64 results about "Plant phenotyping" patented technology

The invention provides a plant phenotype platform three-dimensional imaging method and device based on multi-sensor fusion, and the method comprises the steps: collecting the original three-dimensional point cloud data of a plant through a plurality of heterogeneous imaging sensors carried on a plant phenotype platform, and carrying out the time synchronization processing and space registration processing. And dynamically calculating the confidence coefficient weight of each data point according to the environment illumination condition, the measurement attribute of each data point in the registration point cloud set and the characteristics of the imaging sensor to which the data point belongs, and obtaining a point cloud set with a point-level confidence coefficient weight. Performing adaptive weighted fusion processing according to the spatial distribution of the point cloud set to generate a preliminary fusion point cloud model; and according to a quality index of the model and a real-time environment illumination condition, dynamically adjusting and calculating a weight parameter of a confidence coefficient weight, and finally outputting an optimized three-dimensional point cloud model after iterative optimization. The defect that a fusion result of an existing plant three-dimensional phenotype accurate acquisition scheme is insufficient in stability and precision in a complex and changeable field environment is overcome.

The invention discloses an infrared corn drought identification method based on wavelet boundary enhancement, and belongs to the field of agricultural informatization and plant phenotype identification, and the method comprises the following steps: S1, obtaining an infrared image of a corn plant; s2, executing two-dimensional discrete wavelet transform to generate a boundary response diagram; s3, generating a binary mask according to the characteristics of the boundary response diagram to the blade edge; s4, connected domain analysis is executed, candidate leaves are obtained, and independent single-leaf masks are generated; s5, the small holes are removed, a mask is obtained, and the blade curvature is quantified; s6, constructing a multi-dimensional feature vector reflecting the curling characteristics of the blade; and S7, outputting the drought grade corresponding to the corn leaf through the multi-layer perceptron model. By adopting the method, automatic identification and early warning of the early-stage water shortage state of the corn are realized.

The application discloses a kind of artificial climate chamber for plant phenotype monitoring and climate automatic regulation, it is related to artificial climate chamber technical field, artificial climate chamber is built-in several cultivation frame, artificial climate chamber includes cabin, regulating device and air distribution device, imaging device is built-in cabin, cabin is equipped with heat preservation cavity and air inlet, air inlet and heat preservation cavity are communicated, air inlet inlet section is equipped with compensation groove, compensation groove and heat preservation cavity are communicated, regulating device includes temperature control assembly and compensation component, compensation component and temperature control assembly are electrically connected, compensation component includes temperature sensing piece, temperature sensing piece includes two sections of material of different thermal expansion coefficients, air distribution device includes fan, fan and air inlet inlet pipeline are communicated, temperature sensing piece one end is inserted into air inlet, and the other end is inserted into compensation groove.In the process of plant cultivation, the plant to be cultivated is placed on the cultivation frame, double-row cultivation frame can be used in the artificial climate chamber, and multiple layers are provided on each cultivation frame to ensure the cultivation efficiency.

This invention provides a handheld method and device for detecting vegetation index of crop leaves, relating to the field of plant phenotyping technology. The method includes: controlling a contact linear array image sensor to scan the leaf of the crop to be tested line by line in the shaded leaf channel to obtain multi-channel reflectance grayscale digital quantities corresponding to each band; sequentially performing radiation baseline drift suppression, fluorescence cross-channel crosstalk compensation, and crosstalk bias correction on the multi-channel reflectance grayscale digital quantities to obtain multi-channel response data; determining the vegetation index of the leaf of the crop to be tested based on the multi-channel response data, and extracting the leaf phenotypic parameters of the leaf of the crop to be tested; determining the growth status of the leaf of the crop to be tested based on the vegetation index and leaf phenotypic parameters. This method can suppress the influence of ambient light interference, fluorescence crosstalk, and multi-band crosstalk on the measurement results, improve the accuracy of the vegetation index and the reliability of the leaf phenotypic parameters, and is beneficial for high-frequency, continuous, single-person handheld operation under field conditions.

The utility model discloses a hyper-spectral imaging analysis system for a RhizoTronn root phenotype, which relates to the technical field of plant phenotype and spectral imaging and comprises an X-axis, an imaging unit and an imaging light source are arranged on the X-axis, and the X-axis is equipped with a plant culture module. The technical problem that an existing plant root system phenotype imaging observation device cannot conduct in-situ imaging analysis on the side face of an overground plant and an underground root system at the same time is solved. The function of performing in-situ reflected light hyperspectral and UV-MCF biological fluorescence hyperspectral imaging analysis on the overground plant side surface and the underground root system of the plant at the same time through one set of imaging unit is realized; a root phenotype analysis method is expanded to comprehensive analysis of physicochemical properties, biochemical components and physiological states of plants and root samples from two hyperspectral dimensions of reflected light and fluorescence, synchronous imaging and coupling analysis of canopy phenotypes and root phenotypes are realized, phenotype analysis cost is reduced, phenotype imaging analysis efficiency is improved, and the method is suitable for popularization and application. And the space is also saved.

The application discloses an infrared corn drought identification method based on wavelet boundary enhancement and belongs to the field of agricultural informatization and plant phenotype identification, and comprises the following steps: S1, acquiring an infrared image of a corn plant; S2, performing two-dimensional discrete wavelet transform to generate a boundary response graph; S3, generating a binary mask according to the characteristics of the leaf edge of the boundary response graph; S4, performing connected domain analysis to obtain candidate leaves and generate independent single-leaf masks; S5, removing small holes to obtain a mask and quantifying leaf curvature; S6, constructing a multi-dimensional feature vector containing the curling characteristics of the leaves; and S7, outputting the corresponding drought grade of the corn leaves through a multilayer perception machine model. The above method realizes automatic identification and early warning of the early water shortage state of corn.

The invention provides an automatic rail changing type plant phenotype measurement platform and a rail changing method, and the platform comprises a measurement rail group which comprises at least two measurement rails arranged in a first direction, and the measurement rails extend in a second direction; the rail changing rail is arranged on at least one side of the measuring rail set in the second direction, and the rail changing rail extends in the first direction; the rail changing carrier is arranged on the rail changing track, and a carrier track extending in the second direction is arranged on the rail changing carrier; the measuring platform main body comprises a main frame body and a collecting and measuring unit arranged on the main frame body; the measuring platform main body has a switchable measuring state and a switchable rail switching state; when the measuring platform main body is in a rail changing state, the main frame body is located on the carrier rail, and the rail changing carrier can drive the measuring platform main body to move along the rail changing rail. According to the automatic rail changing type plant phenotype measuring platform and the rail changing method, the manual participation degree in the rail changing process is reduced, and the rail changing efficiency is improved.

A corn tassel branch counting method based on digital images and semi-supervised learning belongs to the field of plant phenotype measurement. The method is: the construction of a corn tassel branch identification model YOLOv5-C3CA: a new network model YOLOv5-C3CA is constructed based on YOLOv5 for corn tassel branch identification; a corn tassel branch identification model based on YOLOv5-C3CA model and introduction of semi-supervised learning; corn tassel identification and branch number extraction: the YOLOv5-C3CA detection model based on semi-supervised learning is applied to identify the corn tassel; after identifying the tassel, the YOLOv5-C3CA training model based on semi-supervised learning is continuously applied to extract the number of corn tassel branches. The corn tassel identification and tassel branch identification of the application are relatively accurate, which shows that the corn tassel branch counting scheme based on semi-supervised learning is feasible, which not only ensures the detection performance of the model, but also saves the data labeling workload.

The application belongs to the technical field of agricultural monitoring, and discloses a plant phenotype high-throughput monitoring system based on matrix control, which comprises a matrix monitoring module, a matrix plant phenotype detection environment is constructed, and to-be-detected plant samples are arranged in N plant cultivation units; each plant cultivation unit is respectively arranged with a corresponding environment regulation and control device, and is mapped into an addressable matrix node based on a logical addressing rule, and the matrix nodes form an environment control matrix through a distributed parallel communication link; a multi-factor monitoring module is arranged, a multi-factor collaborative monitoring network is built, a distributed sensor array arranged in the matrix nodes is arranged, a preset matrix sampling time sequence table is used, and each matrix node is uniformly triggered to synchronously collect SPAC parameters based on the preset matrix sampling time sequence table; the application breaks through the limitations of traditional plant phenotype collection means in terms of processing scale, regulation accuracy and data space-time consistency, and realizes automatic and intelligent management of the whole process of plant phenotype monitoring.

The invention discloses a multi-omics plant phenotype prediction method fused with small RNA, and relates to the field of bioinformatics and agricultural biology, and the method comprises the following steps: obtaining and preprocessing genotype, transcript expression and small RNA expression data to obtain a standardized feature matrix; performing feature screening based on Pearson correlation on the standardized feature matrix to obtain corresponding genotype, transcript expression and small RNA expression feature subsets; and based on a pre-trained MirGP prediction model, through multi-branch feature extraction, multistage efficient channel attention fusion and regression prediction, outputting a plant phenotype prediction result. According to the method, multiple types of small RNA regulation and control features are introduced and a hierarchical deep learning fusion architecture is constructed, so that multiple omics feature dimensions are expanded, data redundancy is reduced, and effective integration and deep mining of cross-modal features are realized.

A phenotyping device, method and system for measuring phenotypic traits of one or more plants in a target canopy, wherein the comprises a canopy spreader for spreading the target canopy, a camera unit, a means for controlling the angle and the distance of the camera unit in relation to the canopy spreader, an electronic control unit for controlling the camera unit. The phenotyping device, method and system enables recording the raw data from the target canopy during a data collection session comprising information of the plants that are not visible to a camera unit without spreading the canopy.

The invention relates to the technical field of plant phenotype information acquisition, and discloses a crawler-type plant phenotype information acquisition device and method.The crawler-type plant phenotype information acquisition device comprises a vehicle frame, a lifting adjusting mechanism is arranged on the vehicle frame, and the lifting adjusting mechanism is used for conducting lifting adjustment during information acquisition so as to adapt to plant phenotype information acquisition of different heights; the lifting adjusting mechanism is connected with a distance adjusting mechanism, the distance adjusting mechanism is used for adjusting the distance when information is collected, the whole machine is driven by an electric system, two sets of driving motors are used for landing on the tracks on the two sides, one-way orbital transfer and pivot steering are achieved, and a battery box body can move transversely; the cameras share one lead screw guide rail and can independently ascend and descend to meet the shooting requirements of crops at different angles under the same walking cross section, camera lead screws are installed on an integral installation frame, the installation frame can synchronously ascend and descend through electric cylinders on the two sides, and integral adjustment under the same camera arrangement state is achieved.

The invention discloses a plant phenotype extraction-oriented point cloud calibration method and a plant phenotype extraction-oriented point cloud calibration system. The method comprises the following steps of arranging a 3D printing cylindrical disc with known sizes and clear plane normal characteristics and a standard color card in a plant point cloud acquisition scene; extracting point cloud data of the cylindrical disc region, and carrying out point cloud scale and direction calibration based on plane fitting and normal vector calculation of the cylindrical disc region; cutting the calibrated point cloud data to obtain a color card area point cloud, and generating a point cloud top view; performing color block detection on the point cloud top view, extracting an effective color block, constructing a matching relationship between a reference color block of a standard color card and a center point of the effective color block of the point cloud, calculating a rotation matrix and a translation vector between the reference color block and the point cloud color block based on a two-dimensional rigid transformation model, mapping the reference color block into the point cloud top view, and performing missing color block prediction; and constructing a color correction matrix to realize global color calibration. According to the invention, the consistency of the scale, direction and color of the point cloud can be improved.

The invention relates to the technical field of plant phenotype information acquisition, and discloses a mobile plant phenotype information acquisition device and a method thereof.The mobile plant phenotype information acquisition device comprises a main body frame, the main body frame is provided with a movement mechanism, the movement mechanism is used for driving the device to move, walking can be achieved, and linear walking and in-situ turning can be achieved; in addition, transverse movement can be achieved in the movement process, adjustment can be conducted in the movement process according to the flatness of the ground, it is guaranteed that four wheels touch the ground all the time, and the movement stability is guaranteed; according to the plant phenotype information acquisition device, plant phenotype information can be acquired, the acquisition direction and angle can be adjusted in the acquisition process, the acquisition height and interval can be adjusted, and the plant phenotype information acquisition device is suitable for acquisition of plants with different heights; according to the device, plants can be centralized during collection, inclined plants can be centralized, information collection is facilitated, and the collection efficiency is improved.

The invention discloses a plant phenotype image acquisition method based on a wheeled phenotype platform, which relates to the technical field of plant image acquisition, and comprises the following steps: deploying the wheeled phenotype platform and planning an acquisition route, acquiring crop image data and environment data in real time, and acquiring historical phenotype information data of crop growth at the same time; the method comprises the following steps: preprocessing collected image data and environment data, extracting phenotypic characteristics of crops for quantitative description and measurement, and carrying out statistical analysis on the extracted phenotypic characteristics to obtain an illumination interference factor and a shielding interference factor; according to the invention, quantitative analysis is carried out on the growth speed, morphological change, leaf color and the like of the plant through image processing and data analysis technologies, a scientific basis is provided for growth monitoring and early warning of plant diseases and insect pests, and in addition, the platform can combine the collected image and data with a plant growth model, so that the early warning of plant diseases and insect pests is realized. Growth trend prediction and yield estimation are carried out, and accurate decision support is provided for agricultural production.

The invention provides a fireproof design method and device which are suitable for plant phenotypic facility or indoor facility agriculture, and the method comprises the steps: determining the safety distance of each functional region based on critical radiation heat simulation; according to the safety distance, the fire scale of each functional area under the most unfavorable working condition is obtained by combining the characteristics of combustible materials in each functional area or through analogue simulation; calculating a fire probability prediction value of each functional area under different second-class hazard sources and a fire loss prediction value under different fire scales, and evaluating a fire risk evaluation level of each functional area through the risk matrix; and in combination with the fire scale, the fire risk evaluation grade and the safety distance, matching fire prevention control measures corresponding to each functional area. The method can effectively improve the fireproof control of plant phenotypic facility or indoor facility agriculture, and achieves the real-time monitoring and prediction functions of fire risks.

Method and system for plant phenotyping. Sequence of images of agricultural scene are captured along trajectory using image sensor of mobile computing device. Position and orientation data is obtained for each captured image using inertial sensor of mobile computing device. Point cloud generated in captured images, point cloud including coordinates defining distances to respective features in image. An object is selected, object belonging to plant hierarchy level of: plant organ; plant; plant grouping; or plant field. For each selected first object in first plant hierarchy level, object is identified and tracked in successive images of sequence, distance coordinates for object are supplemented when point cloud insufficient; visually undetectable portions of object are supplemented in images; spatial distance of object in respective images is determined based on point cloud coordinates and supplemented distance coordinates; and at least one phenotype of object is determined, based on determined spatial distance of object.

The invention relates to a plant rapid breeding regulation and control system and method. The system comprises a plant growth information acquisition module, a measurement module and an information processing module, the information processing module detects the development stage of a plant through panoramic image data of the plant, and when the plant enters a new development stage, the measurement module is started and obtains gene expression quantity information of a primary plant tissue; and when the genes with abnormal expression quantity detected by the measurement module are the genes influencing the growth of the plant tissues needing to be harvested, the information processing module acquires plant phenotypic defect related information and generates an environment regulation instruction. The method comprises the following steps: collecting panoramic image data of a plant, and determining the development stage of the plant through comparison; acquiring gene expression quantity information of primary plant tissues; and when the gene with abnormal expression quantity is the gene influencing the growth of the plant tissue needing to be harvested, adjusting one or more environmental factors in the plant growth environment in a manner of positively guiding the plant growth.

This invention relates to the fields of computer vision and soybean plant phenotyping technology, and discloses a method, system, and apparatus for soybean plant phenotyping analysis based on 3D reconstruction. The method includes the following steps: The invention uses a device for multi-view image acquisition of soybean plants to collect images from multiple perspectives, and constructs an end-to-end soybean plant leaf phenotyping model based on PointNet Transform. This model can directly predict key phenotypic features such as leaf area and leaf perimeter from high-precision leaf point cloud data. Before constructing the model, discrete point cloud voxelization processing is required to provide conditions for semantic segmentation. This invention achieves automated acquisition of soybean plant point clouds, automated segmentation of soybean plant point clouds, and automated measurement of soybean plant leaf point cloud phenotyping, providing a data foundation for soybean plant phenotyping analysis and realizing high-throughput, intelligent soybean plant phenotyping analysis.

This invention relates to the technical field of plant phenotypic information acquisition equipment, and discloses a dual-purpose (rail and road) wheeled phenotypic information acquisition platform, including a long-distance platform frame. The long-distance platform frame is equipped with a spacing adjustment mechanism, enabling the acquisition of plant phenotypic information. During acquisition, it can capture images of the entire plant, performing rotating circular acquisition, resulting in a wide range of acquisition efficiency. This overcomes the problem of existing acquisition methods that cannot capture images of the entire height of the plant. It can move on railway tracks and highways, broadening its application range. Furthermore, it can simultaneously acquire images of multiple plants, resulting in high acquisition efficiency. The wheel spacing can be adjusted during acquisition to accommodate acquisition movements of different widths.

The invention provides a plant phenotype missing data interpolation method based on an auto-encoder, and the method comprises the steps: inputting plant sample phenotype data with a missing value into a trained optimal auto-encoder model, and outputting a missing phenotype prediction value. The acquisition method of the optimal auto-encoder model comprises the following steps: hiding each character of plant sample data according to a preset proportion to obtain a test set, and recording a hiding position; performing pre-interpolation processing on the test set at a position corresponding to the hidden position to obtain a training set; constructing missing samples having similar features with the test set as a mask data set; inputting the mask data set and the training set into a pre-constructed auto-encoder model for iterative training to obtain an optimized auto-encoder model; and inputting the test set into the optimized auto-encoder model, and outputting a missing phenotype prediction value. The problem that the interpolation accuracy and feasibility are limited due to the fact that an existing plant phenotype data interpolation method excessively depends on a reference data set is solved.

The invention relates to a multi-factor regulation and control system and method for accelerated breeding of plants. The system comprises a collection module, a processing module and a measurement module. Based on a comparison result generated by a functional gene sequence provided by the measurement module and a standard functional gene sequence stored in a database, the processing module generates an instruction for controlling the collection module to collect the plant image and provides environmental parameters conforming to the plant growth tendency for the regulation and control module. The method comprises the following steps: comparing plant functional gene sequences; determining an abnormal site of the functional gene sequence; when the abnormal site of the functional gene sequence is consistent or partially consistent with the abnormal site which is stored in the database and affects the expression of the functional gene, the processing module generates an instruction for controlling the collection module to collect the image of the first plant phenotype. By comparing the plant functional gene sequences, the abnormal sites of the functional gene sequences can be quickly and accurately identified, so that whether the plant has gene mutation or epigenetic change which affects the growth and development of the plant can be judged.

The application discloses a plant phenotype analysis intelligent agent system based on a large language model. The system adopts a star topological structure, takes a planning intelligent agent as a hub, and is provided with special intelligent agents for vision, model fine tuning, code generation, GWAS analysis and paper writing, so that an overall process from multi-scale plant image processing, zero sample plant image segmentation, plant three-dimensional reconstruction and point cloud segmentation, neural network model optimization, visual chart drawing, genotype-phenotype analysis to scientific research paper writing is realized, and a one-stop intelligent solution is provided for accelerating high-throughput phenomics and genomics correlation research.

The invention relates to the technical field of plant phenotype imaging, and discloses a high-flux hydroponic ginseng 3D time sequence imaging system and method. According to the system, an original image flow is obtained from a multispectral image acquisition device of a water culture container, and a ginseng main body structure image and a root system microenvironment image are separated. Executing voxelization growth element decoupling on the main body structure image to obtain three growth components of structure, form and metabolism; and performing microenvironment element extraction on the microenvironment image to obtain a nutrient flow field, dissolved oxygen and root exudates distribution. Based on the components, a primary metabolic pressure field, a secondary metabolic pressure field and a feedback suppression field are respectively constructed. And inputting the three fields into a growth situation deduction engine, and outputting a three-dimensional growth situation deduction result comprising a growth hotspot area map, a hysteresis area map and a competition area map. According to the method, synchronous, global and refined analysis and prediction of ginseng growth and rhizosphere environment are realized.

The application provides a kind of semi-automatic field scene under the phenotype data measurement method and system, belong to the phenotype data measurement field under the outdoor field scene, the method comprises: collecting crop image, and according to crop image, obtain the target phenotype center coordinate point specified by user;Crop image and target phenotype center coordinate point specified by user are segmented by pre-trained SAM model, and target phenotype mask image is obtained;The morphological characteristics of the mask image are obtained by judging the bending characteristics of the mask image, wherein the morphological characteristics include straight shape and curved shape.The application can be widely applied to various plant, plant phenotype data calculation method under the scene without data collection, calibration and model training, and can avoid a large amount of cost caused by data annotation and model training.

The invention discloses a One Health-based crop disease and insect pest intelligent diagnosis system, which comprises a data acquisition module, a diagnosis prediction module, a prevention and control module, a data storage module and a man-machine interaction module, and comprises the steps of 1, acquiring a crop canopy image and farmland three-dimensional point cloud data, 2, carrying out in-situ monitoring and establishing a soil health index model, step 3, predicting the probability of disease and pest outbreak, step 4, triggering a corresponding prevention and control strategy according to a prediction result, and step 5, recording a prevention and control process and generating a traceability data chain. According to the method, a plant phenotype-microbiome-environmental factor ternary analysis framework is put forward for the first time, multi-source heterogeneous data feature level fusion is performed through a Transform model, cross-domain data deep fusion is realized, the diagnosis accuracy is effectively improved, meanwhile, a deep reinforcement learning adaptive learning mechanism is introduced into the system, a classifier is dynamically updated according to new pest and disease damage data, and the diagnosis accuracy is improved. And the manual annotation demand is effectively reduced.

The application discloses a kind of small RNA fusion multiomics plant phenotype prediction method, it is related to bioinformatics and agricultural biology field, including the following steps: obtaining and pre-processing genotype, transcript expression and small RNA expression data, obtain standardized feature matrix;The standardized feature matrix is screened based on the feature of pearson correlation, and the corresponding genotype, transcript expression, small RNA expression feature subset is obtained;Based on the pre-trained MirGP prediction model, by multi-branch feature extraction, multi-level efficient channel attention fusion and regression prediction, output plant phenotype prediction result.The method introduces multiple small RNA regulation characteristics and constructs hierarchical deep learning fusion architecture, expands the multiomics feature dimension, reduces data redundancy, realizes the effective integration and deep mining of cross-modal features.

The application relates to the technical field of agricultural information detection, and particularly discloses an intelligent breeding phenotype collection and mobile measurement platform, which comprises an image collection unit, a positioning and attitude sensing unit and a central processing unit carried on a mobile device. The central processing unit is used for performing the following operations: collecting a plant image sequence through a multispectral camera, and performing waveband selection and pretreatment based on plant leaf reflectance spectrum characteristics; performing space-time alignment on the image, and identifying and repairing a blurred area caused by wind; performing image enhancement and noise removal on a low-contrast plant contour area; and integrating position and visual data by using a sensor fusion algorithm to generate integrated pose data. The application effectively solves the problems of data distortion and incomplete coverage caused by wind blur, noise interference, mobile device shaking and path fixation in a dynamic field environment, and improves the precision and efficiency of plant phenotype collection.

The application relates to the technical field of three-dimensional reconstruction, and particularly discloses a semantic plant three-dimensional reconstruction and phenotype extraction method and system, which comprises the following steps: acquiring a multi-view image sequence of a target plant; performing organ-level semantic segmentation on the image of each view to obtain semantic segmentation masks of each organ; based on the principle of geometric consistency, performing cross-view organ instance ID matching on the multi-view semantic segmentation masks to generate unified instance identification of the same organ under different views; fusing the semantic information of the unified instance identification and performing a three-dimensional Gaussian splashing reconstruction process to generate a three-dimensional point cloud model with organ-level semantic labels; and extracting predetermined plant phenotype parameters based on the semantic three-dimensional point cloud model. The application solves the problems of three-dimensional reconstruction without semantics, a fragmented phenotype extraction process, and cross-view organ correlation confusion, realizes full-process automation and high-precision extraction from images to phenotype parameters, and significantly improves the efficiency and objectivity of plant phenotype analysis.

The application discloses a plant phenotype-oriented three-dimensional reconstruction method and a collection system. The method first acquires multi-view image data of a plant through a 360-degree ring collection system integrated with an RGB camera and a depth camera; then generates an accurate foreground mask by using a foreground semantic segmentation model; then combines structure from motion (SfM) and depth point cloud, and performs fusion under the guidance of the foreground mask to obtain an initial point cloud; then performs mask-guided 3D Gaussian splashing (3DGS) optimization reconstruction based on the point cloud, and the optimization process improves the identification of small plant organs by using a mask weighted loss function and a semantic-guided density control strategy; and finally derives a color point cloud from the optimized Gaussian model, which can be directly used for extraction of phenotype parameters such as plant height and leaf area. The application solves the problems of large noise and detail loss in the prior art when reconstructing plants in a complex background, and realizes fast three-dimensional reconstruction which can be directly used for phenotype analysis.