CN-121982703-A - Corn high light efficiency character recognition method, device, electronic equipment and storage medium

CN121982703ACN 121982703 ACN121982703 ACN 121982703ACN-121982703-A

Abstract

The invention discloses a method and a device for identifying corn high light efficiency characteristics, electronic equipment and a storage medium. The method comprises the steps of extracting three-dimensional phenotype characters from single-plant corn three-dimensional grid models of a plurality of varieties in different growth periods, constructing corn-like three-dimensional scenes of different growth periods consistent with a real planting scene according to the single-plant corn three-dimensional grid models, conducting time-by-time radiation transmission simulation on the corn-like three-dimensional scenes of each growth period by combining radiation parameters of different moments through a calibrated three-dimensional radiation transmission model, determining the daily cumulative photosynthetic effective radiation absorption amount of each canopy of each growth period, selecting a plurality of candidate three-dimensional phenotype characters which are most important for the calculated daily cumulative photosynthetic effective radiation absorption amount of the canopy according to the three-dimensional phenotype characters of the varieties in the whole growth period, and conducting redundant filtration on the candidate three-dimensional phenotype characters in a colinear removing mode to obtain the residual candidate three-dimensional phenotype characters as high-light-efficiency characters of corn.

Inventors

QI NING
YANG GUIJUN
YANG HAO
CHENG JINPENG
YANG XIAODONG
XU BO
LONG HUILING
WANG BAOLIN
BAO JUNWEI

Assignees

北京市农林科学院信息技术研究中心
内蒙古自治区农牧业科学院

Dates

Publication Date: 20260505
Application Date: 20260116

Claims (10)

1. The method for identifying the high light efficiency character of the corn is characterized by comprising the following steps: Acquiring three-dimensional grid models of single-plant corns of a plurality of varieties in different growth periods, and extracting three-dimensional phenotypic characters of preset quantity from the three-dimensional grid models of the single-plant corns; According to the three-dimensional grid model of the single corn, constructing three-dimensional scenes of corn sample squares with different growth periods consistent with the real planting scenes according to the row spacing, plant spacing and plant coordinates of the actual planting in the field; Carrying out time-by-time radiation transmission simulation on corn sample side three-dimensional scenes in each growth period by combining a three-dimensional radiation transmission model calibrated by leaf layering photosynthetic pigment content calibration values and radiation parameters at different moments output by a preset atmospheric correction model, and determining the daily accumulated photosynthetic effective radiation absorption quantity of a canopy; selecting a plurality of candidate three-dimensional phenotypic traits which are most important to the calculated daily cumulative photosynthetic effective radiation absorption of the canopy according to the three-dimensional phenotypic traits of the plurality of varieties in the whole growth period; And redundant filtering is carried out on the candidate three-dimensional phenotypic traits in a non-collinear processing mode, and the rest candidate three-dimensional phenotypic traits are used as high-light-efficiency traits of the corn.
2. The method according to claim 1, wherein the modeling of time-by-time radiation transmission of three-dimensional scenes of corn-like parties in each growth period by combining radiation parameters at different moments output by a preset atmospheric correction model through a three-dimensional radiation transmission model calibrated by leaf layering photosynthetic pigment content calibration values, and determining the daily accumulated photosynthetic effective radiation absorption of a canopy comprises: According to the height layering of corn leaves in the canopy, the content of chlorophyll, chlorophyll a, chlorophyll b and carotenoid in each layer is measured; Taking the measurement result as a leaf layered photosynthetic pigment content calibration value, and substituting the leaf layered photosynthetic pigment content calibration value into the three-dimensional radiation transmission model to adjust the pigment absorption coefficient of the three-dimensional radiation transmission model; Inputting the atmospheric corrected radiation parameters at each moment in the photosynthetic effective period into a calibrated three-dimensional radiation transmission model according to a set time step aiming at a corn-like-side three-dimensional scene in the whole growth period, simulating the incidence, reflection, transmission and scattering processes of radiation in a canopy, and calculating the instantaneous photosynthetic effective radiation absorption quantity of corn plant groups in the corn-like-side three-dimensional virtual scene in the whole growth period at the moment; And accumulating the instantaneous photosynthetic active radiation absorption quantity at all times in the photosynthetic active period of the current day to obtain the crown layer daily accumulated photosynthetic active radiation absorption quantity.
3. The method of claim 1, wherein selecting a plurality of candidate three-dimensional phenotypic traits that are most important for the calculated cumulative photosynthetic effective radiation uptake on canopy days for the three-dimensional phenotypic traits of the plurality of cultivars throughout the growth period comprises: The method comprises the steps of taking the cumulative photosynthetic effective radiation absorption capacity of the canopy as a target quantity, taking three-dimensional phenotypic characters of the whole growth period as independent variables, and determining the influence weight of each three-dimensional phenotypic character on the cumulative photosynthetic effective radiation absorption capacity of the canopy through elastic network regression combined with a recursive feature elimination algorithm; And screening a plurality of candidate three-dimensional phenotypic traits according to the influence weight of each three-dimensional phenotypic trait on the daily accumulated photosynthetic effective radiation absorption of the canopy and a preset weight threshold.
4. The method of claim 1, wherein redundant filtering of the plurality of candidate three-dimensional phenotypic traits by a non-collinear process and taking the remaining candidate three-dimensional phenotypic traits as high light efficiency traits for corn comprises: Filtering the candidate three-dimensional phenotypic traits by a non-collinear mode of iterative elimination of the variance expansion factors until the variance expansion factors of the residual candidate three-dimensional phenotypic traits are smaller than a preset factor threshold; And taking the residual candidate three-dimensional phenotype character after filtering as the high light efficiency character of the corn.
5. The method as recited in claim 1, further comprising: Preprocessing the three-dimensional phenotype character of the whole growth period and the corresponding daily accumulated photosynthetic effective radiation absorption quantity of the canopy, and dividing a training set and a verification set according to a preset proportion; Constructing a first prediction model and a second prediction model based on any algorithm of elastic network regression, random forests and XGBoost; Taking the three-dimensional phenotype character in the training set as the input of the first prediction model, and performing iterative training of the first prediction model; taking the high light efficiency character included in the three-dimensional phenotypic character in the training set example as the input of the second prediction model, and performing iterative training of the second prediction model; taking the three-dimensional phenotype character in the verification set as the input of a first prediction model after training, and outputting a first predicted value of the daily accumulated photosynthetic effective radiation absorption quantity of the canopy; Taking the high light efficiency character included in the three-dimensional phenotypic character in the verification set as the input of a trained second prediction model, and outputting a second predicted value of the daily accumulated photosynthetic effective radiation absorption quantity of the canopy; And determining that the high light efficiency property of the screened corn is accurate in response to the difference value of the first predicted value and the second predicted value meeting a preset condition.
6. The method of claim 1, wherein the obtaining a plurality of individual corn three-dimensional grid models of different growth periods and extracting a predetermined number of three-dimensional phenotypic traits from the individual corn three-dimensional grid models comprises: Scanning individual corns of a plurality of varieties in different growth periods by adopting a three-dimensional scanning technology, collecting three-dimensional coordinate data of plant surfaces, and reconstructing through a three-dimensional software to generate a three-dimensional grid model of the individual corns; and extracting a preset number of three-dimensional phenotypic traits based on the reconstructed three-dimensional grid model through image processing and a geometric calculation algorithm, wherein the three-dimensional phenotypic traits comprise leaf-level traits, hierarchical traits, plant-level traits and population-level traits.
7. The method of claim 1, further comprising, after constructing the three-dimensional scene of the corn-like party at different growth periods consistent with the real planting scene: Four-way tiling treatment is respectively carried out on three-dimensional scenes of corn-like squares in different growth periods so as to eliminate boundary effects.
8. A maize high light efficiency trait identification apparatus, comprising: The character extraction module is used for obtaining three-dimensional grid models of single-plant corns of a plurality of varieties in different growth periods and respectively extracting three-dimensional phenotype characters of preset quantity from the three-dimensional grid models of the single-plant corns; The sampling square construction module is used for constructing corn sampling square three-dimensional scenes with different growth periods consistent with the real planting scenes according to the line spacing, plant spacing and plant coordinates of the field actual planting according to the single-plant corn three-dimensional grid model; The absorption amount calculation module is used for carrying out time-by-time radiation transmission simulation on the corn sample side three-dimensional scene in each growth period by combining the three-dimensional radiation transmission model calibrated by the leaf layering photosynthetic pigment content calibration value and the radiation parameters at different moments output by the preset atmospheric correction model, and determining the daily accumulated photosynthetic effective radiation absorption amount of the canopy; the character primary selection module is used for selecting a plurality of candidate three-dimensional phenotypic characters which are most important to the calculated daily cumulative photosynthetic effective radiation absorption of the canopy according to the three-dimensional phenotypic characters of the plurality of varieties in the whole growth period; And the character final selection module is used for performing redundant filtration on the candidate three-dimensional phenotypic characters in a non-collinear processing mode, and taking the remaining candidate three-dimensional phenotypic characters as high-light-efficiency characters of the corn.
9. An electronic device, comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.
10. A computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, the computer instructions for causing a processor to perform the method of any one of claims 1-7 when executed.

Description

Corn high light efficiency character recognition method, device, electronic equipment and storage medium Technical Field The invention relates to the technical field of agricultural information, in particular to a method and a device for identifying high light efficiency properties of corns, electronic equipment and a storage medium. Background Crop uptake of photosynthetically active radiation (Photosynthetically Active Radiation, PAR) (APAR) is a central quantitative indicator driving photosynthesis and yield formation. In corn breeding fields (breeding cells), differences in canopy structure and physiological characteristics of different genotypes can directly affect population daily cumulative APAR and final yield potential through changes in photon interception, penetration and absorption. At present, the estimation of the daily accumulation APAR of the population mainly depends on two approaches, namely, firstly, the ground point measurement (such as a quantum sensor array) obtains the incidence and absorption at a certain moment, and secondly, the estimation of the average absorption capacity of the canopy based on a remote sensing index or a simplified radiation transmission model. The method has the limitation that three dimensions of point generation surface, instantaneous generation day and average generation are adopted, and key high light efficiency characters of accumulating APAR differences on leading days are difficult to stably screen in a breeding scene. Therefore, a method for identifying the high light efficiency characteristics of corn is needed. Disclosure of Invention The invention provides a corn high light efficiency character recognition method, a device, electronic equipment, a storage medium and a computer program product. According to one aspect of the invention, a method for identifying high light efficiency traits of corn is provided, which comprises the following steps: Acquiring three-dimensional grid models of single-plant corns of a plurality of varieties in different growth periods, and extracting three-dimensional phenotypic characters of preset quantity from the three-dimensional grid models of the single-plant corns; According to the three-dimensional grid model of the single corn, constructing three-dimensional scenes of corn sample squares with different growth periods consistent with the real planting scenes according to the row spacing, plant spacing and plant coordinates of the actual planting in the field; Carrying out time-by-time radiation transmission simulation on corn sample side three-dimensional scenes in each growth period by combining a three-dimensional radiation transmission model calibrated by leaf layering photosynthetic pigment content calibration values and radiation parameters at different moments output by a preset atmospheric correction model, and determining the daily accumulated photosynthetic effective radiation absorption quantity of a canopy; selecting a plurality of candidate three-dimensional phenotypic traits which are most important to the calculated daily cumulative photosynthetic effective radiation absorption of the canopy according to the three-dimensional phenotypic traits of the plurality of varieties in the whole growth period; redundant filtration is carried out on a plurality of candidate three-dimensional phenotypic traits in a non-collinear processing mode, and the rest candidate three-dimensional phenotypic traits are used as high-light-efficiency traits of corn. According to another aspect of the present invention, there is provided a maize high light efficiency trait identification apparatus comprising: The character extraction module is used for obtaining three-dimensional grid models of single-plant corns of a plurality of varieties in different growth periods and respectively extracting three-dimensional phenotype characters of preset quantity from the three-dimensional grid models of the single-plant corns; The sampling square construction module is used for constructing corn sampling square three-dimensional scenes with different growth periods consistent with the real planting scenes according to the line spacing, plant spacing and plant coordinates of the field actual planting according to the single-plant corn three-dimensional grid model; The absorption amount calculation module is used for carrying out time-by-time radiation transmission simulation on the corn sample side three-dimensional scene in each growth period by combining the three-dimensional radiation transmission model calibrated by the leaf layering photosynthetic pigment content calibration value and the radiation parameters at different moments output by the preset atmospheric correction model, and determining the daily accumulated photosynthetic effective radiation absorption amount of the canopy; the character primary selection module is used for selecting a plurality of candidate three-dimensional phenotypic characters which are most important to