| CPC A01G 25/167 (2013.01) [G05B 13/042 (2013.01)] | 5 Claims |
|
1. An intelligent irrigation control method for rice fields based on a cloud service platform, comprising:
periodically monitoring crop planting conditions in each sub-area of an irrigation area, constructing a planting condition data set, constructing a condition coefficient Fps of crop planting based on the planting condition data set, and issuing a warning instruction if the condition coefficient Fps of crop planting exceeds a state threshold;
after receiving the warning instruction, monitoring a current water supply environment of crops and constructing a water supply coefficient Hps, if the water supply coefficient is lower than a water supply threshold, installing an irrigation system in a planting area, and obtaining a crop growth digital twin model after training;
based on current growth data of crops, matching a plurality of irrigation strategies correspondingly from a pre-constructed crop irrigation knowledge graph, and selecting a targeted strategy from the plurality of irrigation strategies through a growth coefficient Bps constructed;
after implementing the target strategy in the irrigation area, periodically obtaining the water supply coefficient Hps and the growth coefficient Bps of crops during an observation period, then constructing an irrigation observation coefficient set, constructing an irrigation coefficient Pes based on the irrigation observation coefficient set, if the irrigation coefficient Pes is lower than an efficiency threshold, issuing an optimization instruction, wherein constructing the irrigation coefficient Pes based on the irrigation observation coefficient set is as follows:
 wherein, k2 and k1 are weights, 0≤k1≤1, 0≤k2≤1, and k1+k2=1, Bpsi is the i-th growth coefficient, Hpsi is the i-th water supply coefficient, i=1, 2, . . . n, n is the number of observation nodes;
receiving an optimization instruction, using a trained strategy optimization model to optimize the target strategy to obtain an optimized target strategy, and controlling the irrigation coefficient based on the optimized target strategy to irrigate crops in the irrigation area;
dividing the irrigation area into a plurality of sub-areas, selecting an installation point within each sub-area, installing a detection unit at each installation point, and establishing a planting condition monitoring system within the irrigation area; dividing crop growth into a plurality of growth cycles, and setting up a plurality of monitoring nodes spaced equally within a current growth cycle of crops; at each monitoring node, monitoring the planting conditions in the sub-area through the planting condition monitoring system, and summarizing collected monitoring data to construct the planting condition data set;
obtaining the rainfall Rv, the temperature Rt, and the sunshine duration Rg in the irrigation area at each monitoring node from the planting condition data set, performing linear normalization on the above three, mapping the corresponding data values to an interval [0,1], and constructing the condition coefficient Fps of crop planting according to the following formula:
 weight coefficients: 0≤β≤1, 0≤α≤1, 0≤γ≤1, and α+β+γ=1, i=1, 2, . . . , k, k is the number of monitoring nodes, Rvavg is an average value of rainfall, Rv is a qualified standard value for rainfall, Rtavg is an average value of temperature, Rt is the qualified standard value for temperature, Rgavg is an average value of sunshine duration, and Rg is a qualified standard value for sunshine duration;
setting up a detection point in each sub-area to monitor the air humidity and the soil moisture content in each sub-area, and summarizing the monitoring data to construct a water supply condition set; constructing the water supply coefficient Hps from the water supply condition set is as follows: performing linear normalization on the air humidity Rp and the soil moisture content Tp, and mapping the corresponding data values to an interval [0, 1] according to the following formula:
 weight coefficient: 0≤ρ≤1, 0≤ζ≤1; if the water supply coefficient Hps is lower than the water supply threshold, installing the irrigation system in the planting area;
after determining a type and a current growth stage of crops, taking crop irrigation and related keywords as search terms to construct a graph data set through deep search, and generating a crop irrigation knowledge graph after completing the entity relationship construction;
performing feature recognition on crops data and planting status data of crops to obtain data features correspondingly; after determining the irrigation requirements of crops, based on a correspondence between data features and irrigation strategies of crops, using a trained matching model to match a plurality of irrigation strategies correspondingly from the crop irrigation knowledge graph;
obtaining rainfall prediction information and combining it with the irrigation strategies obtained as an input, using the crop growth digital twin model to predict the growth status of crops, obtaining predicted growth data correspondingly, summarizing the predicted growth data to construct a predicted data set; constructing the growth coefficient Bps of crops based on the predicted data set, marking each of irrigation strategies based on the growth coefficient Bps obtained, and taking an irrigation strategy with the highest growth coefficient Bps as the target strategy, an irrigation strategy with a second-highest growth coefficient Bps as a backup strategy.
|