43 results about "Species classification" patented technology

A crop species identification system based on satellite telemetry data is disclosed, which includes: a receiver module, receiving plural telemetric vegetation indexes of a target area; a data cleaning module, cleaning anomalous data in the telemetric vegetation indices, to correspondingly generate cleaned index data; a feature extraction module, including at least two different convolution kernels for mapping the cleaned index data into at least two feature scale mapping data which respectively correspond to the convolution kernels, performing a pooling operation of the feature scale mapping data to generate a pooled data, and concatenating the at least two feature scale mapping data and the pooled data into concatenated data; and a classification module, including a fully-connected layer, for extracting features of the concatenated data, to generate a species classification result for the target area.

The invention provides a method and a system for identifying migrated insects in the air based on an insect radar and a high-altitude lamp. The method comprises the following steps of: 1, receiving four biological parameters of flapping frequency, body length, body width and body weight of each migrated insect in the air monitored by the insect radar in real time; 2, the received four biological parameters are input into an optimal insect type classification model, the type of each aerial migratory fly insect is obtained, and the optimal insect type classification model is determined according to the counted types of the insects trapped by the high-altitude lamp within the recent T3 time. According to the invention, real-time accurate identification of insect individual types can be realized for target insects captured by an insect radar; the real insect condition information is monitored in real time according to the received data of the insects trapped by the high-altitude lamp, the optimal insect type classification model is determined according to the real insect condition information, it is guaranteed that the optimal insect type classification model is matched with the real insect condition, and good expansibility and broad-spectrum adaptability are achieved.

The invention discloses a vegetation species identification method for enhancing AI computing power and time sequence tracing, and the method comprises the steps: collecting and preprocessing multi-source remote sensing data, and constructing a vegetation classification data set with time-space alignment and uniform resolution; generating a vegetation mask based on the vegetation classification data set, obtaining a standardized sample slice, and constructing a training data set in combination with spectral features; performing time-phase-sharing processing on the training data set based on vegetation phenological characteristics, and outputting a preliminary vegetation classification result; performing object-level optimization on the preliminary vegetation classification result to obtain an optimized vegetation classification result; and correcting an optimized vegetation classification result in combination with topographic data, generating a vegetation species classification map of a target year, and realizing vegetation dynamic change inversion in a specified time period through transfer learning. Therefore, traditional resolution limitation can be broken through, the classification error problem caused by independent use of multi-source data is solved, the distinction degree of complex vegetation types is remarkably improved, and historical vegetation dynamic backtracking analysis is supported.

The invention relates to the technical field of metagenomics, discloses an automatic analysis method and system for pit mud metagenomic data, and aims at solving the problem that an existing method is poor in efficiency and accuracy, and the scheme mainly comprises the steps that a sequencing data type, a file path and analysis parameters are received; performing quality control on the original offline data; sequence assembly is carried out, and a contigs file is generated; carrying out assembly quality evaluation on the contigs file; carrying out genome binning by using at least two binning tools; integrating output results of the binning tool, and performing optimization based on a preset integrity threshold value and a preset pollution degree threshold value to obtain an optimized binning genome data set; evaluating and optimizing the integrity, the pollution degree and the strain heterogeneity of the binning genome; calculating coverage and relative abundance; performing species classification annotation and function annotation; and integrating the result data of the previous steps to generate an analysis report. According to the method, automatic analysis of metagenome data is realized, and the analysis efficiency and accuracy are improved.

The invention relates to the technical field of plant ecological monitoring and computer vision, and discloses a grassland dominant species identification method based on unmanned aerial vehicle remote sensing and deep learning, and the method comprises the steps: carrying out the fusion processing of a to-be-identified unmanned aerial vehicle remote sensing image of a target grassland region, and obtaining a to-be-identified fusion image; inputting the fusion image to be identified into a trained species identification model to obtain a species classification result of the target grassland area; combining the species classification result with the space coordinates of the target grassland area to generate a species distribution diagram of the target grassland area; according to the method, the problem of insufficient species identification precision in a complex grassland environment can be solved, manual investigation cost and subjective deviation are reduced, grassland dominant species identification accuracy and spatial analysis capability are improved, and refined data support is provided for grassland ecological protection.

This invention relates to a method for species classification in wildlife monitoring images based on point context, belonging to the field of image classification technology. The method involves modeling the correlation between monitoring points and species, assigning weights to samples based on the strength of the correlation, and decoupling the point classification and species classification tasks to learn environmental and species characteristics. By using the environmental information represented by the points as context for species classification, the method enhances the model's understanding of the relationship between species and habitat environments, achieving effective generalization to out-of-distribution data.

This application relates to the field of computer vision technology, and in particular to a method for wildlife identification and tracing using infrared image enhancement and multi-scale fusion. The method includes: acquiring raw images from an infrared camera and enhancing the raw images using Retinex and contrast-limited adaptive histogram equalization; performing animal target detection, feature extraction, and feature fusion by optimizing the YOLOv8 model and using a multi-modal fusion Transformer to obtain multi-scale fused features; identifying species categories using a pre-defined primary species classification model and extracting and matching individual biometric features of key species using a pre-defined secondary individual identification model to generate individual identifiers for corresponding animals; and automatically associating the species identification results, individual identifiers, and multi-source forestry data to generate an intelligent decision-making report for wildlife population analysis and conservation management. This application contributes to achieving high-precision, automated wildlife identification and tracing.

The invention provides an urban tree species classification method, system and device, and belongs to the technical field of tree species classification, and the method comprises the steps: firstly, selecting candidate wavebands in a red-edge spectrum range, traversing all dual-waveband combinations, and defining a vegetation index through a preset formula; secondly, for each index, generating a vegetation mask by using a plurality of candidate thresholds, and calculating a confusion matrix evaluation index by comparing the vegetation mask with a real mask; and finally, based on the index, automatically screening out an index with optimal performance and a threshold pair. Wherein the index defined by the optimal dual-band combination is the final vegetation index, and according to the final vegetation index, classification and identification of urban tree species are carried out through a machine learning model. The method solves the problem that classification of the classic vegetation index on the urban tree species is not accurate.

The invention discloses an active living object-oriented fresh food weighing data edge calculation optimization method, and belongs to the technical field of fresh food weighing, and the method comprises the following steps: collecting weighing time sequence data of fresh food species in high, medium and low vigor states, extracting quantitative characteristics, and calculating the weight of the fresh food species in the high, medium and low vigor states; meanwhile, a freshness classification rule and a species classification rule are trained through machine learning to establish a rule base, and then filtering and calculation strategies are set according to the freshness rule; the method comprises the following steps: acquiring an image through a camera to judge the type of a living object, loading a corresponding rule library from an edge device species feature library, and when the classification confidence is lower than 90%, secondarily deducing and matching species classification rules through a weight fluctuation mode; the invention discloses an active living object-oriented fresh food weighing data edge calculation optimization method, which can improve the weighing precision and greatly improve the weighing efficiency through species classification, freshness judgment and secondary identification.

The invention discloses an animal parasite intelligent diagnosis system and method based on artificial intelligence. The system comprises a terminal device, a terminal device used for obtaining a microscopic image of an animal parasite, and a cloud server used for receiving the microscopic image. The cloud server deploys a parasite target detection model and a parasite type classification model; the parasite target detection model processes the received microscopic image and outputs bounding box coordinate information in the image; cutting out corresponding suspected area image blocks based on the bounding box coordinate information; inputting the suspected region image blocks into a parasite type classification model to obtain an authenticity discrimination result and a parasite type identification result of each suspected region; according to the authenticity judgment result and the parasite type recognition result, a diagnosis report is generated and returned to the terminal equipment for display, the problems that special equipment is expensive and terminal application computing power is insufficient are solved, and unification of high performance, low cost and high portability is achieved.

The application discloses a pangolin species identification method based on nail piece element characteristics, which comprises the following steps: step S1: collecting nail piece samples covering existing 10 kinds of pangolin species, and obtaining initial data of multi-element characteristics of the nail piece samples; step S2: identifying the species of the nail piece samples by a molecular biology method, and obtaining the species labels of each nail piece sample; step S3: constructing an "element characteristic-species label" standard data set; step S4: selecting a table prior fitting network as a core machine learning algorithm, training a model by using training set data, and constructing the model; and step S5: identifying whether a to-be-detected sample belongs to a pangolin nail piece according to morphological judgment. The method has a single sample data collection time of less than 60 seconds, does not need to damage the sample, is used for nail piece identification of samples without prior clues, and the overall accuracy of a global ten-species classification model reaches 0.93.

The application provides an aerial migration insect identification method and system based on an insect radar and a high-altitude lamp, the method comprising the following steps: 1, receiving the wing beat frequency, body length, body width and body weight of each aerial migration insect monitored by the insect radar in real time; 2, inputting the four biological parameters into an optimal insect species classification model to obtain the species of each aerial migration insect, wherein the optimal insect species classification model is determined according to the insect species captured by the high-altitude lamp in the latest T3 time. The application can realize real-time accurate identification of the species of the target insect captured by the insect radar, can monitor real insect information in real time according to the received insect data captured by the high-altitude lamp, and can determine the optimal insect species classification model according to the real insect information, so that the optimal insect species classification model is matched with the real insect information, and the application has good expansibility and broad spectrum adaptability.

The invention relates to the technical field of smart agriculture, and provides an unmanned aerial vehicle hyperspectral technology-based pecan tree species classification method, which comprises the following steps of: obtaining a hyperspectral image, and performing image cutting on the hyperspectral image to obtain a plurality of block images with specified sizes; analyzing a single hickory tree variety accurately, performing first feature extraction on the block image to obtain first feature information corresponding to the block image, and performing second feature extraction on the spatial feature to obtain second feature information corresponding to the spatial feature; feature information corresponding to the target is accurately collected through two times of feature extraction of different dimensions, the block images are classified based on the first feature information and the second feature information, classification information corresponding to the block images is obtained, and the classification information comprises pecan tree species information; based on the block images and the classification information, the pecan tree species classification image is obtained, an accurate pecan tree species classification result can be obtained, planting distribution of pecan tree species in the forest land is analyzed, and sustainable development of the forest land is promoted.

The invention discloses a pathogen gene sequence detection and analysis method based on nanopore sequencing, and relates to the field of customs quarantine, and the method comprises the following steps: carrying out nanopore sequencing on a collected sample to obtain original sequencing data, and carrying out quality screening to obtain a high-quality sequence. Species classification and mutation recognition are completed, and pathogen types and mutation characteristics are extracted. And constructing a propagation map based on the classification information, the mutation information, the sampling time and the sampling place, and forming a dynamic propagation structure chart. And calculating a propagation path score between the samples through graph diffusion modeling. And constructing a Bayesian causal model in combination with mutation characteristics and propagation scores, and reasoning the posterior probability of high-risk events caused by pathogens. And generating a risk level label according to the posterior probability, outputting a pathogen analysis report and early warning information, and realizing integrated analysis of rapid detection and risk assessment.

The invention discloses an algae community detection method and system based on polarized light characteristics, and relates to the field of water environment monitoring, and the method comprises the steps: obtaining an algae single particle polarization scattering signal through multiple observation angles, resolving a Stokes parameter, further extracting a polarization phase to construct a polarization phase divergence matrix, and carrying out the structural representation of the polarization difference between the angles; meanwhile, a morphological complexity index is obtained through multi-angle scattering intensity statistics and is used for depicting scattering distribution characteristics; on the basis, a polarization-form coupling coefficient is provided, polarization phase differences of different angle combinations are subjected to weighted convergence and modulated by form complexity, and stable fusion of polarization information and form statistical information is achieved; and the coupling coefficients are subjected to statistical convergence in a time window to form an algal group cooperation index, so that unified output of single-particle species classification and community risk early warning is realized. Therefore, the distinction degree, the stability and the time sequence characterization capability of algae community detection are improved.

The application discloses a pathogen gene sequence detection and analysis method based on nanopore sequencing, and relates to the field of customs quarantine.The method comprises the following steps: nanopore sequencing is performed on collected samples, original sequencing data is acquired, quality screening is performed, and high-quality sequences are obtained; species classification and mutation recognition are completed; pathogen types and mutation characteristics are extracted; a transmission map is constructed based on classification information, mutation information, sampling time and location, and a dynamic transmission structure diagram is formed; a graph diffusion modeling calculation is performed to calculate the transmission path score between samples; a Bayesian causal model is constructed by combining the mutation characteristics and the transmission score, and the posterior probability of a pathogen causing a high-risk event is inferred; a risk level label is generated according to the posterior probability, a pathogen analysis report and early warning information are output, and integrated analysis of rapid detection and risk assessment is realized.

The application belongs to the technical field of molecular biology, and discloses a method for evaluating the species / typing level identification capability of universal primers. The method comprises the following steps: constructing a microorganism database, aligning the universal primers to the microorganism database to obtain target fragments and species / typing information corresponding to the universal primers, obtaining sequencing data by cutting the target fragments from the F end to a specific length, aligning the sequencing data to the microorganism database for species classification and identification after deduplication and statistics, calculating the sequence return rate of the universal primers for a specific species / typing, and evaluating the identification capability of the universal primers for the specific species / typing level. The application provides a method for evaluating the species level identification capability of universal primers for pathogenic microorganism identification, and evaluates the performance of the universal primers for identifying species / typing.

The application discloses an open set fine-grained image recognition method based on retrieval-enhanced multi-modal reasoning, and comprises the following steps: obtaining a query image to be recognized; using the query image to be recognized to perform candidate species recall in a pre-established species reference retrieval library, so as to obtain k candidate species and example images corresponding to each candidate species; all example images corresponding to all candidate species form a retrieval-enhanced context; inputting the obtained query image to be recognized and the retrieval-enhanced context into a pre-constructed multi-modal reasoning model to perform chained thinking comparison reasoning, so as to obtain a reasoning result. The application can solve the technical problem that the existing closed set classification method often separates the known species classification and unknown species discovery for processing, and is difficult to realize simultaneously in a unified framework.

The invention relates to the field of image processing, and particularly discloses a multi-dimensional feature-based wetland species classification method, which is characterized in that radiation mutual calibration is carried out on satellite and unmanned aerial vehicle image data sources by introducing a domain self-adaption thought, so that a fused image with high spatial resolution and high spectral fidelity is generated; the problem of spectrum inconsistency of satellite and unmanned aerial vehicle images caused by sensor and environment differences is solved, and the quality of fused image data is fundamentally improved. On the basis of the high-quality data, the limitation of dependence on a single spectral feature traditionally is further broken through, a vertical structure (namely height information) of vegetation is introduced as a key distinguishing feature by mining an elevation model generated by unmanned aerial vehicle data, and the vertical structure of the vegetation is distinguished by combining spectrum, texture and specific physiological indexes. A multi-dimensional feature space capable of comprehensively reflecting the internal difference of the wetland species is constructed, and a reliable technical path is provided for fine classification of the wetland species.

The application provides a pet behavior training system based on AI vision, and belongs to the technical field of pet training. The system comprises a pet locator, a visual acquisition module, a data processing module, a strategy matching module and an execution feedback module. By integrating positioning information, physiological signals and image data, the system realizes pet individual differentiation, behavior characteristic extraction, behavior type and confidence determination, and analyzes physiological information to obtain basic emotional state and quantitative emotional intensity and stability parameters. In combination with a species classification model, the pet species is determined, a species-specific reward and punishment strategy library is called, the reward and punishment intensity is dynamically adjusted according to behavior attributes, confidence and emotional parameters, and a complete closed-loop training mechanism is formed. The application solves the problems of subjective misjudgment, poor species adaptability and neglect of pet emotions in traditional training, improves the behavior recognition accuracy, emotional perception objectivity and reward and punishment strategy precision, and is suitable for scientific and efficient pet behavior training in multiple species and multiple scenes.

The invention provides a voiceprint recognition-based migrant bird species classification method, and relates to the technical field of migrant bird species classification, and the method comprises the steps: collecting an original voiceprint signal of migrant bird buzzing in a migrant bird migration region; carrying out adaptive beam forming processing on the original voiceprint signal; performing space-time joint voiceprint feature separation processing on the directional audio signal; inputting the three-dimensional voiceprint feature body into a federal migration capsule network for model training; inputting the three-dimensional voiceprint feature body collected in real time into a species classification model; and when the confidence coefficient is lower than a set threshold value, inputting the three-dimensional voiceprint feature body collected in real time into a zero sample learning module. According to the invention, through integrating the three-dimensional features of the time-frequency diagram, the space azimuth angle and the frequency modulation track, the distinction degree of the sound features in the complex field environment is obviously improved; through combination of a confidence coefficient threshold value and a zero sample learning module, when new species outside a training set are encountered, preliminary classification is automatically generated through biological characteristic space mapping, and the rejection rate of unknown species is greatly reduced.

This invention provides a computer-aided analysis system based on high-throughput sequencing data and knowledge graphs, including a data access agent, a quality control analysis agent, a species annotation agent, a knowledge graph reasoning agent, a report generation agent, and an agent scheduling center. The data access agent acquires sequencing data and clinical phenotypic data; the quality control analysis agent performs quality control using the Illumina NovaSeq platform, V3-V4 primers, the SILVA v144 database, Q30 ≥ 90%, and sequences ≥ 50 bp; the species annotation agent performs species classification annotation based on a reference database; the knowledge graph reasoning agent performs association analysis based on a microbiome-disease-intervention knowledge graph, using ET_MGNN and RTGN models, including 264 search keywords covering dental caries, periodontitis, oral cancer, Alzheimer's disease, diabetes, and heart disease; the report generation agent generates personalized analysis reports and intervention suggestions; and the agent scheduling center manages and dynamically schedules all agents, forming a complete closed loop of "perception → analysis → diagnosis → intervention → execution → feedback".

The present application relates to the technical field of forest and grass resources identification, and particularly relates to a forest and grass resource accurate identification method based on image analysis, comprising the following steps: step S1: collecting multi-view sequence images of target forest and grass sample plots, synchronously acquiring real-time positioning data, shooting posture parameters and environment site basic data corresponding to the sample plots; step S2: uniformly preprocessing the collected sequence images; step S3: matching feature points, positioning and posture data based on the sequence images; step S4: multi-scale segmentation of the standardized images, extracting the contours and feature parameters of different types of vegetation targets such as trees, shrubs and herbaceous plants; step S5: completing forest and grass species classification, resource parameter inversion and health grade determination based on the extracted feature parameters, and outputting the forest and grass resource investigation results matched with geographic coordinates. The present application is free from the limitation of physical scales, is suitable for multi-plant structure detection, and realizes full-plant parameter calculation without the constraint of close-range targets.

The invention discloses a metagenome species level classification method based on a Debrueine graph, which takes a linear genome and a generic genome as a complementary reference framework, and adopts a two-stage process: firstly, executing approximate member query by utilizing HIBF, and screening out candidate species and strains from a database in a high-recall and low-overhead manner; and then ccDBG is constructed around the candidates, and the attribution of each read length is finely judged on the strain scale. The method has the rapid screening capability of linear reference and high-resolution capture of a generic genome on intra-population differences, so that the problems of fuzzy and confusion of species classification in a traditional process are remarkably relieved.

The invention provides a pet behavior training system based on AI vision, and belongs to the technical field of pet training. The system comprises a pet positioner, a visual acquisition module, a data processing module, a strategy matching module and an execution feedback module, and realizes pet individual distinguishing, behavior feature extraction and behavior type and confidence judgment by integrating positioning information, physiological signals and image data. Analyzing the physiological information to obtain a basic emotional state and quantized emotional intensity and stability parameters; and pet types are determined in combination with a species classification model, a species exclusive reward and punishment strategy library is called, the reward and punishment strength is dynamically adjusted according to behavior attributes, confidence coefficients and emotion parameters, and a complete closed-loop training mechanism is formed. The method solves the problems of subjective misjudgment, poor species adaptability, neglect of pet emotions and the like in traditional training, improves the behavior recognition accuracy, emotion perception objectivity and reward and punishment strategy precision, and is suitable for multi-species and multi-scene scientific and efficient pet behavior training.

The invention provides a method for constructing a clinical-grade respiratory tract pathogenic microorganism detection genome database. Key functions of screening of high-quality genome sequences of species, removal of redundant genome sequences, classification of wrong species of the genome, filtration of pollution sequences in the genome and the like are integrated.

The invention discloses a species classification method based on a hybrid expert model and comparative learning, and relates to the technical field of bioinformatics and artificial intelligence crossing, in particular to a species classification method. The invention aims to solve the problems of difficulty in separation, inaccurate classification, incapability of identifying novel pathogens and the like when a mixed genome sample is processed in the prior art. The method comprises the following steps: step 1, defining a genome deep learning model; 2, training the genome deep learning model to obtain a trained genome deep learning model for pathogenic microorganism separation and classification; and step 3, obtaining to-be-detected genome sequence data, inputting the to-be-detected genome sequence data into the trained genome deep learning model for pathogenic microorganism separation and classification obtained in the step 2, and outputting a genome sequence data source and a hierarchical classification task result by the model.

The invention discloses a metagenome virus classification method based on a weighted graph network and label propagation, and relates to a metagenome virus classification method. The method aims at solving the problems that in the prior art, the dynamic adaptability of ICTV classification standards is insufficient, data categories are unbalanced, and dark matter classification is difficult. The method comprises the following steps: in a database updating and constructing process, acquiring latest virus classification sample data, and constructing a latest virus protein database; in the weighted graph network construction process, firstly, a to-be-detected virus nucleotide sequence is translated into a protein sequence, then a protein cluster is obtained through Markov clustering, finally, the weight of an edge between the to-be-detected virus sequence and a reference sequence in a database is calculated through an edge weight calculation algorithm, and a weighted graph network is constructed; in a label propagation virus classification process, comprehensive virus species classification is carried out on metagenome viruses in three stages through a category weight normalization label propagation algorithm. The invention belongs to the technical field of gene engineering.

This application relates to the fields of intelligent measurement and machine vision technology, and discloses an intelligent sensing and measurement system for biological body length based on image recognition. The system includes modules for multimodal data acquisition, dynamic behavioral feature extraction, adaptive weight filtering, bidirectional collaborative locking based on apparent weight, and image length measurement and data fusion. It dynamically adjusts the filtering window based on the frequency of struggling to smooth physical impact noise, and combines the rate of weight change with body curvature for bidirectional collaborative locking. Based on pre-set species classification labels, it matches corresponding feature point detection strategies to achieve absolutely synchronous and accurate measurement of the weight and dimensions of multiple body parts of different species such as fish, shrimp, and crabs. Through adaptive filtering based on action frequency, bidirectional collaborative locking based on apparent weight, and density-based hardware linkage adjustment, the system automatically uploads time-aligned structured data packets directly to the computer system, eliminating manual recording and improving the accuracy and automation efficiency of dynamic live animal measurement.

The invention provides an urban tree species classification and mapping method based on LiDAR and hyperspectral fusion. The method comprises the steps of multi-source data acquisition and collaborative preprocessing; vegetation area extraction and non-tree ground object masking; individual tree crown segmentation and contour optimization are carried out; performing multi-dimensional feature extraction and feature engineering; feature selection and classification model construction; multi-category tree species identification and classification result generation; carrying out precision verification and uncertainty analysis; the advantages of LiDAR data in the aspect of representing a canopy three-dimensional structure and the characteristics of hyperspectral data in the aspect of detecting vegetation biochemical parameters are fully utilized, information complementation is achieved through feature level fusion, and the classification precision is remarkably improved.