231 results about "Rice growth" patented technology

The invention discloses a rice leaf tip small target detection and counting method based on global information enhancement, and the method comprises the steps: firstly building a field rice leaf tip image data set as a basis, and constructing an improved YOLO-DCL model; the core innovation of the method lies in that a global context information enhancement module, an inverted residual-cascade grouping attention module and a re-parameterization shared convolution detection head are introduced, key feature expression is enhanced and detection efficiency is optimized through effective fusion of global information, the recognition precision and processing speed of the model on a tiny rice leaf tip target are significantly improved, and the method is suitable for large-scale popularization and application. After training parameters are set, an optimization model is trained, then system performance verification and module optimization are carried out on the model, and a version with the optimal rice leaf tip detection performance is screened out. The technical bottlenecks that small target leaf tips are prone to missing detection and false detection and insufficient in real-time performance under the complex farmland background are effectively overcome, and reliable and efficient technical support is provided for rice growth early-stage yield prediction.

The invention discloses a fertilization recommendation method combined with a rice seedling raising fertilizer carrying technology in the technical field of agricultural production. The fertilization recommendation method comprises the following steps: S1, acquiring geology and seedling raising states; s2, presetting rice growth requirements; s3, calculating to obtain an additional nutrient amount; s4, dividing the growth cycle of the rice; s5, recommending the fertilization amount according to stages; according to the method, nutrient supply of seedling raising fertilizer is brought into a whole-growth-period fertilization plan, and accurate matching is achieved; on the premise that the yield is not affected, the fertilizing amount in the early and whole growth periods of rice is remarkably reduced, and the fertilizer utilization rate is increased; through dynamic matching of the early-stage fertilizer loading and supplying amount and the middle and later-stage fertilizer needing amount, resource waste caused by excessive fertilization is avoided, and the fertilizer cost is reduced; the nutrient proportion is scientifically distributed according to the nutrient absorption efficiency of each stage, and the fertilizer utilization rate is increased; the final yield can be improved; soil fertility indexes and historical data are combined, the profit and loss trend of soil nutrients is recognized, and cyclic utilization of the soil nutrients is achieved through dynamic matching of the sustained and controlled release fertilizer and the organic fertilizer.

The invention relates to the field of artificial intelligence, in particular to a rice growth vigor dynamic monitoring method and system based on modal fusion and deep learning. The method comprises the following steps: for a plurality of target regions in the same administrative division, acquiring multi-dimensional agricultural data related to rice growth vigor in each target region; performing space-time fusion on the multi-dimensional agricultural data of the plurality of target areas to obtain a linkage monitoring map; performing rice growth prediction on each target area in the linkage monitoring map through a DSW Transform model to obtain an independent growth prediction result of each target area; a correlation weight is dynamically calculated through a deep learning model of a cross-modal attention mechanism in combination with geographic data, meteorological data and a linkage monitoring map, and rice growth prediction is performed on a rice block formed by linkage of each target area based on the correlation weight to obtain a comprehensive growth prediction result; and pushing rice growth information in each target area to the user according to the independent growth prediction result and the comprehensive growth prediction result.

The invention provides a rice variable rate fertilization method and device, electronic equipment and a storage medium, and relates to the technical field of agricultural intelligent information process.The method comprises the steps that non-rice plant interferents in a multispectral remote sensing image are removed, and a rice plant canopy mask is generated; based on the rice plant canopy mask, multi-dimensional features of the rice plant canopy are extracted and fused, and a fused growth feature vector is obtained; grading the rice growth vigor according to the fused growth vigor feature vector through a growth vigor grade prediction model to obtain a predicted growth vigor grade; performing standardization processing on the predicted growth vigor level to obtain a standardized growth vigor level; determining a nitrogen fertilizer adjustment factor based on the standardized growth vigor grade; and according to the nitrogen fertilizer adjustment factor, adjusting a standard nitrogen application amount to obtain a recommended nitrogen application amount. According to the invention, accurate recommendation of the nitrogen application amount in the rice planting process can be realized, so that accurate fertilization is realized.

The invention discloses a microbial agent and application thereof in preventing and treating rice blast and promoting rice growth, active ingredients of the microbial agent are formed by compounding cold bacillus with the preservation number of CGMCC (China General Microbiological Culture Collection Center) NO: 33492 and bacillus velezensis with the preservation number of CGMCC NO: 33493, and the microbial agent aims at solving the synergistic problem of preventing and treating rice blast and promoting rice growth. According to the complex microbial inoculant, through the synergistic interaction between the two strains, the metabolite of the complex microbial inoculant can completely inhibit the spore germination of magnaporthe oryzae; meanwhile, the fungicide can degrade cellulose, obviously promote the growth of plant height, fresh weight and dry weight of rice, and realize the integration of disease prevention and growth promotion. The invention further discloses a preparation method of the fungicide and an application method of the fungicide for preventing and treating rice blast and promoting growth in a root irrigation mode in the three-leaf stage of rice. The microbial agent disclosed by the invention has dual functions of efficiently preventing diseases and stably promoting growth, and an efficient and safe biological solution is provided for green and sustainable development of agriculture.

The invention discloses a rice growth period distinguishing method and an intelligent irrigation control device. Belongs to the technical field of crossing of intelligent agriculture and irrigation and water conservancy engineering and particularly relates to the technical field of rice growth period judgment and intelligent irrigation. The technical problem that an accurate and effective rice growth period judgment method for guiding intelligent rice irrigation is lacked at present is solved. The method specifically comprises the following steps: calculating a rice growth process index through an improved accumulated temperature model; and constructing a multi-factor coupling growth period judgment method, and judging the current rice growth period through the rice growth process index. The system comprises a sensing unit, a judgment and decision-making unit, an electric execution unit and an input and power supply unit, the judgment and decision-making unit receives the environment data of the rice field and the water layer depth data of the rice field, judges the current growth period of the rice according to a rice growth period judgment method, calculates the current water layer deviation and generates a control instruction to the electric execution unit.

The invention relates to the technical field of agricultural intelligent prediction, and discloses a fertilization decision-making system fusing soil and rice growth vigor information data acquisition and analysis, which is characterized in that a microcontroller controls an irrigation valve to apply an irrigation pulse with a preset duration, and a soil humidity sensor is used to monitor a time interval when moisture falls back from a peak value to a baseline; according to the method, irrigation behaviors are reconstructed into an active diagnostic tool, traditional chemical index detection is replaced by physical process dynamic monitoring, lag adjustment is converted into a pre-prediction mode, meanwhile, acoustic verification and a differential correction mechanism are combined, and the crop transpiration rate is predicted. And an anti-environmental interference accurate decision-making system is constructed on the basis of low-cost hardware.

The invention relates to a rice character parameter inversion method combining an optical-microwave collaborative model and a variational auto-encoder. The method comprises the following steps: acquiring multiband optical reflectivity, a microwave backscattering coefficient and observation geometric information of a research area; inputting the data into a multi-mode encoder, and encoding the data into probability distribution of a potential space; an optical and microwave physical model is used as a differentiable decoder, potential variables sampled from a potential space are received, and remote sensing data are simulated and generated; a self-supervised learning normal form is adopted, and end-to-end training is carried out on the whole frame by minimizing a composite loss function; a trained encoder is used as a rapid inversion model, vegetation biological physicochemical parameters are extracted from new remote sensing data, rapid inversion of the parameters is achieved, and uncertainty of the parameters is quantified. The rapid and accurate vegetation character parameter inversion method is established by fusing multi-modal data, a physical model and a deep learning technology, important reference is provided for rice growth monitoring and precision agriculture, and the method has popularization prospects.

The invention discloses a rice whole growth period growth dynamic monitoring system based on multi-source sensing fusion, and relates to the field of agricultural intellectualization, and the system comprises an acquisition module which is used for taking a rice root development core area as a target, adaptively arranging a multi-dimensional sensing array and dynamically adjusting the operation frequency of the array, continuously collecting root system microenvironment and morphological characteristic original data; the preprocessing module is used for receiving the original data acquired by the acquisition module, performing space-time alignment and noise filtering on the root system microenvironment and morphological feature sensing data, and synchronously extracting core data dimensions for driving development analysis through a feature enhancement algorithm; by comprehensively evaluating the generated growth potential and yield potential report, the rice growth state is comprehensively reflected, a planter is assisted to intervene abnormal conditions in advance, the growth stress loss is reduced, and a scientific basis is provided for accurate regulation and control, so that the fine management level and yield potential of rice planting are improved.

According to the application of the nano silicon dioxide in relieving rice temperature rise stress, the nano silicon dioxide has a relieving effect on rice growth damage caused by high environment temperature. According to the method, the nano silicon dioxide is used for relieving the temperature rise stress of rice, a new method is provided for crops to cope with the temperature rise stress, the yield loss caused by the temperature rise stress can be remarkably reversed by spraying the nano silicon dioxide on the leaf surfaces in the tillering stage and the jointing stage, and experiments prove that the yield is improved by 34% compared with that of an untreated high-temperature group.

The method comprises the following steps: monitoring seedling condition and growth vigor information in a rice growth process, and constructing a seedling condition index spatial distribution diagram; according to the seedling condition index spatial distribution map and a preset rice suitable growth index dynamic curve, generating a topdressing application amount operation prescription map; determining a navigation path diagram of the rice agricultural unmanned aerial vehicle; and fusing the topdressing application amount operation prescription map with the navigation path map, and carrying out topdressing operation through a rice agricultural unmanned aerial vehicle based on a fusion result. Tedious steps such as hyperspectral unmixing and dimensionality reduction modeling are not needed, and the decision-making efficiency is improved. The fertilizer demand is accurately matched, and excessive fertilization is avoided; a navigation path is fused with a prescription map, variable spraying with finer granularity is achieved, rasterization is not needed to restrain a minimum operation unit, and spraying uniformity is improved.

The invention discloses a chloroplast RNA edited gene OsPPR7 for temperature-sensitive regulation and control of rice and application of the gene OsPPR7, and belongs to the technical field of gene engineering. The CDS sequence of the OsPPR7 gene is as shown in SEQ ID NO. 1. The OsPPR7 is an RNA (Ribonucleic Acid) editing factor which has a bidirectional regulation and control function and is sensitive to temperature. Editing of multiple sites on the rpoB gene is negatively regulated and controlled, and meanwhile RNA editing of multiple genes such as ndhB and ndhD which are crucial to the chloroplast function is positively regulated and controlled under high-temperature stress. Under the stress of high temperature of 35 DEG C, the loss of functions can lead to obvious reduction of RNA editing efficiency of a plurality of ndh genes, and serious yellowing of seedlings is caused. The key function of the rice OsPPR7 gene in regulation and control of rice growth and development and high-temperature response is clarified for the first time, the importance of the OsPPR7 gene in maintenance of chloroplast functions under high-temperature stress is clarified, and a new gene resource and a technical path are provided for cultivation of new varieties of high-temperature-resistant rice.

The invention provides a double-cropping rice field annual rotation fertility dynamic monitoring system based on the Internet of Things, and relates to the technical field of fertility monitoring, and the system comprises a sensing collection module which is used for collecting soil fertility under the annual scale, and synchronously recording double-cropping rice growth indexes and rotation management information; the data processing module is used for preprocessing the multi-dimensional original monitoring data flow and generating a structured data packet; the trend prediction module is used for predicting fertility change through a Prophet algorithm; and the scheme optimization module is used for constructing a fertility regulation and control model and generating an optimized fertility regulation and control scheme. According to the method, the Prophet algorithm is used for feature extraction and model training, so that accurate pre-judgment and early warning of fertility change are realized; the fertility regulation and control rule base is constructed based on the prediction result and the rotation tillage characteristics, personalized regulation and control schemes are generated according to the risk levels of different field parcels, the scheme effect is optimized through historical scene verification, and the pertinence and the implementation effect of fertility regulation and control are effectively improved.

The present invention relates to the technical field of rice yield prediction. Disclosed is a rice yield prediction method based on unmanned aerial vehicle remote sensing image processing. In the method, vegetation indices (VIs), spectral indices, and color indices of rice at different growth stages are collected by an unmanned aerial vehicle, and the rice yield is predicted by using single-date VIs and multi-date VIs based on multispectral (MS) and digital (RGB) images. Both MS and RGB digital sensors on an unmanned aerial vehicle are reliable platforms for rice growth and yield prediction, and it is feasible to estimate the rice yield on the basis of dual-sensor data and temporal accumulation across different growth stages, so that full-growth-stage monitoring of rice can be performed, the influence of spectral indices and color indices at different growth stages on the rice yield can be evaluated, and the pre-harvest grain yield can be predicted in a timely and accurate manner, thereby facilitating rice production management and grain trade.

The utility model relates to the technical field of rice planting sowing, and discloses a telescopic quantitative sowing mechanism for rice planting field fertilization, which comprises a chassis, the left side of the top of the chassis is fixedly connected with a fixed support, the top end of the right side of the fixed support is fixedly connected with a motor II, and the output end of the motor II is fixedly connected with a transmission wheel. The outer walls of the two transmission wheels are in transmission connection through a transmission belt, a second sliding groove is formed in the right side of the outer wall of the fixing disc, the inner wall of the second sliding groove is in sliding connection with the outer wall of an adjusting handle, and a telescopic mechanism is arranged at the bottom of the bottom disc and used for adjusting a discharging port. According to the adjustable quantitative feeding device, the hole II and the hole I form a leakage hole under the rotation action of the adjusting handle, so that the falling amount of a fertilizer is adjusted, and the requirement of rice growth on overall nutrients can be better met through the adjustable quantitative feeding device according to specific operation requirements.

The invention relates to the technical field of agricultural biology, in particular to a rational design method of a rice growth-promoting synthetic microbial community based on collaborative optimization of directed evolution and a metabolism model. The method comprises the following steps: continuously subculturing a multi-source soil microbial community to be stable, co-culturing the multi-source soil microbial community and rice seeds, and selecting an optimal community based on a growth phenotype; the method comprises the following steps: selecting an optimal community, applying a bottleneck to the optimal community, externally adding flora disturbance, constructing a filial generation community, repeating'screening-disturbance-subculture 'for five rounds to obtain a growth promoting function enhanced directed evolution synthesis microbial community, dividing the microbial community into a plurality of modules, and constructing a genome scale metabolism model to simulate the influence of interaction between different modules on rice growth. And combining the directed evolution synthesized microbial communities in pairs according to different proportions so as to obtain rational design synthesized microbial communities with better growth promoting effects. According to the method, the problem of model prediction deviation caused by lack of a host-microorganism interaction mechanism in an existing method is solved.

The invention relates to the technical field of agricultural pest control, and particularly discloses a rice pest infection process inversion device, which comprises a base, crawler belt bodies are arranged on both sides of the bottom end of the base, and through high-frequency acquisition and external data import of an ASD spectrometer, in combination with labeling and data enhancement processing of a spectrum curve, the inversion result of the rice pest infection process is obtained. The comprehensive coverage rice blast database is constructed, deep learning training is carried out based on the comprehensive database, the formed rice blast diagnosis model can accurately capture subtle changes of substances in rice on a spectrum curve, and accurate judgment of the rice blast infection degree is achieved. In addition, by means of the rice growth whole process image reconstruction technology, the WOFOST growth model and the hyperspectral reconstruction technology, the spectral information change curve of the sensitive wave band of the rice in the whole phenological period can be inverted, the rice blast infection process can be visually presented, and the method can be used for identifying the rice blast early-stage recessive characteristics. And a scientific decision basis is provided for pest control.

The invention discloses a biochar rice field water carbon and nitrogen cycle process simulation method based on a DNDC model, which comprises the following steps: firstly obtaining basic data of rice growth and test monitoring data of a rice ecosystem under biochar treatment, and then constructing a biochar organic fertilizer module to improve the DNDC model; the method comprises the following steps: establishing a Biochar-DNDC model according to basic data of rice growth and test monitoring data, then calibrating and verifying the Biochar-DNDC model by adopting the test monitoring data, and finally simulating the water carbon and nitrogen cycle process of the biochar rice field by adopting the verified Biochar-DNDC model to obtain the water carbon and nitrogen cycle process of a rice field ecosystem under the condition that biochar is applied to a target area. The simulation accuracy of the DNDC model on the charcoal rice field water carbon nitrogen cycle process is improved, and the simulation result of the charcoal rice ecosystem carbon sequestration emission reduction synergy potential is more real and accurate.

This invention discloses a cultivation and monitoring device for low-temperature cold-resistant growth of rice, relating to the field of rice cultivation technology. The invention includes a three-layer frame for mounting the equipment; a supplementary lighting system is installed on the upper layer of the frame; a rice cultivation trough is installed on the middle layer; and a water tank and a chiller are installed on the lower layer. The water tank and the chiller's inlet are connected via a first pipe. The rice cultivation trough is internally divided into a cold water inlet chamber, a cold water cultivation chamber, and a cold water return chamber. Water level and flow are controlled by inlet and outlet weirs. The device is also equipped with adjustable weirs to flexibly adjust the water level to meet the needs of different rice growth stages. Simultaneously, a video recording device and a temperature recording device are installed to record the entire rice growth process and monitor water temperature in real time, providing researchers with accurate experimental data. Furthermore, the device includes a room-temperature cultivation chamber as a control, which helps evaluate the low-temperature growth ability of genetically modified or hybridized rice seeds.

The invention relates to the technical field of soil remediation, in particular to an iron-modified cow dung biochar material and a preparation method thereof and application in remediation of cadmium-copper contaminated soil. Cow dung is used as a raw material, pretreated cow dung is obtained through natural airing, dewatering and smashing, then the pretreated cow dung and a ferrous chloride solution are mixed, stirred, filtered, dried and ground, and the iron-modified cow dung biochar material is obtained. And pyrolyzing under the protection of inert gas. The iron-modified biochar has a porous layered structure, rich oxygen-containing functional groups and iron-based active sites, has a pH value close to neutrality, and can effectively adapt to the farmland soil environment. The material can remarkably reduce the effective state content of cadmium and copper in soil, promote conversion of heavy metal to a stable residue state, reduce migration of heavy metal in pore water and improve the enzyme activity and nutrient effectiveness of the soil at the same time, the grain cadmium accumulation reduction amplitude reaches 71.65% and the copper accumulation reduction amplitude reaches 29.36% in the rice growth process, and therefore the cadmium and copper combined polluted soil is efficiently repaired, and the quality of the rice is improved. And crop safety and soil health are guaranteed.

The application discloses a kind of based on soil compactness machine harvesting regenerated rice loss reduction stable yield method, belong to rice cultivation technical field, its method includes following steps, S1, the soil environment of different gradient soil compactness is constructed: before the first season rice harvest 15~17d, 12~14d, 9~11d, 6~8d, 3~5d respectively to target field block is drained field water treatment, while setting the control field block of first season rice growth period flooding irrigation, by different drainage timing control paddy field soil moisture content, the machine harvesting operating environment of different soil compactness is constructed;S2, determination soil compactness and soil moisture content;S3, mechanical harvesting operation;S4, yield and loss rate measurement;S5, quantitative relationship model construction;S6, soil compactness classification and water management strategy formulation.The application first takes soil compactness as core index, establishes its quantitative relationship model with soil moisture content, machine harvesting regenerated rice single yield of regeneration season, yield loss rate, realizes the quantitative, precision management of machine harvesting regenerated rice paddy field operating environment soil moisture.

The present invention relates to a Single Nucleotide Polymorphism (SNP) marker for determining the degree of protein content of rice seeds and a method for determining the degree of protein content of rice seeds using said marker. Specifically, the present invention relates to an SNP marker for determining protein content and a method for determining protein content using said marker, which constructs a machine learning-based rice protein content prediction model by utilizing SNP markers associated with high and low protein content, thereby enabling the prediction of rice seeds with high protein content in the early stages of growth.

The invention discloses application of bacillus megaterium and a biological fertilizer thereof in rice planting, and relates to the technical field of microorganisms. The bacillus megaterium is preserved in the China General Microbiological Culture Collection Center on July 10, 2025, and the preservation number of the bacillus megaterium is CGMCC (China General Microbiological Culture Collection Center) NO.35179. The bacillus megaterium has the advantages that the bacillus megaterium is preserved in the China General Microbiological Culture Collection Center on July 10, 2025; the strain is a plant growth-promoting bacterium and has a good rice growth-promoting effect, and when the prepared biological fertilizer is applied to the rice planting process, compared with a control group, the tiller number of rice is increased by 64.44%, the plant height is increased by 2.77%, and the leaf chlorophyll content is increased by 8.38%. The bacillus megatherium provided by the invention and the biological fertilizer prepared by strain fermentation have the advantages of greenness, safety, high efficiency and the like, and have a better application prospect in the aspect of agricultural production.

The invention discloses a construction method of an unmanned aerial vehicle multi-spectral remote sensing basic model based on agricultural application, and relates to the technical field of intelligent agricultural equipment, and the method comprises the following steps: S1, collecting a multi-band remote sensing image of a rice region through a multi-spectral camera carried by an unmanned aerial vehicle; s2, calculating an anti-flooding vegetation index, and generating a vegetation coverage rate initial value according to the anti-flooding vegetation index; s3, extracting a rice plant area image, and inputting the rice plant area image into the trained rice state recognition model; s4, judging whether the flooding stress probability is greater than a flooding threshold value or not; and S5, determining the rice growth potential grade based on the vegetation coverage rate initial value or the corrected vegetation coverage rate. According to the method, through the anti-flooding vegetation index and the self-adaptive calculation algorithm in the turbid water body, the strong absorption interference of the water body in the near-infrared band is effectively stripped, so that the initial value of the vegetation coverage rate can truly reflect the survival state of underwater vegetation, and the situation that part of flooded healthy rice is misjudged as poor growth or death is fundamentally avoided.

The invention discloses a rice waterlogging disaster dynamic monitoring model construction method and device, equipment and a medium. The method comprises the following steps: acquiring historical waterlogging disaster situation, rainfall, digital elevation and rice growth period data of a target area; calculating an effective cumulative rainfall capacity and a continuous rainfall day number index based on rainfall data, calculating a gradient index based on elevation data, performing normalization, and constructing a sample data set with a disaster grade; inputting the sample set into a random forest model, and calculating a feature weight to generate a comprehensive index calculation model; grouping the samples in combination with the growth period data, and constructing each growth period sample subset; performing normality test on the sample subsets, and determining grading thresholds of different disaster grades in each growth period by using a t-distribution interval estimation method; and associating the calculation model with the grading threshold value to form a model capable of directly and dynamically monitoring the rice waterlogging disaster. According to the embodiment of the invention, the model for monitoring the rice waterlogging disaster is provided, the disaster grade can be quickly output, and a basis is provided for timely drainage and disaster reduction.

The invention discloses a growth-promoting methylobacterium R5018 strain and an application of the growth-promoting methylobacterium R5018 strain. The strain R5018 is obtained by separation from a rice phyllospheric microflora and methanol domestication and screening, and is preserved in the China Center for Type Culture Collection (CCTCC), and the preservation number is CCTCC NO: M20252544. The rice growth-promoting microbial inoculum prepared from the strain can be stably colonized at phyllosphere and promote the growth of overground and underground parts of rice through seed soaking, foliage spraying or combined application of the seed soaking and the foliage spraying, and the ear number, the grain number and the yield are remarkably increased. Compared with the prior art, the strain disclosed by the invention is derived from a rice phyllospheric bacterial community, is high in adaptability, environment-friendly and efficient, provides a new microbial technical scheme for high-yield cultivation of rice, and has a wide popularization and application prospect.

The invention discloses a method and a system for detecting rice bacterial leaf blight based on unmanned aerial vehicle spatial spectrum fusion. The method comprises the following steps: firstly, constructing a rice bacterial leaf blight detection data set; a rice bacterial leaf blight severity grading detection double-branch model is constructed, and the model adopts a spectrum-space double-branch parallel coding and fusion framework and comprises a spectrum branch module, a space branch module and a cross attention fusion mechanism module. And inputting the complete image of the rice field into the trained rice bacterial leaf blight severity grading detection double-branch model, and obtaining severity grading detection results of the bacterial leaf blight in different rice growth periods of the target field. According to the invention, deep complementation and dynamic information balance of spectrum and spatial features of the multispectral unmanned aerial vehicle are realized, so that the accuracy of disease detection and the generalization ability of the model are effectively improved.

The invention discloses an iron-phosphorus biochar-based fertilizer suitable for saline-alkali soil, which has a good nutrient slow-release characteristic, and also provides a preparation method and an application approach of the fertilizer. According to the product, biomass which is wide in source and low in cost is taken as a basic raw material, and nutrient elements such as nitrogen, phosphorus and potassium are successfully converted into various forms with a slow-release function by introducing iron-phosphorus composite modification and combining a high-temperature pyrolysis technology, so that the utilization efficiency and the lasting supply capacity of nutrients are greatly improved. The prepared biochar-based fertilizer can effectively optimize the physical and chemical environment of saline-alkali soil, for example, increase soil organic matters and enhance cation exchange capacity, also can relieve adverse effects of salt stress on rice growth, and supports healthy development of saline-alkali tolerant rice. The method is simple and convenient in process, high in environmental compatibility and remarkable in economic benefit, and has wide popularization value and application potential in the aspects of saline-alkali soil improvement and crop safety production.

This invention relates to a method for rice growth status perception and agricultural decision-making based on multimodal semantic guidance, belonging to the fields of smart agriculture, computer vision, and artificial intelligence. The method includes acquiring rice canopy images; constructing a dataset; utilizing a small number of samples labeled by agricultural experts and their corresponding descriptive semantic information; introducing a multimodal semantic-guided auxiliary labeling mechanism to generate growth status labels for the remaining samples, thereby reducing data labeling costs and improving labeling consistency; based on the labeled dataset, constructing a multi-task recognition model to collaboratively identify multi-dimensional growth states of rice, such as phenological stage, nutritional status, disease type, disease impact degree, and population growth, within a unified feature space. Based on the multi-task recognition results, rice growth status analysis information is generated, forming agricultural guidance strategies such as fertilization management and pest and disease control, which are then distributed to field execution equipment, realizing an intelligent management closed loop combining rice growth status perception and agricultural decision-making.