CN117035618B - Management method and system for storage bin of organic water-soluble fertilizer - Google Patents
Management method and system for storage bin of organic water-soluble fertilizer Download PDFInfo
- Publication number
- CN117035618B CN117035618B CN202311293783.2A CN202311293783A CN117035618B CN 117035618 B CN117035618 B CN 117035618B CN 202311293783 A CN202311293783 A CN 202311293783A CN 117035618 B CN117035618 B CN 117035618B
- Authority
- CN
- China
- Prior art keywords
- fertilizers
- fertilizer
- stack
- dimensional vector
- obtaining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000003337 fertilizer Substances 0.000 title claims abstract description 548
- 238000003860 storage Methods 0.000 title claims abstract description 118
- 238000007726 management method Methods 0.000 title claims description 15
- 239000013598 vector Substances 0.000 claims abstract description 167
- 230000006866 deterioration Effects 0.000 claims abstract description 57
- 230000008859 change Effects 0.000 claims abstract description 44
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000002159 abnormal effect Effects 0.000 claims description 28
- 238000004422 calculation algorithm Methods 0.000 claims description 20
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 238000012937 correction Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000004321 preservation Methods 0.000 abstract 1
- 239000003895 organic fertilizer Substances 0.000 description 10
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000002035 prolonged effect Effects 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 239000007788 liquid Substances 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 238000003973 irrigation Methods 0.000 description 2
- 230000002262 irrigation Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 230000002542 deteriorative effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 239000007921 spray Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention relates to the technical field of bin management, in particular to a method and a system for managing an organic water-soluble fertilizer storage bin, wherein the method comprises the following steps: collecting humidity data and weight change of each stack of fertilizer in the storage bin so as to obtain storage safety coefficients of the fertilizer; obtaining a first high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry; constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions; based on the first high-dimensional vector and the second high-dimensional vector, the local outlier coefficient of each stack of fertilizers is obtained, and then the deterioration tendency coefficient is obtained through the local outlier coefficient of each stack of fertilizers, the first high-dimensional vector, the second high-dimensional vector and the warehouse-in time; and acquiring edge weights based on the metamorphic tendency coefficients to further obtain a plurality of groups of matched pairs, and carrying out position adjustment on the metamorphic fertilizer by taking all the matched pairs as references. The invention can efficiently manage the fertilizer warehouse and maximally prolong the preservation time of the fertilizer.
Description
Technical Field
The invention relates to the technical field of bin management, in particular to a method and a system for managing an organic water-soluble fertilizer storage bin.
Background
The water-soluble fertilizer is a novel chemical fertilizer developed in China in recent years, is a multi-element compound fertilizer which can be completely dissolved in water, can be rapidly dissolved in water, is easier to be absorbed by crops, has high absorption and utilization rate, and can meet the nutrition requirements of high-yield crops in the rapid growth period; the water-soluble fertilizer is applied to the combination of the micro-spray irrigation, drip irrigation and the like in facility agriculture, so that the integration of the water and the fertilizer is realized, the effects of saving water, fertilizer and labor are achieved, and when water resources are in shortage, the application of the water-soluble fertilizer becomes one of measures of agricultural efficiency improvement and income improvement of farmers.
However, the organic water-soluble fertilizer is more severe in storage condition, needs to be stored in a place selected to be dried in a storehouse and needs to be waterproof and dampproof, but a part of areas which are easy to wet are inevitably present in the storehouse, and the organic fertilizer is easy to dissolve in water, is easy to agglomerate or become liquid to flow off after being wetted or soaked in water, so that the fertilizer efficiency is seriously affected.
Since the environments are different and the deterioration tendency of the organic fertilizer is not linear, it is difficult to determine the deterioration tendency through experience and simple physicochemical analysis; the existing method for judging whether the fertilizer has the deterioration trend mainly depends on simply counting the weight after moisture absorption, however, the method cannot well analyze the influence caused by space factors, and cannot accurately estimate the fertilizer to be delivered under the condition of mass mixing, so that the analysis difficulty of storage management is high, the management is rough, and the storage time of the fertilizer cannot be optimally prolonged.
Disclosure of Invention
In order to solve the problem that storage management of organic water-soluble fertilizers is rough and fertilizer storage time cannot be optimally delayed at present, the invention aims to provide a storage bin management method and system for organic water-soluble fertilizers, and the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides a method for managing an organic water-soluble fertilizer storage bin, including the steps of:
acquiring humidity data of the position of each stack of fertilizer in the storage bin, acquiring weight change of each stack of fertilizer, and acquiring storage safety coefficients of the fertilizer based on the weight change of each stack of fertilizer and the humidity data of the position;
obtaining a first high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry; constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions; acquiring local outlier coefficients of corresponding fertilizers based on the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizers;
obtaining a deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehousing time of each stack of fertilizers;
acquiring edge weights between two stacks of fertilizers according to the deterioration tendency coefficients corresponding to each two stacks of fertilizers, and acquiring matching pairs in all the fertilizers based on the edge weights; and identifying all the spoiled fertilizers according to all the matching pairs and the trained neural network, and adjusting the positions of the spoiled fertilizers based on the matching pairs.
Preferably, the step of obtaining the storage safety coefficient of the fertilizer based on the weight change of each stack of fertilizer and the humidity data of the position comprises the following steps:
obtaining maximum humidity data of the position of the fertilizer in a preset time period and a difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period, and marking the difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period as a maximum weight change value;
taking the reciprocal of the maximum value of the weight change of the fertilizer in a preset time period under normal conditions as a correction coefficient, obtaining the product of the maximum value of the weight change corresponding to the fertilizer and the correction coefficient, and selecting the minimum value between the product and a constant 1; taking the difference value between the constant 1 and the minimum value as a molecule, and taking the maximum humidity data of the position where the fertilizer is located as a denominator to obtain a ratio result which is the storage safety coefficient.
Preferably, the step of obtaining the first high-dimensional vector based on the position change of each stack of fertilizer from the warehouse entry includes:
meshing and dividing the bin, and marking the grid position of the fertilizer and the eight neighborhood grid position of the fertilizer as 1 and marking the other grid positions as 0 for any fertilizer; and marking grid positions according to all position changes at the beginning of fertilizer warehouse entry, and sequentially arranging to obtain a first high-dimensional vector.
Preferably, the step of constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions comprises:
constructing an initial vector with a preset length, wherein all element values in the initial vector are 0;
acquiring worst storage safety coefficients of the fertilizer at each position after the storage of the fertilizer begins, and replacing element values in the initial vector with the worst storage safety coefficients of the fertilizer at each position to obtain a second high-dimensional vector of the fertilizer;
wherein, the worst storage safety coefficient of fertilizer corresponding to each position is: and obtaining a difference value between the constant 1 and the minimum storage safety coefficient of the fertilizer at the position, wherein the difference value is the worst storage safety coefficient.
Preferably, the step of obtaining the local outlier coefficient of the corresponding fertilizer based on the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizer comprises the following steps:
for any two stacks of fertilizers, acquiring a hamming distance between the first high-dimensional vectors corresponding to the two stacks of fertilizers, arranging elements in the second high-dimensional vectors corresponding to each stack of fertilizers in a descending order to obtain a new second high-dimensional vector, acquiring a cosine distance between the new second high-dimensional vectors corresponding to the two stacks of fertilizers, and taking the product of the hamming distance and the cosine distance as the distance between the two stacks of fertilizers;
Taking any fertilizer as a target fertilizer, and obtaining k nearest adjacent fertilizers of the target fertilizer based on the distance between the target fertilizer and other fertilizers, wherein k is a positive integer; adding k adjacent distances between the target fertilizer and the k nearest adjacent fertilizers corresponding to the target fertilizer to obtain a summation result, and calculating the ratio of the summation result to the maximum distance among the k adjacent distances as the local reachable density of the target fertilizer;
and obtaining the average value of the local reachable densities of the k nearest fertilizers of the target fertilizer, wherein the ratio of the average value of the local reachable densities of the k nearest fertilizers to the local reachable density of the target fertilizer is the local outlier coefficient of the target fertilizer.
Preferably, the step of obtaining the deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector, and the warehouse-in time of each stack of fertilizers includes:
calculating the average value of non-zero elements in the second high-dimensional vector of the fertilizer, and calculating the product result of the average value and the warehousing time, wherein the warehousing time is the total number of days of warehousing of the fertilizer; accumulating all elements in the first high-dimensional vector to obtain an accumulated value, and marking the ratio of the product result to the accumulated value as a first ratio;
And obtaining a summation result of the local outlier coefficient and a constant 1, and taking the product of the summation result and the first ratio as a deterioration tendency coefficient of the fertilizer.
Preferably, the step of obtaining the edge weight between two stacks of fertilizers according to the deterioration tendency coefficients corresponding to each two stacks of fertilizers and obtaining the matching pairs in all fertilizers based on the edge weight includes:
acquiring the absolute value of the difference value of the deterioration tendency coefficients corresponding to each two stacks of fertilizers and the Euclidean distance between the corresponding positions of the two stacks of fertilizers in the storage bin; the ratio of the absolute value of the difference to the Euclidean distance is the edge weight between two stacks of fertilizers;
and obtaining the matched fertilizers of each stack of fertilizers according to the edge weight between any two stacks of fertilizers and based on a K-M matching algorithm, wherein each stack of fertilizers and the matched fertilizers form a group of matching pairs.
Preferably, the step of identifying all spoiled fertilizers based on all the matched pairs and the trained neural network comprises:
the neural network is a twin network, the twin network after training outputs normal fertilizers in all matching pairs, and outputs high-dimensional vectors of each stack of fertilizers, the normal fertilizers in all matching pairs are used as reference groups, and the abnormal fertilizers in all matching pairs are fertilizers to be judged;
Acquiring cosine distances between each stack of fertilizers to be discriminated and high-dimensional vectors corresponding to all fertilizers in the reference group, and constructing a columnar histogram according to all cosine distances between each stack of fertilizers to be discriminated and the fertilizers in the reference group; obtaining corresponding distribution vectors through the columnar histograms of each stack of fertilizers to be discriminated, and obtaining L1 distance between the distribution vectors of each two stacks of fertilizers to be discriminated;
obtaining local abnormal factors of each stack of fertilizers to be discriminated through the L1 distance between each two stacks of fertilizers to be discriminated and based on an LOF algorithm, wherein when the local abnormal factors are larger than an abnormal threshold value, the corresponding fertilizers to be discriminated are metamorphic fertilizers.
Preferably, the step of adjusting the position of the spoiled fertilizer based on the matching pair includes:
and on the basis of the matching pairs, obtaining the matched fertilizer of each metamorphic fertilizer, and replacing and adjusting the positions of the metamorphic fertilizer and the matched fertilizer corresponding to the metamorphic fertilizer.
In a second aspect, another embodiment of the present invention provides an organic water-soluble fertilizer storage bin management system, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of any one of the methods described in the foregoing methods for organic water-soluble fertilizer storage bin management when the computer program is executed.
The invention has the following beneficial effects: the existing method for judging the storage condition of the organic water-soluble fertilizer mainly depends on weight change, but the method cannot accurately estimate the condition of the fertilizer, so that the storage bin is roughly managed; in order to prolong the storage time of the fertilizer and improve the management effect on the storage bin, the embodiment of the invention firstly obtains the storage safety coefficient through the humidity data of the position of each stack of the fertilizer and the weight change of the fertilizer so as to preliminarily evaluate the condition of the environment where the fertilizer is positioned; then, respectively constructing a first high-dimensional vector and a second high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry and the storage safety coefficients of different positions, acquiring local outlier coefficients through the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizers, and combining the storage safety coefficients corresponding to different positions on the basis of considering the position change, wherein the acquired local outlier coefficients are more reliable and accurate; the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehousing time are combined and analyzed to obtain a metamorphic tendency coefficient, the metamorphic tendency coefficient is obtained by combining the first high-dimensional vector, the second high-dimensional vector and the local outlier coefficient, the position change of the fertilizer, the storage safety coefficient of the position and the local outlier coefficient of the fertilizer are comprehensively considered, the obtained metamorphic tendency coefficient of the fertilizer is more persuasive, and the data is more accurate; the edge weight is further obtained through the modification tendency coefficient to obtain the matching pair, a reliable basis is provided for position adjustment of modified fertilizers, the modified fertilizers are obtained by recognition based on a neural network, the recognition effect is better, the efficiency is higher, the storage time of the fertilizers is prolonged due to the matching of the position of the modified fertilizers, the management of the storage bin is more reliable, and the management effect of the organic water-soluble fertilizer storage bin is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for managing an organic water-soluble fertilizer storage bin according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the method and system for managing the organic water-soluble fertilizer storage bin according to the invention in combination with the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a method and a system for managing an organic water-soluble fertilizer storage bin by combining a drawing.
Referring to fig. 1, a flowchart of a method for managing an organic water-soluble fertilizer storage bin according to an embodiment of the invention is shown, and the method includes the following steps:
and S100, collecting humidity data of the positions of the fertilizers in the storage bin, obtaining weight change of the fertilizers in the storage bin, and obtaining storage safety coefficients of the fertilizers based on the weight change of the fertilizers in the storage bin and the humidity data of the positions.
Because the organic water-soluble fertilizer is characterized by being rapidly dissolved in water so as to be better absorbed by crops, when the organic water-soluble fertilizer is stored, the first condition is that the environment of a storage bin is dry, waterproof and moistureproof; the organic water-soluble fertilizer is easy to dissolve in water, is easy to agglomerate or become liquid after being wetted or soaked in water, and is lost, so that the fertilizer efficiency of the organic water-soluble fertilizer is seriously affected, and the humidity condition of each position in a storage bin of the organic water-soluble fertilizer needs to be monitored.
Specifically, in the embodiment of the invention, humidity sensors are arranged in a warehouse to detect humidity data, and the humidity sensors are uniformly distributed in a grid shape, in the embodiment, grid division is performed on a bin area according to a mode of 2 x 2, namely, one grid position comprises 4 stacks of fertilizers, each grid position is provided with one humidity sensor, namely, 4 stacks of fertilizers share the same humidity data, in other embodiments, an implementer can adjust a distribution interval according to actual requirements, the size of grids can be changed, and the sizes among grids can be non-uniform. Because the fertilizers in direct contact with the ground are more easily disturbed by the environment, and the fertilizers are often stacked and stored when being stored, in the embodiment, each stack of fertilizers is taken as an analysis unit, the humidity data of the position of each stack of fertilizers at the current moment can be directly read out by using the humidity sensor, the acquisition interval of the humidity data is set to be 1 minute, and then the humidity data of the position of each stack of fertilizers can be obtained every minute.
Considering that the weight of each stack of fertilizers is basically consistent when the fertilizers are put in storage, if obvious deterioration of the fertilizers does not occur, the weight of the fertilizers is basically unchanged in the storage process, and if the fertilizers are affected by serious moisture, the fertilizers become liquid and flow out, so that obvious difference in weight can occur; therefore, the storage condition of the environment where each stack of fertilizers is located can be reflected through the weight change of each stack of fertilizers.
Obtaining maximum humidity data of the position of the fertilizer in a preset time period and a difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period, and marking the difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period as a maximum weight change value; taking the reciprocal of the maximum value of the weight change of the fertilizer in a preset time period under normal conditions as a correction coefficient, obtaining the product of the maximum value of the weight change corresponding to the fertilizer and the correction coefficient, and selecting the minimum value between the product and a constant 1; taking the difference value between the constant 1 and the minimum value as a molecule, and taking the maximum humidity data of the position where the fertilizer is located as a denominator to obtain a ratio result which is a storage safety coefficient.
Tracking and observing the weight change of each stack of fertilizers, setting a preset time period to be 1 day, and acquiring the storage safety coefficient of the fertilizers within 1 day: determining the weight after the start of 0 hour per day in real time, namely the weight of the fertilizer at each moment in a day; meanwhile, the maximum humidity data in one day is obtained through the humidity sensor, the storage safety coefficient of the fertilizer is obtained through the maximum humidity data and weight change of each stack of fertilizer in one day, and the calculation of the storage safety coefficient is as follows:
Wherein,representing the storage safety coefficient of the current fertilizer; />Maximum humidity data representing the current fertilizer location within a day; />Representing the correction factor, the calculation of the correction factor in this embodiment is +.>，/>The maximum value representing the weight change of the fertilizer in one day under normal conditions is determined empirically by an implementer; />Representing to take the minimum value; />Indicating weight change; />The difference between the maximum weight and the minimum weight within a day is indicated, that is, the maximum value of the weight change within a day.
Maximum weight change of fertilizer in one dayThe larger the value, the more likely the fertilizer is affected by moisture, when the maximum weight change in one day is larger than that in the normal case, namelyWhen the value of (2) is more than 1, the fertilizer is seriously affected by moisture, soTaking a minimum value of 1, and taking a value of 0 corresponding to a molecular item, wherein the obtained storage safety coefficient is zero; the maximum weight change corresponding to the fertilizer in a day +.>The value is smaller than the maximum weight change under normal condition +.>When (I)>The value of (2) is less than 1, the molecular termThe larger the value of the (C) is, the higher the corresponding storage safety coefficient is; meanwhile, when the maximum humidity data of the position of the fertilizer in one day is smaller, the position is drier, and the storage method is suitable for storing the organic water-soluble fertilizer, so that the corresponding storage safety coefficient is higher.
Step S200, obtaining a first high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry; constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions; and obtaining the local outlier coefficient of the corresponding fertilizer based on the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizer.
In the actual storage process, if a slight quality abnormality occurs in the organic fertilizer at one storage point, further loss of the quality of the fertilizer is generally avoided by replacing the position of the fertilizer, or the fertilizer with the abnormal quality is arranged to be timely discharged for timely use, but when the quality abnormality occurs in more time, the use requirement may not exist, so that when the deterioration trend occurs in the organic fertilizer, the position of the organic fertilizer needs to be timely replaced, and further deterioration of the quality of the organic fertilizer is avoided.
Meshing and dividing the bin, and marking the grid position of the fertilizer and the eight neighborhood grid position of the fertilizer as 1 and marking the other grid positions as 0 for any fertilizer; and marking grid positions according to all position changes at the beginning of fertilizer warehouse entry, and sequentially arranging to obtain a first high-dimensional vector. Constructing an initial vector with a preset length, wherein all element values in the initial vector are 0; acquiring worst storage safety coefficients of the fertilizer at each position after the storage of the fertilizer begins, and replacing element values in the initial vector with the worst storage safety coefficients of the fertilizer at each position to obtain a second high-dimensional vector of the fertilizer; wherein, the worst storage safety coefficient of fertilizer corresponding to each position is: and obtaining a difference value between the constant 1 and the minimum storage safety coefficient of the fertilizer at the position, wherein the difference value is the worst storage safety coefficient. For any two stacks of fertilizers, acquiring the Hamming distance between the first high-dimensional vectors corresponding to the two stacks of fertilizers and the cosine distance between the second high-dimensional vectors corresponding to the two stacks of fertilizers, and taking the product of the Hamming distance and the cosine distance as the distance between the two stacks of fertilizers; taking any fertilizer as a target fertilizer, obtaining k nearest adjacent fertilizers of the target fertilizer based on the distance between the target fertilizer and other fertilizers, wherein k is a positive integer; adding k adjacent distances between the target fertilizer and the k nearest adjacent fertilizers corresponding to the target fertilizer to obtain a summation result, and calculating the ratio of the summation result to the maximum distance among the k adjacent distances as the local reachable density of the target fertilizer; and obtaining the average value of the local reachable densities of the k nearest neighboring fertilizers of the target fertilizer, wherein the ratio of the average value of the local reachable densities of the k nearest neighboring fertilizers to the local reachable density of the target fertilizer is the local outlier coefficient of the target fertilizer.
Specifically, the environmental conditions experienced by each stack of fertilizers from warehousing up to now are analyzed, two high-dimensional vectors corresponding to each stack of fertilizers are respectively a first high-dimensional vector and a second high-dimensional vector, wherein the first high-dimensional vector is obtained by position codes of each stack of fertilizers from warehousing up to now, the position codes are carried out according to the grid sequence divided in the step S100, the grid positions of the fertilizers are coded as 1, the eight neighborhood positions of the network positions of the fertilizers are coded as 1, namely the positions of the fertilizers between the self-warehousing and the eight neighborhood positions are coded as 1, the other non-related positions are 0, and the position codes of the positions of each stack of fertilizers from the position of each stack of fertilizers from warehousing up to now are spliced line by line to obtain the first high-dimensional vector, so that the element values in the first high-dimensional vector only comprise 0 and 1; the second high-dimensional vector is obtained based on the storage safety coefficient of the position where the fertilizer is located, an initial vector with a preset length is firstly constructed, each element value in the initial vector is 0, the length of the initial vector is set by the corresponding frequency of moving the fertilizer, namely, the position of each element in the initial vector of the fertilizer corresponds to the change of the position where the fertilizer is located once, in order to facilitate the subsequent comparative analysis, the embodiment of the invention sets the length of the initial vector to be 10, and in other embodiments, the embodiment can be adjusted by an operator; then updating each element value in the initial vector, wherein each element in the initial vector corresponds to the position of the fertilizer once, so that the worst storage safety coefficient of the fertilizer corresponding to the position is used as the element value of the corresponding element in the initial vector, and the worst storage safety coefficient is obtained by subtracting the smallest storage safety coefficient at the current position from a constant 1; and by analogy, updating each element value in the initial vector by using the worst storage safety coefficient to obtain a new vector which is a second high-dimensional vector.
Further, the distance between every two stacks of fertilizers is obtained through the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizers, and the specific calculation of the distance is as follows:
wherein,indicate->Stacking fertilizer and->The distance between the fertilizer stacks; />Indicate->Stacking a first high-dimensional vector corresponding to the fertilizer; />Indicate->Stacking a first high-dimensional vector corresponding to the fertilizer; />Represent the firstStacking a second high-dimensional vector corresponding to the fertilizer; />Indicate->Stacking a second high-dimensional vector corresponding to the fertilizer;indicate->Stacking a first high-dimensional vector and a first high-dimensional vector corresponding to fertilizerThe hamming distance between the first high-dimensional vectors corresponding to the stacked fertilizer is calculated by a known means and is not described in detail; />Indicate->Stacking the second high-dimensional vector corresponding to the fertilizer with the first high-dimensional vector>Second highest corresponding to the stacked fertilizersCosine distances between the dimension vectors.
In this embodiment, before the cosine distance is calculated for the second high-dimensional vectors, the element values in each second high-dimensional vector are arranged from large to small, and because there is no necessary relationship between the position variation trend and the storage safety coefficient variation trend, the arrangement of the elements in the second high-dimensional vectors can eliminate the interference of the position variation, and only the analysis is performed on whether the quality of the fertilizer storage safety coefficient is consistent or not.
The cosine distance between the second high-dimensional vectors is obtained by subtracting the cosine similarity between the two second high-dimensional vectors from a constant 1, and the cosine similarity between the vectors is calculated as a known means and will not be described in detail; if at firstStacking the second high-dimensional vector corresponding to the fertilizer with the first high-dimensional vector>The cosine similarity between the second high-dimensional vectors corresponding to the stacked fertilizer is larger, the +.>Stacking the second high-dimensional vector corresponding to the fertilizer with the first high-dimensional vector>Cosine distance between second high-dimensional vectors corresponding to stacked fertilizers +.>The smaller the same, the +.>Stacking fertilizer and->The smaller the distance between the stacks of fertilizer; correspondingly, the->Stacking the first high-dimensional vector corresponding to the fertilizer and the first high-dimensional vector>Hamming distance between first high-dimensional vectors corresponding to stacked fertilizersThe larger, the description of->Stacking fertilizer and->The more the positions of the stacked fertilizers are not similar, the +.>Stacking fertilizer and->The greater the corresponding distance between the stacks of fertilizer.
And by analogy, obtaining the distance between every two stacks of fertilizers in the storage bin, regarding each stack of fertilizers as one sample, and obtaining an assumed space according to the positions of all samples and the distance between the samples, wherein the assumed space can be referred to an LOF algorithm to obtain a local outlier coefficient corresponding to each sample; the LOF algorithm is an algorithm for identifying abnormal points, is evolved on the basis of the K-NN algorithm, and can identify points with larger differences from other points, namely abnormal points; in the embodiment of the invention, an LOF algorithm can be used for identifying the environmental change condition experienced by each stack of fertilizers, so as to evaluate the deterioration condition of each stack of fertilizers.
Taking any sample in the assumed space as a target sample, firstly, selecting a k-adjacent distance corresponding to the target sample in the assumed space, namely, the distances between k nearest adjacent samples corresponding to the target sample and the target sample, wherein k is a positive integer, and the operator can set the distance to be 10 in the embodiment; then, according to the k-adjacent distance corresponding to the target sample, the local reachable density LRD of the target sample can be correspondingly obtained, namely, the k-adjacent distance of each sample in the k nearest-adjacent samples corresponding to the target sample is added to obtain a summation result, the summation result is divided by the farthest distance in the k-adjacent distances, and the obtained ratio is taken as the local reachable density LRD of the target sample; finally, obtaining local outlier coefficients according to the local reachable density LRDs corresponding to the target sample and the local reachable densities LRDs of k nearest neighbor samples corresponding to the target sample, wherein the local outlier coefficients of the target sample are the ratio of the average value of the local reachable densities LRDs of the k nearest neighbor samples corresponding to the target sample to the local reachable densities LRDs of the target sample; the local outlier coefficient of the target sample is obtained based on the local reachable densities of the target sample and k nearest neighbor samples, and the larger the local outlier coefficient of each stack of fertilizers is, the more abnormal the fertilizer and other fertilizers are, the more likely the deterioration trend is generated.
And step S300, obtaining a deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehousing time of each stack of fertilizers.
The method in the step S200 can obtain local outlier coefficients corresponding to each stack of fertilizers, and simultaneously correspondingly obtain a first high-dimensional vector and a second high-dimensional vector of each stack of fertilizers, so that deterioration tendency coefficients corresponding to the fertilizers can be estimated according to the first high-dimensional vector, the second high-dimensional vector, the local outlier coefficients and the warehouse-in time of each stack of fertilizers, the average value of non-zero elements in the second high-dimensional vector of the fertilizers is calculated, the product result of the average value and the warehouse-in time is calculated, and the warehouse-in time is the total warehouse-in days of the fertilizers; accumulating all elements in the first high-dimensional vector to obtain an accumulated value, and marking the ratio of the product result to the accumulated value as a first ratio; and obtaining a summation result of the local outlier coefficient and the constant 1, and taking the product of the summation result and the first ratio value as a deterioration tendency coefficient of the fertilizer. The concrete calculation method of the deterioration tendency coefficient comprises the following steps:
wherein,indicating the coefficient of deterioration tendency of the fertilizer; />Indicating the total number of days in warehouse entry of the current fertilizer, namely all the days in warehouse entry of the stack of fertilizer; / >Representing the local outlier coefficient of the fertilizer; />A second high-dimensional vector representing fertilizer correspondence;representing the average value of the elements in the second high-dimensional vector corresponding to the current fertilizer, considering the condition that the elements in the second high-dimensional vector are zero, and considering that the average value is more referential, the embodiment only considers the non-zero elements in the second high-dimensional vector when calculating the average value of the elements; />A first high-dimensional vector representing fertilizer correspondence;representing that the first high-dimensional vector corresponding to the current fertilizer is subjected to bit 1 counting, namely accumulating all element values in the first high-dimensional vector to obtain an accumulated value; />Representing a first ratio.
The value obtained by counting the first high-dimensional vector corresponding to the fertilizer by bit 1 reflects the position of the stack of the fertilizer in the bin, and the larger the value is, the more corresponding positions are indicated, the possibility that the stack of the fertilizer is deteriorated due to the fact that the stack of the fertilizer is not in the same position for a long time is reduced, namely, the value of the bit 1 count of the first high-dimensional vector corresponding to the fertilizerThe smaller the fertilizer is, the greater the possibility of deterioration in storage of the fertilizer is, and the larger the corresponding deterioration tendency coefficient is; non-zero elements in the second high-dimensional vector corresponding to the fertilizer stack Average value->The larger the stack of fertilizers is, the larger the worst storage safety coefficient corresponding to the stack of fertilizers is, the larger the possibility of corresponding fertilizers going bad is, and the corresponding deterioration tendency coefficient is larger; correspondingly, when the local outlier coefficient corresponding to the fertilizer is larger, the difference between the stacked fertilizer and other fertilizers is larger, the deterioration tendency is higher, and meanwhile, the longer the organic fertilizer is put in storage, the larger the possibility of deterioration is also higher, so that the longer the organic fertilizer is put in storage, and the larger the local outlier coefficient of the fertilizer is, the larger the deterioration tendency coefficient corresponding to the fertilizer is.
Step S400, acquiring edge weights between two stacks of fertilizers according to deterioration tendency coefficients corresponding to each two stacks of fertilizers, and acquiring matching pairs in all the fertilizers based on the edge weights; and identifying all the spoiled fertilizers according to all the matched pairs and the trained neural network, and adjusting the positions of the spoiled fertilizers based on the matched pairs.
The method in step S300 can obtain the corresponding deterioration tendency coefficient of each stack of fertilizers, if there is no need for a useful fertilizer, when there is a slightly deteriorated fertilizer in the storage bin, further deterioration of the fertilizer needs to be avoided by adjusting the position of the deteriorated fertilizer, and in order to obtain the most suitable adjustment position of each stack of fertilizers, the fertilizers are paired based on the deterioration tendency coefficient of the fertilizer at each storage point; acquiring the absolute value of the difference value of the deterioration tendency coefficients corresponding to each two stacks of fertilizers and the Euclidean distance between the corresponding positions of the two stacks of fertilizers in the storage bin; the ratio of the absolute value of the difference to the Euclidean distance is the edge weight between two stacks of fertilizers; and obtaining the matched fertilizers of each stack of fertilizers according to the edge weight between any two stacks of fertilizers and based on a K-M matching algorithm, wherein each stack of fertilizers and the matched fertilizers form a group of matching pairs.
Firstly, obtaining an edge weight between every two stacks of fertilizers, wherein the edge weight is calculated as follows:
wherein,representing the edge weight between fertilizer A and fertilizer B; />Indicating the deterioration tendency coefficient corresponding to the fertilizer A; />Indicating the deterioration tendency coefficient corresponding to the fertilizer B; />The Euclidean distance between the corresponding storage points of the fertilizer A and the fertilizer B is represented, the Euclidean distance in the embodiment is calculated by grid distribution in a warehouse, the value is a positive number, and the calculation method is a known means and is not repeated; />Representing absolute value calculations.
The absolute value of the difference between the deterioration tendency coefficients corresponding to the fertilizer A and the fertilizer B is reflected, and the larger the value is, the larger the difference between the fertilizer A and the fertilizer B in the deterioration condition is, the larger the edge weight between the corresponding fertilizer A and the fertilizer B is; meanwhile, in the embodiment, the Euclidean distance between the fertilizers is referred to when the edge weight calculation is performed, so as to reduce the difficulty of position adjustment between the fertilizers and improve the efficiency of fertilizer adjustment.
Then, matching is carried out according to the edge weight between every two stacks of fertilizers, a K-M matching algorithm is adopted in the matching method, and in other embodiments, different matching algorithms can be adopted by an implementer; the edge weight obtained by using the deterioration tendency coefficient is matched with a K-M matching algorithm to obtain a plurality of groups of matching pairs, so that the most proper fertilizer adjustment position can be ensured while the moving manpower loss is reduced, so that the storage time of the fertilizer is prolonged as much as possible, namely, the fertilizer with larger deterioration tendency coefficient is matched with the fertilizer with smaller deterioration tendency coefficient, and the Euclidean distance between the positions of the fertilizer is relatively close on the basis of large deterioration tendency coefficient difference, so that the position adjustment is facilitated.
Further, selecting a fertilizer with obvious deterioration and a fertilizer with larger deterioration tendency, determining the fertilizer needing position exchange through a training neural network, wherein the neural network is a twin network, outputting normal fertilizers in all matching pairs through the twin network after training, outputting high-dimensional vectors of each stack of fertilizers, taking the normal fertilizers in all matching pairs as a reference group, and taking abnormal fertilizers in all matching pairs as fertilizers to be discriminated; acquiring cosine distances between the high-dimensional vectors corresponding to all fertilizers in each stack of fertilizers to be discriminated and the fertilizers in the reference group, and constructing a columnar histogram according to all cosine distances between each stack of fertilizers to be discriminated and the fertilizers in the reference group; obtaining corresponding distribution vectors through columnar histograms of each stack of fertilizers to be discriminated, and obtaining L1 distance between the distribution vectors of each two stacks of fertilizers to be discriminated; and obtaining local abnormal factors of each stack of fertilizers to be discriminated on the basis of an LOF algorithm through the L1 distance between each two stacks of fertilizers to be discriminated, and when the local abnormal factors are larger than an abnormal threshold value, the corresponding fertilizers to be discriminated are metamorphic fertilizers.
Specifically, in the embodiment of the invention, the neural network adopts a twin network, and the twin network can automatically measure the dissimilarity of the two high-dimensional characteristics in a comparison learning mode; according to the K-M matching algorithm, all fertilizers can be matched to obtain the matched fertilizers of each fertilizer, each fertilizer and the matched fertilizers corresponding to the fertilizer form a group of matched pairs, multiple groups of matched pairs can be obtained according to the K-M algorithm, and as corresponding side weights exist among the fertilizers in each group of matched pairs, the first 25% of matched pairs with larger side weights are selected as training samples for twin network training in the embodiment of the invention.
Because the selected matching pairs comprise two stacks of fertilizers, the edge weight between the two stacks of fertilizers is larger, the two stacks of fertilizers are shown to be positioned in a storage bin at a relatively short distance, and the corresponding deterioration tendency coefficients of the two stacks of fertilizers are relatively large in difference, namely, the fertilizers with one relatively large deterioration tendency coefficient and the fertilizers with one relatively small deterioration tendency coefficient exist in the matching pairs, the selected fertilizers with relatively large deterioration tendency coefficient in each group of matching pairs are marked as a first type, the fertilizers with relatively small deterioration tendency coefficient in each group of matching pairs are marked as a second type, and the twin networks are trained through the feature vectors corresponding to the fertilizers in the first type and the second type respectively, so that the twin networks can learn the environmental features undergone by each stack of fertilizers from the moment of warehousing, the training process of the twin networks is a known technology means, and the embodiment is not repeated; the characteristic vector of the fertilizer is a vector obtained by directly splicing a first high-dimensional vector and a second high-dimensional vector corresponding to the fertilizer. In the embodiment, the twin network is trained by the matching pair, so that the twin network training can be thinned, the boundaries of most organic fertilizers with deterioration tendency can be divided according to the twin network after the training is finished, and compared with the judgment of fixed proportion, the fertilizer with obvious deterioration tendency or fertilizer with deterioration tendency can be planned and found more flexibly.
The twin network based on the training completion can obtain more normal fertilizers in all matching pairs obtained by the K-M matching algorithm, all normal fertilizers are extracted to serve as a reference group, and the other abnormal fertilizer in the matching pair is recorded as the fertilizer to be distinguished; the twin network outputs a 128-dimensional high-dimensional vector to each sample as a description of the position of the corresponding fertilizer in the high-dimensional space, so that the 128-dimensional high-dimensional vector corresponding to the characteristic vector of each pile of organic fertilizer can be obtained based on the twin network after training, the high-dimensional vector is more effectively mapped in the high-dimensional space by the twin network, the dimension problem and the unbalanced distribution problem of the characteristic vector among different fertilizers are avoided, and meanwhile, the twin network learns how to measure the relative distance between the normal fertilizer and the abnormal fertilizer in the space and maps the relative distance in the high-dimensional space.
Further, the position of the fertilizer to be discriminated is adjusted by referring to the normal fertilizer in the group: firstly, obtaining cosine distances between each stack of fertilizers to be discriminated and each stack of normal fertilizers in a reference group, wherein the cosine distances are obtained by subtracting cosine similarity between high-dimensional vectors corresponding to the fertilizers from a constant 1; then, representing all cosine distances corresponding to the fertilizer to be discriminated and the fertilizer in the reference group by using columnar histograms, and setting the number of columns in the columnar histograms to be 10 in the embodiment of the invention so as to reflect the difference between the fertilizer to be discriminated and the fertilizer in the reference group for the convenience of analysis and calculation; when the cosine distance between a certain stack of fertilizers to be discriminated and all fertilizers in the reference group is large, it may be an organic fertilizer which is obviously deteriorated or has a tendency to deteriorate.
Because each stack of fertilizers to be discriminated corresponds to one columnar histogram distribution, and the number of columns in the columnar histogram distribution is 10, a distribution vector corresponding to the columnar histogram can be constructed according to the value corresponding to each column in the columnar histogram, the value of each element in the distribution vector is the height value corresponding to each column in the columnar histogram, so that the corresponding L1 distance can be calculated based on the distribution vector of each stack of fertilizers to be discriminated, the local abnormal factor of each stack of fertilizers to be discriminated is obtained based on the L1 distance of the distribution vector between different fertilizers, the local abnormal factor is obtained by adopting a traditional LOF algorithm, namely, the L1 distance between each two stacks of fertilizers to be discriminated is used as the distance between two points in a hypothetical space, and the local abnormal factor corresponding to each stack of fertilizers to be discriminated is obtained according to an LOF algorithm, namely, the LOF value at the position; the LOF algorithm is a well-known means, and will not be described in detail herein.
The larger the local abnormal factor of each stack of fertilizers to be discriminated is, the more likely the stacks of fertilizers to be discriminated are abnormal, the fertilizers to be discriminated with the local abnormal factor larger than an abnormal threshold are recorded as metamorphic fertilizers, the empirical value of the abnormal threshold given in the embodiment of the invention is 1.05, and when the local abnormal factor of the fertilizers to be discriminated is larger than the abnormal threshold 1.05, the corresponding fertilizers to be discriminated are metamorphic fertilizers; when the modified fertilizer is produced but the fertilizer is not required, the position of the modified fertilizer needs to be replaced in time so as to prevent the condition of the modified fertilizer from further deteriorating; because the edge weights between every two stacks of fertilizers obtain a plurality of groups of matching pairs, each spoiled fertilizer corresponds to one matching fertilizer, the positions of the spoiled fertilizer and the corresponding matching fertilizer are adjusted and replaced, the two positions are respectively dried during adjustment, and subsequent statistical monitoring is carried out after position adjustment; through the adjustment between rotten fertilizer and the corresponding fertilizer position, can avoid leading to the condition that fixed position fertilizer continuously takes place rotten inefficacy because of the problem of depositing the some position in the feed bin, through continuous alternate position between the fertilizer, also increased the save duration of fertilizer when keeping the dry of fertilizer storage position department, improved the whole storage capacity of feed bin.
In summary, according to the embodiment of the invention, the humidity data of the position of each stack of fertilizer in the storage bin is collected, the weight change of each stack of fertilizer is obtained, and the storage safety coefficient of the fertilizer is obtained based on the weight change of each stack of fertilizer and the humidity data of the position; obtaining a first high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry; constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions; acquiring local outlier coefficients of the fertilizers corresponding to the fertilizers based on the first high-dimensional vectors and the second high-dimensional vectors corresponding to each stack of fertilizers; obtaining a deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehousing time of each stack of fertilizers; acquiring an edge weight according to the deterioration tendency coefficient corresponding to each two stacks of fertilizers, and acquiring matching pairs in all the fertilizers based on the edge weight; all the spoiled fertilizers are identified according to the trained network, and the positions of the spoiled fertilizers are adjusted based on the matched pair, so that the storage time of the amount of the organic water-soluble fertilizer is prolonged, and the storage capacity of the storage bin and the storage bin management effect are improved.
Based on the same inventive concept as the method embodiment, the embodiment of the invention also provides an organic water-soluble fertilizer storage bin management system, which comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor. The steps in the embodiment of the method for managing the organic water-soluble fertilizer storage bin are implemented by the processor when the processor executes the computer program, for example, the steps shown in fig. 1. The method for managing the storage bin of the organic water-soluble fertilizer is described in detail in the above embodiments, and will not be described again.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (6)
1. The management method of the storage bin of the organic water-soluble fertilizer is characterized by comprising the following steps of:
acquiring humidity data of the position of each stack of fertilizer in the storage bin, acquiring weight change of each stack of fertilizer, and acquiring storage safety coefficients of the fertilizer based on the weight change of each stack of fertilizer and the humidity data of the position;
Obtaining a first high-dimensional vector based on the position change of each stack of fertilizers from the warehouse entry; constructing a second high-dimensional vector according to the storage safety coefficients of each stack of fertilizers at different positions; acquiring local outlier coefficients of corresponding fertilizers based on the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizers;
obtaining a deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehousing time of each stack of fertilizers;
acquiring edge weights between two stacks of fertilizers according to the deterioration tendency coefficients corresponding to each two stacks of fertilizers, and acquiring matching pairs in all the fertilizers based on the edge weights; identifying all spoiled fertilizers according to all the matching pairs and the trained neural network, and adjusting the positions of the spoiled fertilizers based on the matching pairs;
the step of obtaining the storage safety coefficient of the fertilizer based on the weight change of each stack of fertilizer and the humidity data of the position, comprises the following steps:
obtaining maximum humidity data of the position of the fertilizer in a preset time period and a difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period, and marking the difference value between the maximum weight and the minimum weight of the fertilizer in the preset time period as a maximum weight change value;
Taking the reciprocal of the maximum value of the weight change of the fertilizer in a preset time period under normal conditions as a correction coefficient, obtaining the product of the maximum value of the weight change corresponding to the fertilizer and the correction coefficient, and selecting the minimum value between the product and a constant 1; taking the difference value between the constant 1 and the minimum value as a molecule, and taking the maximum humidity data of the position where the fertilizer is located as a denominator to obtain a ratio result which is the storage safety coefficient;
the step of obtaining the local outlier coefficient of the corresponding fertilizer based on the first high-dimensional vector and the second high-dimensional vector corresponding to each stack of fertilizer comprises the following steps:
for any two stacks of fertilizers, acquiring a hamming distance between the first high-dimensional vectors corresponding to the two stacks of fertilizers, arranging elements in the second high-dimensional vectors corresponding to each stack of fertilizers in a descending order to obtain a new second high-dimensional vector, acquiring a cosine distance between the new second high-dimensional vectors corresponding to the two stacks of fertilizers, and taking the product of the hamming distance and the cosine distance as the distance between the two stacks of fertilizers;
taking any fertilizer as a target fertilizer, and obtaining k nearest adjacent fertilizers of the target fertilizer based on the distance between the target fertilizer and other fertilizers, wherein k is a positive integer; adding k adjacent distances between the target fertilizer and the k nearest adjacent fertilizers corresponding to the target fertilizer to obtain a summation result, and calculating the ratio of the summation result to the maximum distance among the k adjacent distances as the local reachable density of the target fertilizer;
Obtaining the average value of the local reachable densities of k nearest neighboring fertilizers of the target fertilizer, wherein the ratio of the average value of the local reachable densities of the k nearest neighboring fertilizers to the local reachable density of the target fertilizer is the local outlier coefficient of the target fertilizer;
the step of obtaining the deterioration tendency coefficient based on the local outlier coefficient, the first high-dimensional vector, the second high-dimensional vector and the warehouse-in time of each stack of fertilizers comprises the following steps:
calculating the average value of non-zero elements in the second high-dimensional vector of the fertilizer, and calculating the product result of the average value and the warehousing time, wherein the warehousing time is the total number of days of warehousing of the fertilizer; accumulating all elements in the first high-dimensional vector to obtain an accumulated value, and marking the ratio of the product result to the accumulated value as a first ratio;
obtaining a summation result of the local outlier coefficient and a constant 1, and taking the product of the summation result and the first ratio as a deterioration tendency coefficient of the fertilizer;
the step of obtaining the edge weight between two stacks of fertilizers according to the deterioration tendency coefficients corresponding to each two stacks of fertilizers and obtaining the matching pairs in all the fertilizers based on the edge weight comprises the following steps:
Acquiring the absolute value of the difference value of the deterioration tendency coefficients corresponding to each two stacks of fertilizers and the Euclidean distance between the corresponding positions of the two stacks of fertilizers in the storage bin; the ratio of the absolute value of the difference to the Euclidean distance is the edge weight between two stacks of fertilizers;
and obtaining the matched fertilizers of each stack of fertilizers according to the edge weight between any two stacks of fertilizers and based on a K-M matching algorithm, wherein each stack of fertilizers and the matched fertilizers form a group of matching pairs.
2. The method for managing an organic water-soluble fertilizer storage silo according to claim 1, wherein the step of obtaining the first high-dimensional vector based on the position change of each stack of fertilizers from the beginning of warehousing comprises:
meshing and dividing the bin, and marking the grid position of the fertilizer and the eight neighborhood grid position of the fertilizer as 1 and marking the other grid positions as 0 for any fertilizer; and marking grid positions according to all position changes at the beginning of fertilizer warehouse entry, and sequentially arranging to obtain a first high-dimensional vector.
3. The method of claim 1, wherein the step of constructing a second high-dimensional vector from the storage safety coefficients of each stack of fertilizers at different locations comprises:
Constructing an initial vector with a preset length, wherein all element values in the initial vector are 0;
acquiring worst storage safety coefficients of the fertilizer at each position after the storage of the fertilizer begins, and replacing element values in the initial vector with the worst storage safety coefficients of the fertilizer at each position to obtain a second high-dimensional vector of the fertilizer;
wherein, the worst storage safety coefficient of fertilizer corresponding to each position is: and obtaining a difference value between the constant 1 and the minimum storage safety coefficient of the fertilizer at the position, wherein the difference value is the worst storage safety coefficient.
4. The method for managing a storage silo for organic water-soluble fertilizers according to claim 1, wherein the step of identifying all spoiled fertilizers based on all the matched pairs and the trained neural network comprises:
the neural network is a twin network, the twin network after training outputs normal fertilizers in all matching pairs, and outputs high-dimensional vectors of each stack of fertilizers, the normal fertilizers in all matching pairs are used as reference groups, and the abnormal fertilizers in all matching pairs are fertilizers to be judged;
acquiring cosine distances between each stack of fertilizers to be discriminated and high-dimensional vectors corresponding to all fertilizers in the reference group, and constructing a columnar histogram according to all cosine distances between each stack of fertilizers to be discriminated and the fertilizers in the reference group; obtaining corresponding distribution vectors through the columnar histograms of each stack of fertilizers to be discriminated, and obtaining L1 distance between the distribution vectors of each two stacks of fertilizers to be discriminated;
Obtaining local abnormal factors of each stack of fertilizers to be discriminated through the L1 distance between each two stacks of fertilizers to be discriminated and based on an LOF algorithm, wherein when the local abnormal factors are larger than an abnormal threshold value, the corresponding fertilizers to be discriminated are metamorphic fertilizers.
5. The method of claim 1, wherein the step of adjusting the position of the spoiled fertilizer based on the matched pair comprises:
and on the basis of the matching pairs, obtaining the matched fertilizer of each metamorphic fertilizer, and replacing and adjusting the positions of the metamorphic fertilizer and the matched fertilizer corresponding to the metamorphic fertilizer.
6. A system for managing storage bins for organic water-soluble fertilizers, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, realizes the steps of the method according to any one of the preceding claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311293783.2A CN117035618B (en) | 2023-10-09 | 2023-10-09 | Management method and system for storage bin of organic water-soluble fertilizer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311293783.2A CN117035618B (en) | 2023-10-09 | 2023-10-09 | Management method and system for storage bin of organic water-soluble fertilizer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117035618A CN117035618A (en) | 2023-11-10 |
| CN117035618B true CN117035618B (en) | 2024-01-12 |
Family
ID=88645225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311293783.2A Active CN117035618B (en) | 2023-10-09 | 2023-10-09 | Management method and system for storage bin of organic water-soluble fertilizer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117035618B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5220876A (en) * | 1992-06-22 | 1993-06-22 | Ag-Chem Equipment Co., Inc. | Variable rate application system |
| KR20080096620A (en) * | 2007-03-23 | 2008-10-31 | 진재민 | Refrigerator system |
| KR20140086703A (en) * | 2012-12-28 | 2014-07-08 | 세종대학교산학협력단 | Apparatus and method for detemining putrefaction, refrigeration system and smart building system using the apparatus |
| CN106295974A (en) * | 2016-08-01 | 2017-01-04 | 浙江大学 | A kind of agricultural product quality and safety based on Internet of Things management system |
| CN108154267A (en) * | 2017-12-25 | 2018-06-12 | 深圳春沐源控股有限公司 | The fertilizer parameter regulation means and device of a kind of fertilizer applicator |
| CN108280561A (en) * | 2017-01-06 | 2018-07-13 | 重庆邮电大学 | A kind of discrete manufacture mechanical product quality source tracing method based on comentropy and Weighted distance |
| WO2020012473A1 (en) * | 2018-07-13 | 2020-01-16 | Blau Avi | System and method for composting monitoring and verification |
| CN113095269A (en) * | 2021-04-22 | 2021-07-09 | 云南中烟工业有限责任公司 | Method for judging moisture degree of cigarette blasting bead based on twin neural network |
| KR20220001172A (en) * | 2020-06-29 | 2022-01-05 | 주식회사 꽃팜 | System for managing quality of flower by using temperature and humidity monitoring kit |
| CN215525675U (en) * | 2021-01-26 | 2022-01-14 | 南京汉广生物科技有限公司 | Organic fertilizer humidity detection preprocessing device |
| CN113962638A (en) * | 2021-12-22 | 2022-01-21 | 广州骏天科技有限公司 | Intelligent discount estimation and intelligent promotion method and system |
| CN115034698A (en) * | 2022-05-05 | 2022-09-09 | 汉海信息技术(上海)有限公司 | Fresh product monitoring method and system and server |
| CN116022477A (en) * | 2023-02-03 | 2023-04-28 | 广东米粒农业发展有限公司 | Intelligent storage management system before unhulled rice |
| CN116147278A (en) * | 2023-01-03 | 2023-05-23 | 珠海格力电器股份有限公司 | Method and device for monitoring freshness preservation of food materials in refrigerator and refrigerator |
| CN116659589A (en) * | 2023-07-25 | 2023-08-29 | 澳润(山东)药业有限公司 | Donkey-hide gelatin cake preservation environment monitoring method based on data analysis |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090143923A1 (en) * | 2000-09-08 | 2009-06-04 | Breed David S | Arrangement and Method for Monitoring Shipping Containers |
| US20190387375A1 (en) * | 2018-06-14 | 2019-12-19 | Candibell, Inc. | System and method for food quality monitoring and intelligent restocking |
| BR102019025926A2 (en) * | 2018-12-07 | 2020-06-16 | Imerys Usa, Inc. | ANTIAGLOMERANT AGENT FOR HIGROSCOPIC FERTILIZER |
| US20210144802A1 (en) * | 2019-11-07 | 2021-05-13 | TeleSense, Inc. | Systems and Methods for Advanced Grain Storage and Management Using Predictive Analytics and Anomaly Detection |
-
2023
- 2023-10-09 CN CN202311293783.2A patent/CN117035618B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5220876A (en) * | 1992-06-22 | 1993-06-22 | Ag-Chem Equipment Co., Inc. | Variable rate application system |
| KR20080096620A (en) * | 2007-03-23 | 2008-10-31 | 진재민 | Refrigerator system |
| KR20140086703A (en) * | 2012-12-28 | 2014-07-08 | 세종대학교산학협력단 | Apparatus and method for detemining putrefaction, refrigeration system and smart building system using the apparatus |
| CN106295974A (en) * | 2016-08-01 | 2017-01-04 | 浙江大学 | A kind of agricultural product quality and safety based on Internet of Things management system |
| CN108280561A (en) * | 2017-01-06 | 2018-07-13 | 重庆邮电大学 | A kind of discrete manufacture mechanical product quality source tracing method based on comentropy and Weighted distance |
| CN108154267A (en) * | 2017-12-25 | 2018-06-12 | 深圳春沐源控股有限公司 | The fertilizer parameter regulation means and device of a kind of fertilizer applicator |
| WO2020012473A1 (en) * | 2018-07-13 | 2020-01-16 | Blau Avi | System and method for composting monitoring and verification |
| KR20220001172A (en) * | 2020-06-29 | 2022-01-05 | 주식회사 꽃팜 | System for managing quality of flower by using temperature and humidity monitoring kit |
| CN215525675U (en) * | 2021-01-26 | 2022-01-14 | 南京汉广生物科技有限公司 | Organic fertilizer humidity detection preprocessing device |
| CN113095269A (en) * | 2021-04-22 | 2021-07-09 | 云南中烟工业有限责任公司 | Method for judging moisture degree of cigarette blasting bead based on twin neural network |
| CN113962638A (en) * | 2021-12-22 | 2022-01-21 | 广州骏天科技有限公司 | Intelligent discount estimation and intelligent promotion method and system |
| CN115034698A (en) * | 2022-05-05 | 2022-09-09 | 汉海信息技术(上海)有限公司 | Fresh product monitoring method and system and server |
| CN116147278A (en) * | 2023-01-03 | 2023-05-23 | 珠海格力电器股份有限公司 | Method and device for monitoring freshness preservation of food materials in refrigerator and refrigerator |
| CN116022477A (en) * | 2023-02-03 | 2023-04-28 | 广东米粒农业发展有限公司 | Intelligent storage management system before unhulled rice |
| CN116659589A (en) * | 2023-07-25 | 2023-08-29 | 澳润(山东)药业有限公司 | Donkey-hide gelatin cake preservation environment monitoring method based on data analysis |
Non-Patent Citations (4)
| Title |
|---|
| Predicting soybean seed-germination during warehouse storage;Tekrony DM等;《Seed science and technology》;第21卷(第1期);第127-137页 * |
| 基于电子鼻的鱼粉新鲜度快速检测方法研究;刘辉;《中国优秀硕士学位论文全文数据库 (农业科技辑)》(第4期);第D050-8页 * |
| 测定肥料防结块有效性的一种有创见性的方法;Ph.Haicour等;精细化工中间体(第04期);第46-49页 * |
| 适于餐厅与家庭的叶菜外部品质在线检测与分级系统;魏文松;邢瑶瑶;李永玉;彭彦昆;张文平;;农业工程学报(第05期);第264-273页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117035618A (en) | 2023-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111126662A (en) | Irrigation decision making method, device, server and medium based on big data | |
| CN115994137A (en) | Data management method based on application service system of Internet of things | |
| CN117035618B (en) | Management method and system for storage bin of organic water-soluble fertilizer | |
| KR20190136774A (en) | Prediction system for harvesting time of crop and the method thereof | |
| CN114638536A (en) | Animal health monitoring method and system based on artificial intelligence | |
| CN117147468B (en) | Method and system for detecting anti-nutritional factors of plant source agricultural wastes | |
| CN114077929A (en) | Wind power prediction method and system based on IS-ARIMA-LSTM prediction model | |
| CN113127464A (en) | Agricultural big data environment feature processing method and device and electronic equipment | |
| CN116843085B (en) | Freshwater fish growth monitoring method, device, equipment and storage medium | |
| CN116804668B (en) | Salt iodine content detection data identification method and system | |
| CN116993059A (en) | Internet of things intelligent agricultural plant protection system based on big data | |
| CN115082841B (en) | Method for monitoring abnormity of working area of warehouse logistics robot | |
| Hendrawan et al. | Precision irrigation for Sunagoke moss production using intelligent image analysis | |
| CN114372611A (en) | Water resource optimal allocation method and device | |
| CN116502802A (en) | Data management system based on big data and wireless sensing technology | |
| CN113962638B (en) | Intelligent fruit and vegetable quality detection method and system | |
| CN113702612B (en) | River bank pollution detection and investigation site selection method and system based on artificial intelligence | |
| Ignova et al. | Fermentation seed quality analysis with self‐organising neural networks | |
| CN112232387B (en) | Effective characteristic identification method for disease symptoms of grain crops based on LSELM-RFE | |
| Bhavanandam | Wcp: Weather-based crop yield prediction using machine learning and big data analytics | |
| Vanarase et al. | Crop Prediction Using Data Mining and Machine Learning Techniques | |
| CN117171678B (en) | Soil microbial flora regulation and control method and system in microbial remediation process | |
| AU2021104029A4 (en) | Dryland agriculture and yield gap analysis by machine learning algorithms using iot sensors | |
| Bramley et al. | Yield mapping at different scales to improve fertilizer decision making in the Australian sugar industry | |
| Kumar et al. | Regression-Based Approach for Paddy Crop Assists for Atmospheric Data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |