Honey production prediction method and system

Abstract

The invention relates to the technical field of honey production prediction, in particular to a honey production prediction method and system. Determining a change period, and identifying the target breeding area at the beginning of the change period to obtain flower image data and flower distribution data; constructing a residual flowering phase prediction model to identify and adjust the flower image data to obtain adjusted distribution data; obtaining a clustering center according to the adjusted distribution data and a clustering algorithm; the nectar collection amount of bees and the activity distance of the bees are obtained through recognition; obtaining a plurality of sub-regions according to the clustering center and the bee activity distance, and collecting a region flower data set and a region environment data set; obtaining a first region coefficient according to the region flower data set and the nectar collection amount; obtaining a second region coefficient according to the region environment coefficient set; and obtaining a region prediction index according to the first region coefficient and the second region coefficient. According to the method, the honey production efficiency of different beehive positions is compared, and the optimal beehive position is predicted.

Description

Technical Field

[0001] The invention relates to the technical field of honey production prediction, and in particular to a honey production prediction method and system. Background Art

[0002] The production of honey mainly relies on bees to collect nectar naturally, and the honey obtained by bees is reprocessed to obtain honey products. In order to increase the output and stability of honey production, an industry of breeding bees through beehives has emerged. Beehives can provide a safe and stable environment for bees, effectively protecting them from bad weather, natural enemies and other external threats; they can also realize long-distance transportation of bees in order to track the flowering period and ensure the production of honey.

[0003] In the process of bee breeding and honey production through beehives, the location selection of the beehive is an important consideration; by adjusting the location of the beehive, the bees can be close to abundant nectar sources, avoid interference from harsh natural environments, and increase honey production and quality; therefore, it is necessary to study the location setting of the beehive.

[0004] In the existing related technologies, natural environmental factors are taken into consideration to predict the honey production from a macro level to obtain the honey production; there are few predictions on the location selection and change of beehives in honey production, and the location of beehives cannot be well guided.

[0005] Therefore, a prediction method and system for honey production are proposed. Summary of the invention

[0006] The purpose of the present invention is to provide a prediction method and system for honey production; firstly, a change cycle is determined, and at the beginning of the change cycle, a target breeding area is identified to obtain flower image data and flower distribution data; a remaining flowering period prediction model is constructed to identify and adjust the flower image data to obtain adjusted distribution data; a cluster center is obtained according to the adjusted distribution data and a clustering algorithm; the nectar collection amount of bees and the bee activity distance are identified; a plurality of sub-areas are obtained according to the cluster center and the bee activity distance, and a regional flower data set and a regional environmental data set are collected; a first regional coefficient is obtained according to the regional flower data set and the nectar collection amount; a second regional coefficient is obtained according to the regional environmental coefficient set; a regional prediction index is obtained according to the first regional coefficient and the second regional coefficient; the honey production efficiency of different beehive locations is predicted and compared by the regional prediction index to determine the optimal beehive location.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for predicting honey production, comprising:

[0009] S10. Determine the change cycle of the hive position; at the beginning of the change cycle, identify the target breeding area to obtain flower image data and flower distribution data;

[0010] S20. Constructing a remaining flowering period prediction model to identify the flower image data to obtain the remaining flowering period of the flowers; deleting the flower distribution data whose remaining flowering period is less than the change period to obtain the adjusted distribution data;

[0011] S30. Obtain flower coordinate data according to the adjusted distribution data, perform clustering algorithm identification on the flower coordinate data, and obtain multiple cluster centers;

[0012] S40. Identify the bees raised in the beehive to obtain the nectar collection amount and the bee activity distance; take the cluster center as the center of the circle and the bee activity distance as the radius to obtain multiple sub-areas;

[0013] S50. Collect and predict data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data;

[0014] S60. Calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set;

[0015] S70. Calculate a regional prediction index based on the first regional coefficient and the second regional coefficient of the sub-region; and determine the location of the beehive based on the regional prediction index.

[0016] The remaining flowering period prediction model includes a flower image input layer, a flower image recognition layer, a flower feature recognition layer, a flowering period prediction layer, a flowering period correction layer and a flowering period output layer;

[0017] The flower image input layer is used to input the obtained flower image data into the model, and pre-process the flower image data to obtain a pre-processed flower image;

[0018] The flower image recognition layer recognizes and segments the pre-processed flower image to obtain a flower sub-image; determines the flower species in the flower sub-image, and organizes the flower sub-images of the same species into a flower recognition image set;

[0019] The flower feature recognition layer is used to recognize the color features, texture features and shape features of the flowers in the flower recognition image set;

[0020] The flowering period prediction layer obtains standard flowering period data according to the flower species, and identifies the remaining flowering period of the flowers in the flower recognition image set through the standard flowering period data, color feature data, texture feature data and shape feature data;

[0021] The flowering period correction layer obtains the remaining flowering period of all flowers in the flower recognition image set, and corrects the remaining flowering period of flowers in the flower sub-image according to the remaining flowering period of all flowers;

[0022] The flowering period output layer is used to output the remaining flowering period of the flower.

[0023] The flower quality coefficient is determined by the flower's attraction index to bees and the flower's nectar quality index;

[0024] Identify all flower species in the target breeding area, collect the same number of flowers of various types and set up a closed first experimental environment; select a first bee sample and place it in the first experimental environment, identify the frequency of the first bee sample visiting different types of flowers through a positioning system, and use the frequency as the attraction index of the flower;

[0025] Identify all flower species in the target breeding area, collect a single type of flowers to set up a second closed experimental environment; select a second bee sample and place it in the experimental environment, and perform chemical analysis on the honey produced by the second bee sample to determine the substance ratio; and determine the nectar quality index by having an expert group study the substance ratio.

[0026] The calculation formula of the first regional coefficient is:

[0027]

[0028] Among them, Region 1 represents the first area coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

[0029] The second regional coefficient is calculated as:

[0030]

[0031] Among them, Region 2 represents the second area coefficient; Ter represents the terrain flatness of the sub-area; α 1represents the first impact weight; TH represents the first threshold; D() represents the distance calculation formula, Temp represents the predicted temperature data of the sub-area, ST represents the suitable temperature data of the bees; α 2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-area; SW represents the suitable wind speed data of the bees; exp represents an exponential function with a natural constant as the base.

[0032] The calculation formula of the regional prediction index is:

[0033]

[0034] Among them, Fore represents the regional forecast index; Region 1 Indicates the first regional coefficient; Region 2 Represents the second region coefficient.

[0035] A prediction system for honey production, comprising:

[0036] A breeding area identification module determines the change period of the beehive position; at the beginning of the change period, the target breeding area is identified to obtain flower image data and flower distribution data;

[0037] A flowering period recognition module is used to construct a remaining flowering period prediction model to recognize the flower image data to obtain the remaining flowering period of the flowers; the flower distribution data of the flowers whose remaining flowering period is less than the change period are deleted to obtain the adjusted distribution data;

[0038] A flower clustering recognition module obtains flower coordinate data according to the adjusted distribution data, and performs clustering algorithm recognition on the flower coordinate data to obtain multiple cluster centers;

[0039] The breeding area division module identifies the bees bred in the beehive to obtain the nectar collection amount and the bee activity distance; taking the cluster center as the center of the circle and the activity distance as the radius, a plurality of sub-areas are obtained;

[0040] The regional data identification module collects and predicts data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data;

[0041] A regional coefficient calculation module is used to calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; and to calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set;

[0042] The beehive location determination module calculates a regional prediction index according to the first regional coefficient and the second regional coefficient of the sub-region; and determines the beehive location according to the regional prediction index.

[0043] The calculation formula of the first regional coefficient is:

[0044]

[0045] Among them, Region 1 Represents the first area coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

[0046] The second regional coefficient is calculated as:

[0047]

[0048] Among them, Region 2 represents the second area coefficient; Ter represents the terrain flatness of the sub-area; α 1 represents the first impact weight; TH represents the first threshold; D() represents the distance calculation formula, Temp represents the predicted temperature data of the sub-area, ST represents the suitable temperature data of the bees; α 2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-area; SW represents the suitable wind speed data of the bees; exp represents an exponential function with a natural constant as the base.

[0049] The calculation formula of the regional prediction index is:

[0050]

[0051] Among them, Fore represents the regional forecast index; Region 1 Indicates the first regional coefficient; Region 2 Represents the second region coefficient.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. The present invention constructs a remaining flowering period prediction model based on a convolutional neural network; preprocesses and segments flower image data of an input model to obtain flower sub-images; identifies the flower sub-images and divides them according to flower types to obtain a flower recognition image set; obtains standard flowering period data according to flower types, and obtains the remaining flowering period of flowers through the standard flowering period data, color feature data, texture feature data and shape feature data; corrects the remaining flowering period of flowers in the flower sub-images according to the remaining flowering period of all flowers in the flower recognition image set, thereby accurately predicting the remaining flowering period of flowers.

[0054] 2. The present invention identifies the target breeding area and obtains the types of flowers contained therein; constructs a closed experimental environment according to the types of flowers, and selects bee samples for experiments; obtains the frequency of bees' visits to different types of flowers and the honey produced by a single type of flower, uses the bee visit frequency as the flower attraction index, and obtains the flower nectar quality index according to the honey quality; accurately obtains the flower quality coefficient based on the flower attraction index to bees and the flower nectar quality index.

[0055] 3. The present invention obtains the nectar amount of a single type of flower through the number of flowers and the nectar content of the flowers, and weights the nectar amount of a single type of flower through the quality coefficient of the flowers to obtain the comprehensive nectar content of the sub-region; the total amount of nectar collected by the bees is obtained through the nectar collection amount of the bees and the number of days of the change cycle; the first regional coefficient is obtained according to the comprehensive nectar content of the sub-region and the total amount of nectar collected by the bees; the nectar content of the sub-region can be accurately quantified through the first regional coefficient.

[0056] 4. The present invention takes into account the impact of terrain, temperature and wind speed on honey collection by bees, and reflects the terrain interference factor of the sub-region through the terrain flatness; reflects the temperature interference factor through the difference between the predicted temperature data and the suitable temperature data for bees, and reflects the wind speed interference factor through the difference between the predicted wind speed data and the suitable wind speed data for bees; and obtains the second regional coefficient by comprehensively considering the interference factors of terrain, temperature and wind speed. The second regional coefficient can accurately reflect the interference caused by the sub-region environment to honey collection.

[0057] 5. The present invention comprehensively quantifies the nectar content and nectar content of the sub-region through the first regional coefficient, and quantifies the environmental interference factors in the sub-region through the second regional coefficient; the honey production situation of the sub-region is predicted through the first regional coefficient and the second regional coefficient of the sub-region, and the honey production efficiency of different sub-regions is accurately compared. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of a process for predicting honey production of the present invention;

[0059] Figure 2 It is a structural schematic diagram of the remaining flowering period prediction model of the present invention;

[0060] Figure 3 It is a schematic diagram of sub-area division of the present invention;

[0061] Figure 4 It is a structural schematic diagram of a prediction system for honey production of the present invention;

[0062] In the figure: W represents the target breeding area; b, h and r represent the cluster centers; B, H and R represent the circular areas obtained according to the cluster centers and the bee activity distances. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] In the process of bee breeding and honey production through beehives, the location selection of the beehive is an important consideration; by adjusting the location of the beehive, the bees can be close to the rich nectar source, avoid the interference of the harsh natural environment, and improve the honey production and quality. In order to better guide the location of the beehive, the present invention proposes a prediction method and system for honey production.

[0065] Embodiment 1

[0066] The present invention proposes a method for predicting honey production, the method flow is as follows: Figure 1 As shown, including:

[0067] S10. Determine the change cycle of the beehive position; at the beginning of the change cycle, identify the target breeding area to obtain flower image data and flower distribution data.

[0068] The change period is the period of position change of the beehive, which is determined according to the collection efficiency of the bees and the flowering period of the area where they are located.

[0069] Determine a target breeding area where bee breeding and honey production can be carried out, photograph and identify the target breeding area through unmanned aerial vehicle remote sensing technology or satellite remote sensing technology, and obtain flower image data and flower distribution data; wherein the flower image data is a flower image data set obtained by remote sensing technology; the flower distribution data is the distribution position data of flowers in the overhead view of the target breeding area.

[0070] S20. Construct a remaining flowering period prediction model to identify the flower image data to obtain the remaining flowering period of the flowers; delete the flower distribution data whose remaining flowering period is less than the change period to obtain adjusted distribution data.

[0071] Generally, the nectar content of flowers fluctuates as the flowering period progresses;

[0072] In the early stage of flowering, the flowers have not yet been visited by a large number of pollinators. In order to attract pollinators, such as bees and butterflies, the flowers secrete more and denser nectar. At this time, there is less competition and the collection efficiency is higher. In addition, the newly opened flowers are relatively fresh, and the quality and quantity of the nectar are relatively good.

[0073] In the middle of the flowering period, the physiological activities of the flowers are most vigorous, and the nectar secretion also increases to attract more pollinators and improve pollination efficiency. The nectar content may reach its peak at this time.

[0074] In the late flowering period, as the flowering period progresses, the flowers gradually age, physiological activities weaken, the secretion of nectar begins to decrease, and at the same time they face more competitors.

[0075] Therefore, the flowering period is an important factor affecting the efficiency and quality of honey bees. In order to accurately measure and identify the flowers in the target breeding area, the present invention constructs a remaining flowering period prediction model to predict the remaining flowering period of flowers; the remaining flowering period of flowers is identified by changing the cycle, and flowers in the late flowering period are eliminated.

[0076] The remaining flowering period prediction model is constructed based on a convolutional neural network, and its structure is as follows: Figure 2 As shown; including a flower image input layer, a flower image recognition layer, a flower feature recognition layer, a flowering period prediction layer, a flowering period correction layer and a flowering period output layer;

[0077] The flower image input layer is used to input the obtained flower image data into the model, and pre-process the flower image data to obtain a pre-processed flower image;

[0078] The flower image recognition layer recognizes and segments the pre-processed flower image to obtain a flower sub-image; determines the flower species in the flower sub-image, and organizes the flower sub-images of the same species into a flower recognition image set;

[0079] The flower feature recognition layer is used to recognize the color features, texture features and shape features of the flowers in the flower recognition image set;

[0080] The flowering period prediction layer obtains standard flowering period data according to the flower species, and identifies the remaining flowering period of the flowers in the flower recognition image set through the standard flowering period data, color feature data, texture feature data and shape feature data; wherein the standard flowering period data is obtained by monitoring the growth cycle of the same type of flowers in the same area;

[0081] The flowering period correction layer obtains the remaining flowering period of all flowers in the flower recognition image set, and corrects the remaining flowering period of flowers in the flower sub-image according to the remaining flowering period of all flowers;

[0082] The flowering period output layer is used to output the remaining flowering period of the flower.

[0083] The present invention constructs a remaining flowering period prediction model based on a convolutional neural network; preprocesses and segments flower image data of an input model to obtain flower sub-images; identifies the flower sub-images and divides them according to flower types to obtain a flower recognition image set; obtains standard flowering period data according to the flower types, and obtains the remaining flowering period of the flowers through the standard flowering period data, color feature data, texture feature data and shape feature data; corrects the remaining flowering period of the flowers of the flower sub-images in the flower recognition image set through the remaining flowering periods of all flowers in the flower recognition image set, thereby accurately predicting the remaining flowering period of the flowers.

[0084] S30. Obtain flower coordinate data according to the adjusted distribution data, perform clustering algorithm identification on the flower coordinate data, and obtain multiple cluster centers.

[0085] After the adjusted distribution data is obtained, two-dimensional coordinates are constructed according to the top view of the target breeding area; and flower coordinate data in the adjusted distribution data are obtained according to the two-dimensional coordinates.

[0086] The flower coordinate data is identified by a clustering algorithm to obtain a plurality of cluster centers; the cluster centers are used as candidate locations for beehives.

[0087] S40. Identify the bees raised in the beehive to obtain the nectar collection amount and the bee activity distance; and obtain multiple sub-areas with the cluster center as the center of the circle and the activity distance as the radius.

[0088] The nectar collection amount is obtained by conducting experiments on bees, quantitatively analyzing and taking the average value to obtain the average daily nectar collection amount of worker bees, and then combining the number of worker bees collecting nectar to obtain the nectar collection amount.

[0089] The activity distance of bees is used to indicate the straight-line distance between the honey-collecting worker bees and the beehive without affecting the honey-collecting efficiency; it is obtained by identifying the activity trajectory of the honey-collecting worker bees and the amount of nectar collected.

[0090] The sub-areas are divided as follows: Figure 3 As shown in the figure, W represents the target breeding area, b, h and r represent the cluster centers identified by the clustering algorithm, and the circular areas B, H and R are circular areas obtained according to the cluster centers and the bee activity distances; wherein, when the range of the sub-area exceeds the target breeding area, the overlapping area is taken as the sub-area; Figure 3 In the figure, the circular areas B and H are sub-areas, and the overlapping area of ​​the circular area R and the target breeding area W is taken as the sub-area obtained by the cluster center r.

[0091] S50. Collect and predict data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data.

[0092] The flower quality coefficient is determined by the flower's attraction index to bees and the flower's nectar quality index;

[0093] Identify all flower species in the target breeding area, collect the same number of flowers of various types and set up a closed first experimental environment; select a first bee sample and place it in the first experimental environment, identify the frequency of the first bee sample visiting different types of flowers through a positioning system, and use the frequency as the attraction index of the flower;

[0094] Identify all flower species in the target breeding area, collect a single type of flowers to set up a second closed experimental environment; select a second bee sample and place it in the experimental environment, and perform chemical analysis on the honey produced by the second bee sample to determine the substance ratio; and determine the nectar quality index by having an expert group study the substance ratio.

[0095] Different pollens contain different vitamins, minerals and other nutrients, and the material composition of nectar produced by different pollens is also different. Chemical analysis can be used to determine the material composition of nectar, including impurities, proteins, vitamins and sugars. An expert group conducts a comprehensive evaluation of the material composition and proportion of different honeys to obtain a nectar quality index.

[0096] The calculation formula of the flower quality coefficient is:

[0097] Qua i =δ 1 *Att i +δ 2 *Nma i ;

[0098] Among them, Qua i represents the quality coefficient of the i-th flower; Atti Nma represents the attraction index of the i-th flower; i represents the nectar quality coefficient of the i-th flower; δ 1 First flower weight; δ 2 Second flower weight: The first flower weight and the second flower weight are verified and adjusted through historical data obtained through multiple experiments.

[0099] The present invention identifies a target breeding area to obtain the types of flowers contained therein; constructs a closed experimental environment according to the types of flowers, and selects bee samples for experiments; obtains the visit frequency of bees to different types of flowers and the honey produced by a single type of flower, uses the bee visit frequency as the attraction index of the flower, and obtains the nectar quality index of the flower according to the quality of the honey; and accurately obtains the quality coefficient of the flower according to the attraction index of the flower to the bee and the nectar quality index of the flower.

[0100] S60. Calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; and calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set.

[0101] The calculation formula of the first regional coefficient is:

[0102]

[0103] Among them, Region 1 represents the first area coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

[0104] The present invention obtains the nectar amount of a single type of flower through the number of flowers and the nectar content of the flowers, and weights the nectar amount of the single type of flower through the quality coefficient of the flowers to obtain the comprehensive nectar content of the sub-region; obtains the total amount of nectar collected by the bees through the nectar collection amount and the number of days of the change cycle; obtains the first regional coefficient according to the comprehensive nectar content of the sub-region and the total amount of nectar collected by the bees; the nectar content of the sub-region can be accurately reflected through the first regional coefficient.

[0105] The second regional coefficient is calculated as:

[0106]

[0107] Among them, Region 2represents the second area coefficient; Ter represents the terrain flatness of the sub-area; α 1 represents the first impact weight; TH represents the first threshold; D() represents the distance calculation formula, Temp represents the predicted temperature data of the sub-area, ST represents the suitable temperature data of the bees; α 2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-area; SW represents the suitable wind speed data of the bees; exp represents an exponential function with a natural constant as the base.

[0108] Among them, the distance calculation formula D() is used to obtain the distance between the predicted data and the appropriate data, reflect the difference between the predicted data and the appropriate data through the distance, and then quantify the difference according to the distance threshold.

[0109] The terrain flatness is used to reflect the complexity of the terrain of the sub-region, and is obtained through the three-dimensional terrain data of the sub-region. The terrain flatness value ranges from 0 to 1. The larger the terrain flatness, the more complex the terrain.

[0110] The present invention takes into account the influence of terrain, temperature and wind speed on honey collection by bees, reflects the terrain interference factor of the sub-region through the terrain flatness; reflects the temperature interference factor through the difference between the predicted temperature data and the suitable temperature data for bees, and reflects the wind speed interference factor through the difference between the predicted wind speed data and the suitable wind speed data for bees; and obtains the second regional coefficient by comprehensively considering the interference factors of terrain, temperature and wind speed, and the interference caused by the sub-region environment to honey collection can be accurately reflected through the second regional coefficient.

[0111] S70. Calculate a regional prediction index based on the first regional coefficient and the second regional coefficient of the sub-region; and determine the location of the beehive based on the regional prediction index.

[0112] The calculation formula of the regional prediction index is:

[0113]

[0114] Among them, Fore represents the regional forecast index; Region 1 Indicates the first regional coefficient; Region 2 Represents the second region coefficient.

[0115] The present invention comprehensively quantifies the nectar content and nectar quality of the sub-region through the first regional coefficient, and quantifies the environmental interference factors in the sub-region through the second regional coefficient; the honey production efficiency of the sub-region is quantified through the first regional coefficient and the second regional coefficient of the sub-region, and the efficiency comparison of different sub-regions is accurately achieved.

[0116] The present invention first determines a change cycle, identifies a target breeding area at the beginning of the change cycle, and obtains flower image data and flower distribution data; constructs a remaining flowering period prediction model to identify and adjust the flower image data to obtain adjusted distribution data; obtains a cluster center according to the adjusted distribution data and a clustering algorithm; identifies and obtains the nectar collection amount of bees and the bee activity distance; obtains multiple sub-areas according to the cluster center and the bee activity distance, and collects and obtains a regional flower data set and a regional environmental data set; obtains a first regional coefficient according to the regional flower data set and the nectar collection amount; obtains a second regional coefficient according to the regional environmental coefficient set; obtains a regional prediction index according to the first regional coefficient and the second regional coefficient; predicts and compares the honey production efficiency of different beehive locations through the regional prediction index, and determines the optimal beehive location.

[0117] Embodiment 2

[0118] Company A's main business is honey production. It has created a full-process industrial chain from bee breeding, honey production and honey processing. In order to ensure the efficiency and stability of honey production, it has established multiple bee breeding groups and carried out migratory breeding in accordance with the flowering period.

[0119] Among them, the breeding migration route of bee breeding group Y is shown in Table 1.

[0120] Table 1 The breeding and migration routes of bee breeding group Y

[0121] time Place content March-April Anhui Collecting rapeseed nectar May-June Shandong Collecting acacia and vitex nectar July-August Jiangsu Rice nectar, Koelreuteria paniculata nectar, etc. September-October Zhejiang Tea Tree Nectar November-February Anhui winter

[0122] After arriving at the target breeding area, the location of the beehive is extremely critical to the honey production efficiency. On the one hand, the location of the beehive is affected by the distribution of flowers in the breeding area, and being close to abundant nectar sources can improve the efficiency of bees in collecting nectar. On the other hand, it is affected by terrain and climate factors. Complex terrain conditions and harsh climate will interfere with the honey-collecting behavior of bees.

[0123] To this end, bee breeding group Y uses a honey production prediction system described in the present invention to determine the location of the beehive; the structure of the system is as follows Figure 4 As shown, including:

[0124] A breeding area identification module determines the change period of the beehive position; at the beginning of the change period, the target breeding area is identified to obtain flower image data and flower distribution data;

[0125] Take the example of bee breeding group Y going to Anhui to collect rapeseed nectar in March and April.

[0126] At the beginning of the change cycle, bee breeding group Y uses a drone to identify the target breeding area, obtains image data of the target breeding area, and identifies and processes the image data of the target breeding area to obtain flower image data and flower distribution data of rapeseed flowers in the target breeding area.

[0127] A flowering period recognition module is used to construct a remaining flowering period prediction model to recognize the flower image data to obtain the remaining flowering period of the flowers; the flower distribution data of the flowers whose remaining flowering period is less than the change period are deleted to obtain the adjusted distribution data;

[0128] The remaining flowering period prediction model includes a flower image input layer, a flower image recognition layer, a flower feature recognition layer, a flowering period prediction layer, a flowering period correction layer and a flowering period output layer;

[0129] The flower image input layer is used to input the obtained flower image data into the model, and pre-process the flower image data to obtain a pre-processed flower image;

[0130] The flower image recognition layer recognizes and segments the pre-processed flower image to obtain a flower sub-image; determines the flower species in the flower sub-image, and organizes the flower sub-images of the same species into a flower recognition image set;

[0131] The flower feature recognition layer is used to recognize the color features, texture features and shape features of the flowers in the flower recognition image set;

[0132] The flowering period prediction layer obtains standard flowering period data according to the flower species, and identifies the remaining flowering period of the flowers in the flower recognition image set through the standard flowering period data, color feature data, texture feature data and shape feature data;

[0133] The flowering period correction layer obtains the remaining flowering period of all flowers in the flower recognition image set, and corrects the remaining flowering period of flowers in the flower sub-image according to the remaining flowering period of all flowers;

[0134] The flowering period output layer is used to output the remaining flowering period of the flower.

[0135] The flower image of rapeseed flowers is identified by the remaining flowering period prediction model, and the flower distribution data of flowers whose remaining flowering period is less than the change period are deleted to obtain the adjusted distribution data of rapeseed flowers.

[0136] The flower clustering recognition module obtains flower coordinate data according to the adjusted distribution data, and performs clustering algorithm recognition on the flower coordinate data to obtain multiple cluster centers.

[0137] The breeding area division module identifies the bees bred in the beehive to obtain the nectar collection amount and the bee activity distance; taking the cluster center as the center of the circle and the activity distance as the radius, multiple sub-areas are obtained.

[0138] After the bee breeding group Y arrived at the target breeding area, it identified the area, collected data, and divided it into multiple sub-areas; the sub-area division is shown in Table 2.

[0139] Table 2: Sub-regional division of target breeding areas

[0140] serial number Flower Type Sub-area 1 Rapeseed Sub-area 2 Rapeseed Sub-area 3 Rapeseed

[0141] The regional data identification module collects and predicts data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data;

[0142] The flower quality coefficient is determined by the flower's attraction index to bees and the flower's nectar quality index;

[0143] Identify all flower species in the target breeding area, collect the same number of flowers of various types and set up a closed first experimental environment; select a first bee sample and place it in the first experimental environment, identify the frequency of the first bee sample visiting different types of flowers through a positioning system, and use the frequency as the attraction index of the flower;

[0144] Identify all flower species in the target breeding area, collect a single type of flowers to set up a second closed experimental environment; select a second bee sample and place it in the experimental environment, and perform chemical analysis on the honey produced by the second bee sample to determine the substance ratio; and determine the nectar quality index by having an expert group study the substance ratio.

[0145] The calculation formula of the flower quality coefficient is:

[0146] Qua i =δ 1 *Att i +δ 2 *Nma i ;

[0147] Among them, Qua i represents the quality coefficient of the i-th flower; Att i Nma represents the attraction index of the i-th flower; i represents the nectar quality coefficient of the i-th flower; δ 1 First flower weight; δ 2Second flower weight.

[0148] Bee breeding group Y sets the weight of the first flower and the weight of the second flower to 0.5. Table 3 can be obtained through calculation.

[0149] Table 3 Flower quality coefficient calculation data table

[0150] Flower Type Attraction Index Nectar Quality Index Flower quality factor Rapeseed 1 0.80 0.90

[0151] Since there is only one nectar source, rapeseed flowers, in the target breeding area, its attraction index for bees is 1.

[0152] A regional coefficient calculation module is used to calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; and to calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set;

[0153] The calculation formula of the first regional coefficient is:

[0154]

[0155] Among them, Region 1 represents the first area coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

[0156] The second regional coefficient is calculated as:

[0157]

[0158] Among them, Region 2 represents the second area coefficient; Ter represents the terrain flatness of the sub-area; α 1 represents the first impact weight; TH represents the first threshold; D() represents the distance calculation formula, Temp represents the predicted temperature data of the sub-area, ST represents the suitable temperature data of the bees; α 2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-area; SW represents the suitable wind speed data of the bees; exp represents an exponential function with a natural constant as the base.

[0159] The beehive location determination module calculates a regional prediction index according to the first regional coefficient and the second regional coefficient of the sub-region; and determines the beehive location according to the regional prediction index.

[0160] The calculation formula of the regional prediction index is:

[0161]

[0162] Among them, Fore represents the regional forecast index; Region 1 Indicates the first regional coefficient; Region 2 Represents the second region coefficient.

[0163] Bee breeding group Y obtains multiple sub-areas by identifying and dividing the target breeding area; then collects and measures data on the sub-areas to obtain regional prediction indexes of the sub-areas, and selects the optimal beehive location based on the regional prediction index.

[0164] The present invention first determines a change cycle, identifies a target breeding area at the beginning of the change cycle, and obtains flower image data and flower distribution data; constructs a remaining flowering period prediction model to identify and adjust the flower image data to obtain adjusted distribution data; obtains a cluster center according to the adjusted distribution data and a clustering algorithm; identifies and obtains the nectar collection amount of bees and the bee activity distance; obtains multiple sub-areas according to the cluster center and the bee activity distance, and collects and obtains a regional flower data set and a regional environmental data set; obtains a first regional coefficient according to the regional flower data set and the nectar collection amount; obtains a second regional coefficient according to the regional environmental coefficient set; obtains a regional prediction index according to the first regional coefficient and the second regional coefficient; predicts and compares the honey production efficiency of different beehive locations through the regional prediction index, and determines the optimal beehive location.

[0165] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting honey production, characterized in that: include: S10. Determine the change cycle of the hive position; at the beginning of the change cycle, identify the target breeding area to obtain flower image data and flower distribution data; S20. Constructing a remaining flowering period prediction model to identify the flower image data to obtain the remaining flowering period of the flowers; deleting the flower distribution data whose remaining flowering period is less than the change period to obtain the adjusted distribution data; S30. Obtain flower coordinate data according to the adjusted distribution data, perform clustering algorithm identification on the flower coordinate data, and obtain multiple cluster centers; S40. Identify the bees raised in the beehive to obtain the nectar collection amount and the bee activity distance; take the cluster center as the center of the circle and the bee activity distance as the radius to obtain multiple sub-areas; S50. Collect and predict data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data; S60. Calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set; S70. Calculate a regional prediction index based on the first regional coefficient and the second regional coefficient of the sub-region; and determine the location of the beehive based on the regional prediction index.

2. A method for predicting honey production according to claim 1, characterized in that: The remaining flowering period prediction model includes a flower image input layer, a flower image recognition layer, a flower feature recognition layer, a flowering period prediction layer, a flowering period correction layer and a flowering period output layer; The flower image input layer is used to input the obtained flower image data into the model, and pre-process the flower image data to obtain a pre-processed flower image; The flower image recognition layer recognizes and segments the pre-processed flower image to obtain a flower sub-image; Determining the types of flowers in the flower sub-images, and organizing flower sub-images of the same type into a flower recognition image set; The flower feature recognition layer is used to recognize the color features, texture features and shape features of the flowers in the flower recognition image set; The flowering period prediction layer obtains standard flowering period data according to the flower species, and identifies the remaining flowering period of the flowers in the flower recognition image set through the standard flowering period data, color feature data, texture feature data and shape feature data; The flowering period correction layer obtains the remaining flowering period of all flowers in the flower recognition image set, and corrects the remaining flowering period of flowers in the flower sub-image according to the remaining flowering period of all flowers; The flowering period output layer is used to output the remaining flowering period of the flower.

3. A method for predicting honey production according to claim 1, characterized in that: The flower quality coefficient is determined by the flower's attraction index to bees and the flower's nectar quality index; Identify all flower species in the target breeding area, collect the same number of flowers of various types and set up a closed first experimental environment; select a first bee sample and place it in the first experimental environment, identify the frequency of the first bee sample visiting different types of flowers through a positioning system, and use the frequency as the attraction index of the flower; Identify all flower species in the target breeding area, collect a single type of flowers to set up a second closed experimental environment; select a second bee sample and place it in the experimental environment, and perform chemical analysis on the honey produced by the second bee sample to determine the substance ratio; and determine the nectar quality index by having an expert group study the substance ratio.

4. A method for predicting honey production according to claim 1, characterized in that: The calculation formula of the first regional coefficient is: Where, Region1 represents the first region coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

5. A method for predicting honey production according to claim 1, characterized in that: The second regional coefficient is calculated as: Among them, Region2 represents the second region coefficient; Ter represents the terrain flatness of the sub-region; α1 represents the first influence weight; TH represents the first threshold; D( ) represents the distance calculation formula, Temp represents the predicted temperature data of the sub-region, ST represents the suitable temperature data of the bees; α2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-region; SW represents the suitable wind speed data of the bees; exp represents the exponential function with the natural constant as the base.

6. A method for predicting honey production according to claim 1, characterized in that: The calculation formula of the regional prediction index is: Among them, Fore represents the regional prediction index; Region1 represents the first regional coefficient; Region2 represents the second regional coefficient.

7. A prediction system for honey production, characterized in that: include: A breeding area identification module determines the change period of the beehive position; at the beginning of the change period, the target breeding area is identified to obtain flower image data and flower distribution data; A flowering period recognition module is used to construct a remaining flowering period prediction model to recognize the flower image data to obtain the remaining flowering period of the flowers; the flower distribution data of the flowers whose remaining flowering period is less than the change period are deleted to obtain the adjusted distribution data; A flower clustering recognition module obtains flower coordinate data according to the adjusted distribution data, and performs clustering algorithm recognition on the flower coordinate data to obtain multiple cluster centers; The breeding area division module identifies the bees bred in the beehive to obtain the nectar collection amount and the bee activity distance; taking the cluster center as the center of the circle and the activity distance as the radius, a plurality of sub-areas are obtained; The regional data identification module collects and predicts data on the sub-region through the Internet of Things device to obtain a regional flower data set and a regional environmental data set; the regional flower data set includes flower species, flower quantity, flower nectar content and flower quality coefficient; the regional environmental data set includes terrain flatness, predicted temperature data and predicted wind speed data; A regional coefficient calculation module is used to calculate the first regional coefficient of the sub-region according to the regional flower data set and the nectar collection amount; and to calculate the second regional coefficient of the sub-region according to the regional environmental coefficient set; The beehive location determination module calculates a regional prediction index according to the first regional coefficient and the second regional coefficient of the sub-region; and determines the beehive location according to the regional prediction index.

8. A honey production prediction system according to claim 7, characterized in that: The calculation formula of the first regional coefficient is: Where, Region1 represents the first region coefficient; Sum i Nec represents the number of flowers of the i-th type; i represents the nectar content of the i-th flower; Qua i represents the quality coefficient of the i-th flower; n represents the number of flower species identified; HOC represents the amount of nectar collected by bees; T represents the number of days of the change cycle; ln represents the logarithmic function with natural constant as the base.

9. A honey production prediction system according to claim 7, characterized in that: The second regional coefficient is calculated as: Among them, Region2 represents the second region coefficient; Ter represents the terrain flatness of the sub-region; α1 represents the first influence weight; TH represents the first threshold; D( ) represents the distance calculation formula, Temp represents the predicted temperature data of the sub-region, ST represents the suitable temperature data of the bees; α2 represents the second influence weight; WH represents the second threshold; Wind represents the predicted wind speed data of the sub-region; SW represents the suitable wind speed data of the bees; exp represents the exponential function with the natural constant as the base.

10. A honey production prediction system according to claim 7, characterized in that: The calculation formula of the regional prediction index is: Among them, Fore represents the regional prediction index; Region1 represents the first regional coefficient; Region2 represents the second regional coefficient.

Images