, planting areas, the acquisition time of the starting image, the acquisition time of the ending image, several pest types and their type feature sets.

[0063] The identification signal in all pest identification record data stored in the cloud pest analysis unit is 1;

[0064] The cloud-based pest analysis unit analyzes all the stored pest identification record data. The analysis steps are as follows:

[0065] S11: First, randomly select a planting area within the target site as the analysis area, select the crops planted in the analysis area as the analysis crops, and label all growth stages according to the growth cycle of the analysis crops in the analysis area as A1, A2, ..., Aa in chronological order of crop growth time, where a≥1;

[0066] S12: Obtain all pest identification records of the crop in the analysis area A1 stored in the cloud pest analysis unit and traverse them to obtain all pest types contained therein, which are marked as B1, B2, ..., Bb, b≥1 respectively;

[0067] S13: Extract all pest identification records containing pest type B1 from all the obtained pest identification record data, and label them as C1, C2, ..., Cc, where c≥1;

[0068] S14: Calculate the identification adaptation evaluation index F1 of pest type B1 in the pest identification record data C1 according to the preset calculation rules. The calculation rules are as follows:

[0069] S141: Extract the type feature map set of pest type B1 from the pest identification record data C1, and extract the acquisition time of the image data located at the leftmost and rightmost ends of the type feature map set respectively, and mark them as D1 and D2 respectively;

[0070] The identification span weight E1 of pest type B1 in the type feature map set is calculated using the formula E1 = (D2 - D1) / D3. It should be noted that the identification span weight is defined by humans and is used to measure the coverage integrity of the image data in the type feature map set that can identify pest type B1.

[0071] S142: Obtain the acquisition time of the start image and the acquisition time of the end image in the pest identification record data C1, and mark them as D4 and D5 respectively;

[0072] S143: Use the formula E2=(D1-D4) -1 The formula ×ɑ1+(D2-D1) / (D5-D4)×ɑ2 calculates the identification weight E2 of pest type B1 in the type feature map set. In the formula, D2-D1 is the initial identification time of pest type B1 in pest identification record data C1, D5-D4 is the final identification time of pest type B1 in pest identification record data C1, and ɑ1 and ɑ2 are the preset first and second feature dimension factors, which are used to adjust features of different calculation dimensions to the same calculation dimension for calculation. It should be noted that the identification weight E2 is defined by humans and is used to measure the degree of feature of the time span of the image data in the type feature map set that identifies pest type B1.

[0073] S144: Use the formula F1=[P1-(E1×β1+E2×β2) 2 [E3] Calculate the identification adaptation evaluation index F1 for pest type B1 in the pest identification record data C1. It should be noted that the identification adaptation evaluation index is artificially defined and is used to comprehensively evaluate the deployment adaptability of the identified pest type F1 based on three dimensions: identification span weight, identification proportion weight, and the data size of the input pest identification map set. In the formula, P1 is the preset standard compensation constant, and E3 is the data size of the input pest identification map set when the crop pest identification model identifies the pest identification record data C1.

[0074] S15: Calculate and obtain the identification and adaptation evaluation indexes F2, F3, ..., Fc of pest type B1 in pest identification record data C1, C2, ..., Cc in sequence according to S13 to S14, and simultaneously obtain the initial identification feature time and the final identification feature time of pest type B1 in pest identification record data C2, C3, ..., Cc;

[0075] The discrete point filtering algorithm is used to process the obtained identification and adaptation evaluation indices F1, F2, ..., Fc, and calculate the average value of all remaining identification and adaptation evaluation indices after data processing. The average value is then calibrated as the identification and adaptation index G1 of pest type B1 of the crop under growth stage A1 in the analysis area. Similarly, the branch initial recognition time and branch final recognition time of pest type B1 of the crop under growth stage A1 in the analysis area are obtained.

[0076] The initial identification time of pest type B1 in the crop growth stage A1 within the analysis area is obtained by using a discrete point filtering algorithm to process the initial identification feature time of pest type B1 in the obtained pest identification record data C1, C2, ..., Cc, and then calculating the average value of all remaining initial identification feature times after data processing for calibration.

[0077] The final identification time of pest type B1 under crop growth stage A1 in the analysis area is obtained by using a discrete point filtering algorithm to process the final identification feature time of pest type B1 in the obtained pest identification record data C1, C2, ..., Cc, and then calculating the average value of all remaining final identification feature times after data processing for calibration.

[0078] In this application, the discrete point filtering algorithm can be one of the Z-score filtering algorithm, IQR filtering algorithm, and density filtering algorithm;

[0079] S16: Extract all pest identification record data containing pest types B2, B3, ..., Bb from all the obtained pest identification record data in sequence, and calculate the identification adaptation index of pest types B2, B3, ..., Bb of the crop under growth stage A1 in the analysis area in sequence according to S14 to S15.

[0080] S17: The identification and adaptation indices of the pest types B1, B2, ..., Bb are compared with P2 in turn. All pest types corresponding to the identification and adaptation indices with values ​​greater than or equal to P2 are obtained and labeled as H1, H2, ..., Hh, respectively, where 1≤h≤b, and P2 is a preset pest identification branch comparison quantity.

[0081] S18: Sequentially obtain the branch initial recognition time and branch final recognition time of pest types H1, H2, ..., Hh under growth stage A1 of the analyzed crop within the analysis area and traverse them. If there are several pest types that simultaneously satisfy the difference between the branch initial recognition time and the branch initial recognition time of pest type H1, and the difference between the branch final recognition time and the branch final recognition time of pest type H1, both of which are less than or equal to P3, then pest type H1 is relabeled as a merged pest type, and these pest types are used as the associated type set of pest type H1. Otherwise, pest type H1 is relabeled as an independent pest type. P3 is the preset collaborative recognition difference judgment standard quantity.

[0082] Similarly, we can determine whether there are several pest types that simultaneously satisfy the following conditions: the difference between the branch initial time and the branch initial time of pest types H2, H3, ..., Hh, and the branch final time of pest types H2, H3, ..., Hh, which are less than or equal to P3. This yields several merged pest types and their associated type sets, as well as several independent pest types.

[0083] Based on the obtained merged pest types and their associated type sets, several independent pest types generate pest identification branch data for the crop in growth stage A1 within the analysis area.

[0084] S19: Generate pest identification branch data for crops in the analysis area at growth stages A2, A3, ..., Aa in sequence according to S11 to S18;

[0085] S110: Select all planting areas within the target site as the analysis area, and calculate and obtain the pest identification branch data of the corresponding crops in several growth stages in all planting areas in accordance with S11 to S19.

[0086] The pest identification data of the corresponding crops at several growth stages in all the planting areas are transmitted to the cloud pest dismantling unit.

[0087] After receiving the transmitted pest identification branch data of corresponding crops at several growth stages in all planting areas, the cloud pest decomposition unit decomposes several pest identification decomposition data of corresponding crops at several growth stages in all planting areas from the crop pest identification model according to the preset decomposition rules. The decomposition rules are as follows:

[0088] S21: Select any planting area within the target site as a branch area, and divide all growth stages into which the crops within the branch area are located according to their growth cycle, in chronological order.

[0089]

[0090] S22: Obtain all pest identification branch data at growth stage I1 from all pest identification branch data of crops corresponding to the branch region at several growth stages, and extract all merged pest types and their associated type sets from the obtained pest identification branch data.

[0091] Based on any extracted merged pest type and its associated type set, a sub-model that can simultaneously identify all pest types within the merged pest type and its associated type set is extracted from the crop pest identification model, and the sub-model is labeled as the pest-specific identification model for the merged pest type.

[0092] S23: Extract all independent pest types contained in the pest identification branch data, and according to any one of the extracted independent pest types, split out a sub-model that can identify the independent pest type from the crop pest identification model, and label the sub-model as the pest-specific identification model for the independent pest type.

[0093] S24: Generate pest identification breakdown data for crops in the growth stage I1 of the branch area based on all pest-specific identification models obtained from S22 to S23 and the several pest types they can identify.

[0094] S25: Generate pest identification and breakdown data for crops in the branch areas at growth stages I2, I3, ..., Ii in sequence according to S21 to S24;

[0095] S26: Select all planting areas within the target site as branch areas in sequence, and generate pest identification and breakdown data for crops in all planting areas within the target site at several growth stages according to S21 to S25.

[0096] The cloud pest decomposition unit transmits the pest identification and decomposition data of crops in all planting areas within the target site at several growth stages to the local branch identification unit of the pest branch identification module.

[0097] After receiving the pest identification and decomposition data of crops in all planting areas within the target location based on several growth stages, the local branch identification unit first divides the crops in all planting areas into several local collection nodes and cloud collection nodes corresponding to all growth stages, and sets the collection duration and assigns a digital number accordingly.

[0098] Specifically, agricultural technicians, for any crop in any planting area, at any growth stage, based on the pest identification breakdown data of the crop at the growth stage (which includes several pest-specific identification models and their respective corresponding pest types), divide the interval of the growth stage into several local collection nodes and cloud collection nodes, and pre-set a collection duration for each local collection node and cloud collection node. In this process, one local collection node corresponds to one pest-specific identification model.

[0099] After the division is completed, the local and cloud data collection nodes divided from this growth stage are numerically numbered, starting from 1 and proceeding sequentially. The larger the number, the farther the corresponding data collection node is from the starting point of this growth stage;

[0100] The local branch identification unit divides all crops in all planting areas into several local collection nodes corresponding to all growth stages, and stores their corresponding pest identification models, collection durations, and numerical numbers, as well as the collection durations and numerical numbers of cloud collection nodes. Simultaneously, it transmits all the local collection nodes, corresponding pest identification models, collection durations, and numerical numbers, as well as the collection durations and numerical numbers of cloud collection nodes, to the local branch identification unit for storage.

[0101] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0102] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A pest control system based on sound and light, characterized in that, include: The image atlas generation unit is used to identify the growth characteristics of crops in the image data after receiving real-time image data of the planting area, and to determine the growth stage of the crop corresponding to the image data. The image atlas generation unit is also used to determine the node number of the image data by determining the corresponding growth stage of the crop and the acquisition time of the image data. The identification atlas generation unit is also used to generate a pest identification atlas for the node number when the sum of the acquisition times of all image data of the planting area stored in it is consistent with the acquisition time of the local acquisition node or cloud acquisition node corresponding to the node number of any image data in all the image data. The local branch identification unit is used to identify pests in the pest identification map set according to the preset identification steps after receiving the pest identification map set with the node number to obtain the pest identification result data of the node number. The pest identification result data includes an identification signal, which is selected from the numbers 1 and 0. When the identification signal is the number 1, the pest identification result data also includes several pest types. The acoustic and light control response module is used to obtain the corresponding acoustic and light control strategy based on several pest types contained in the pest identification result data after receiving pest identification result data with an identification signal of 1, and to adjust the acoustic and light control of the planting area according to the acoustic and light control strategy. The cloud-based pest control service is used to provide identification services for pest control in planting areas. The cloud-based pest control service includes a cloud-based pest identification unit, a cloud-based pest analysis unit, and a cloud-based pest decomposition unit. The cloud-based pest analysis unit analyzes all stored pest identification record data to obtain pest identification branch data of crops in the planting area at several growth stages. The pest identification branch data includes several independent pest types, merged pest types, and their associated type sets. The cloud-based pest identification unit receives pest identification branch data of crops at several growth stages in the planting area and extracts several pest identification decomposition data of crops at several growth stages in the planting area from the crop pest identification model. The decomposition steps are as follows: S21: The planting area is taken as a branch area, and all growth stages divided according to the growth cycle of crops in the branch area are sequentially marked as I1, I2, ..., Ii, i≥1 according to the time sequence of crop growth. S22: Obtain all pest identification branch data at growth stage I1 from all pest identification branch data of crops corresponding to the branch region at several growth stages, and extract all merged pest types and their associated type sets from the obtained pest identification branch data. Based on any extracted merged pest type and its associated type set, a sub-model that can simultaneously identify all pest types within the merged pest type and its associated type set is extracted from the crop pest identification model, and the sub-model is labeled as the pest-specific identification model for the merged pest type. S23: Extract all independent pest types contained in the pest identification branch data, and according to any one of the extracted independent pest types, split out a sub-model that can identify the independent pest type from the crop pest identification model, and label the sub-model as the pest-specific identification model for the independent pest type. S24: Generate pest identification breakdown data for crops in the growth stage I1 of the branch area based on all pest-specific identification models obtained from S22 to S23 and the several pest types they can identify. S25: Generate pest identification and breakdown data for crops in the branch areas at growth stages I2, I3, ..., Ii in sequence according to S21 to S24.

2. The insect pest control system based on claim 1, characterized in that, The different growth stages of crops are defined by agricultural technicians based on the growth cycle of the crops.

3. The insect pest control system based on claim 1, characterized in that, The cloud-based pest identification unit pre-stores crop pest and disease identification models.

4. The insect pest control system based on claim 1, characterized in that, It also includes a cloud-based pest analysis unit that stores several pest identification record data, which includes identification signals, planting areas, the acquisition time of the starting image, the acquisition time of the ending image, several pest types and their type feature sets.

5. The insect pest control system based on claim 1, characterized in that, Agricultural technicians, for crops in the planting area, at any growth stage, divide the interval of the growth stage into several local collection nodes and cloud collection nodes based on the pest identification and decomposition data of the crops based on the growth stage, and pre-set a collection duration for each local collection node and cloud collection node. Among them, one local collection node corresponds to one pest-specific identification model. After the division is completed, the local collection nodes and cloud collection nodes divided from this growth stage are numbered sequentially starting from 1.

6. The insect pest control system based on claim 5, characterized in that, The steps for identifying pests in the pest identification atlas to obtain the pest identification result data of the node number are as follows: S11: If the node number corresponds to a local collection node, then obtain the pest-specific identification model corresponding to the local collection node corresponding to the node number, use the pest identification atlas as the input of the pest-specific identification model, and have the pest-specific identification model identify the pest identification result data of the node number. S12: If the node number corresponds to a cloud collection node, the pest identification map is transmitted to the cloud pest identification unit. After receiving the transmitted pest identification map, the cloud pest identification unit inputs it into the crop pest identification model to obtain the pest identification result data of the node number.

7. The insect pest control system based on claim 5, characterized in that, The steps for determining the node number of the image data are as follows: The pest branch identification module obtains the collection duration and numerical ID of several local collection nodes and several cloud collection nodes of the crop during the growth stage. Then, all image data corresponding to the growth stage currently stored in the pest branch identification module are obtained, and all the obtained image data are sorted from left to right according to the order of the acquisition time of the image data. If the sum Z1 of the image data before the image data after sorting satisfies Y1+Y2+...+Yy1≤Z1≤Y1+Y2+...+Yy1+1, then the node number of the image data is determined to be y1+1, where Y1, Y2, ..., Yy1 are the acquisition durations of the local acquisition node or cloud acquisition node corresponding to the numerical numbers 1, 2, ..., y1, Yy1+1 respectively.

8. A method for controlling pests through sound and light, characterized in that, Includes the following steps: Step 1: After receiving real-time image data of the planting area, the image generation unit identifies the growth characteristics of the crops in the image data to determine the growth stage of the crops corresponding to the image data. Based on the determined growth stage of the crops and the acquisition time of the image data, the node number of the image data is determined. Step 2: When the sum of the acquisition times of all image data of the planting area stored in the identification image generation unit is consistent with the acquisition time of the local acquisition node or cloud acquisition node corresponding to the node number of any image data in all the image data, the identification image generation unit generates the pest identification image set of the node number based on all the image data of the planting area stored in the unit, and transmits it to the local branch identification unit. Step 3: After receiving the pest identification map of the node number, the local branch identification unit performs pest identification on the pest identification map according to the preset identification steps to obtain the pest identification result data of the node number. The pest identification result data contains an identification signal, which is selected from the numbers 1 and 0. When the identification signal is the number 1, the pest identification result data also contains several pest types. When the identification signal is 1, the pest identification result data is transmitted to the sound and light control response module. Step 4: After receiving the pest identification result data with an identification signal of 1, the sound and light control response module obtains the corresponding sound and light control strategy based on the several pest types contained in the pest identification result data, and adjusts the sound and light control measures for the planting area according to the sound and light control strategy. The cloud-based pest control service is used to provide identification services for pest control in planting areas. The cloud-based pest control service includes a cloud-based pest identification unit, a cloud-based pest analysis unit, and a cloud-based pest decomposition unit. The cloud-based pest analysis unit analyzes all stored pest identification record data to obtain pest identification branch data of crops in the planting area at several growth stages. The pest identification branch data includes several independent pest types, merged pest types, and their associated type sets. The cloud-based pest identification unit receives pest identification branch data of crops at several growth stages in the planting area and extracts several pest identification decomposition data of crops at several growth stages in the planting area from the crop pest identification model. The decomposition steps are as follows: S21: The planting area is taken as a branch area, and all growth stages divided according to the growth cycle of crops in the branch area are sequentially marked as I1, I2, ..., Ii, i≥1 according to the time sequence of crop growth. S22: Obtain all pest identification branch data at growth stage I1 from all pest identification branch data of crops corresponding to the branch region at several growth stages, and extract all merged pest types and their associated type sets from the obtained pest identification branch data. Based on any extracted merged pest type and its associated type set, a sub-model that can simultaneously identify all pest types within the merged pest type and its associated type set is extracted from the crop pest identification model, and the sub-model is labeled as the pest-specific identification model for the merged pest type. S23: Extract all independent pest types contained in the pest identification branch data, and according to any one of the extracted independent pest types, split out a sub-model that can identify the independent pest type from the crop pest identification model, and label the sub-model as the pest-specific identification model for the independent pest type. S24: Generate pest identification breakdown data for crops in the growth stage I1 of the branch area based on all pest-specific identification models obtained from S22 to S23 and the several pest types they can identify. S25: Generate pest identification and breakdown data for crops in the branch areas at growth stages I2, I3, ..., Ii in sequence according to S21 to S24.

Images