64 results about "Crop mapping" patented technology

The invention relates to a method and system for intelligent identification and accurate pesticides spraying using unmanned aerial vehicles. The method comprises the following steps: dividing a crop area into a plurality of subregions; controlling an unmanned surveillance aerial vehicle to acquire crop images in each subregion and feeding the images to a control center; automatically comparing the acquired crop images with prestored images of crops subjected to pest and disease damages to identify the presence of pest and disease damages in each subregion, and if pest and disease damages exist, controlling an unmanned spraying aerial vehicle to accurately spray corresponding proportions or corresponding types of pesticides to the corresponding subregions. For the method, pesticides are sprayed only to subregions with crops subjected to pest and disease damages instead of large area pesticide spraying in a full range, thus reducing the waste of pesticides and protecting ecological environment; in addition, corresponding proportions or corresponding types of pesticides are sprayed according to specific circumstances of pest and disease damages, the same ratio of pesticides is not adopted for different pest and disease damages that the crops are subjected to, so that effective spraying treatment effect can be achieved for the crops subjected to different pest and disease damages.

The invention relates to the field of plant protection unmanned planes, and provides an adaptive operation method for a plant protection unmanned plane based on an expert system. The method comprisesthe steps: (1), collecting a crop image of an operation area through an airborne camera; (2), performing image segmentation and feature extraction of the collected image, and analyzing the crop species, the weed species, the pest species and the severity thereof; 3), comparing the data with an expert system database in an airborne control system, and obtaining the amount of liquid medicine to be sprayed in a target area; (4), enabling an airborne liquid spraying mechanism to adaptively spray the liquid medicine according to the index of the amount of sprayed medicine given by the airborne expert system database. The method is advantageous in adaptively carrying out the plant protection spraying operation according to the index of the amount of sprayed medicine given by the airborne expertsystem in real time and the crop species, the weed species, the pest species and the severity thereof in the operation area, effectively improving the pesticide spraying effect and pesticide utilization rate, avoiding the spraying leakage or repeated spraying in the operation area, needing no manual intervention in the whole process, and greatly improving the intelligentization degree of the plantprotection operation process.

The invention discloses a green intelligent agricultural greenhouse planting environment monitoring and management system. The system comprises a region division module, an environment detection module, an environment analysis module, a crop planting module, a crop image acquisition module, a growth stage analysis module, a nutrient solution sampling module, a sample detection module, a remote control terminal, a cloud server and a storage database. Air environment parameter values of a plurality of subareas in the agricultural greenhouse are detected through the environment detection module;the air quality coefficient of each subarea is calculated.subareas suitable for the growth of the hydroponic crops are analyzed; the planting density of the subareas is properly increased; meanwhile,images of crops in all the subareas are collected through the crop image collection module, growth stages corresponding to the crops in all the subareas are analyzed, the content of components in thenutrient solution in all the subareas is detected, whether the nutrient solution in all the subareas is qualified or not is judged through comparison, the unqualified nutrient solution is replaced, and the growth quality of the hydroponic crops is improved.

The invention provides a crop growth trend analysis system comprising an image acquisition module, an internet of things module and an image processing module. The image acquisition module acquires a crop image and then transmits the crop image to the image processing module through the internet of things module. The image processing module performs grayscale processing on the crop image to obtain the pixel amount of each of the three primary colors of R, G and B of the crop image. The pixel amount of each color is compared with the total pixel amount of the image to obtain the duty ratio of each color. The image processing module extracts the crop contour from the crop image, and the image is de-weighted so that that area of the crop is obtained. The area proportion of the crop is obtained by comparing the area of the crop with the total area of the crop image, the monitoring of the growth state of the crop is realized, the growth environment of the crop is timely and automatically adjusted through the terminal equipment and the labor cost is reduced.

The invention discloses a crop disease image recognition method based on a convolutional neural network. The method comprises the following steps: S1, collecting crop images at fixed time and fixed point based on an unmanned aerial vehicle; S2, reading POS data carried in the crop image, realizing detection and positioning of a disease area in the crop image based on a Faster R-CNN model, and generating a disease area image set; S3, realizing detection and recognition of holes, spots, pest tracks and the like in the disease area image based on a DSSD_Xception_coco model; and S4, outputting a disease recognition result based on the detection recognition result of the holes, the spots, the pest tracks and the like and the POS data of the corresponding disease area image, and completing disease condition statistics of each area. According to the invention, automatic detection, recognition and statistical analysis of crop diseases are realized, a corresponding control scheme is further provided, and a foundation is laid for improving crop disease early warning.

The invention discloses a crop disease identification method based on a remote sensing image. The method comprises the following steps: S1, realizing detection of a crop target in the remote sensing image based on a YOLT model; S2, digging a crop image area corresponding to the crop target in the remote sensing image based on a detection result, and generating a crop image; S3, removing crop occlusion information in the crop image based on a disease image occlusion removal algorithm of the thermodynamic diagram; S4, acquiring the saliency map of the crop image by using a saliency map detectionstrategy based on a saliency map disease image segmentation method, and performing complex background segmentation on the crop image by taking the saliency map as a mask image; and S5, realizing detection and identification of holes, spots, pest tracks and the like in the crop image based on the DSSD_Inception-V2_co model. According to the invention, rapid identification and statistical analysisof crop diseases can be realized.

The invention belongs to the technical field of agricultural control, and discloses a water-saving irrigation control system and a control method. The water-saving irrigation control system comprises an agricultural data acquisition module, an environmental data acquisition module, a stored water quantity acquisition module, an image acquisition module, a central control module, an image processing module, a crop image extraction module, an analysis module, an irrigation determination module, an irrigation parameter calculation module, a judgment module, an irrigation module and a recovery module. According to the water-saving irrigation control system provided by the invention, crop images are extracted by collecting environment images of a planting area, crop states are analyzed and processed, whether crops need to be irrigated or not is intelligently judged by combining environment temperature and humidity and soil temperature and humidity data, irrigation parameters are automatically calculated based on the corresponding data, and irrigation is carried out, so that water resources can be saved, the inaccuracy of manual judgment is eliminated, and the intelligent degree of water-saving irrigation control is improved.

The invention relates to the technical field of intelligent agriculture, and discloses an intelligent crop irrigation method, device and equipment and a storage medium. The intelligent crop irrigation method comprises the following steps that first-class region environment information of a region where crops are located is obtained, region division is carried out on the region where the crops are located according to the first-class region environment information, and different crop irrigation regions are obtained; crop images of the crop irrigation regions are obtained, and image recognition is carried out on the crop images to obtain crop attribute information of crops; second-class region environment information of the crop irrigation regions is obtained, and the water demand and fertilizer demand corresponding to the crop irrigation regions are obtained according to the crop attribute information and the second-class region environment information; and a corresponding crop irrigation mode according to the water demand and the fertilizer demand is determined, and the crop irrigation regions are irrigated according to the crop irrigation mode, so that the intelligent degree and irrigation precision are improved during crop water and fertilizer irrigation, and unnecessary water and fertilizer resource waste is reduced.

The invention relates to the technical field of weed detection, and particularly discloses a weed detection method and device based on crop growth characteristics, and the method comprises the steps: obtaining images of all plants in a field and the growth period of crops, and obtaining a to-be-detected image library; inputting the to-be-detected image library into a trained first classifier, carrying out content identification on the images, and obtaining a weed image according to a content identification result; uploading the weed image to a detection image, manually marking the detection image, and storing the image in a first sample library and a third sample library in a classified manner according to a manual marking result; and inputting the weed image into a trained second classifier, performing feature recognition on the image, removing a crop image in the weed image according to a feature recognition result, and reporting a detection result. According to the method, growth characteristics of crops in each growth period are learned through a neural network algorithm, weeds are detected through a reverse weed detection and recognition method for recognizing non-crops, and the recognition accuracy is greatly improved.

The invention relates to a crop disease control scheme recommendation method and a device, a system, equipment and a medium, relates to the technical field of machine learning, and can be applied to ascene of determining a corresponding disease control scheme according to a crop image. The crop disease control scheme recommendation method comprises the steps of obtaining a plurality of to-be-processed images; wherein the to-be-processed images comprise a crop image; inputting the crop image into a feature classification model to determine classification feature data corresponding to the cropimage; wherein the feature classification model is generated based on a deep residual network; inputting the classification feature data into a result recommendation model to determine a prevention and treatment scheme corresponding to the classification feature data; and receiving feedback information for the prevention and control scheme, and adjusting the weights of the parameters of the deep residual network according to the feedback information to optimize the structure of the deep residual network. According to the method, the deep residual network is introduced, so that the disease control scheme corresponding to the crop image can be obtained more accurately.

The invention relates to the technical field of crop image acquisition. The invention discloses a greenhouse crop three-dimensional directed perception and fine-grained automatic collection device andmethod. The device comprises a first telescoping rod which is horizontally arranged, a second telescoping rod which is vertically arranged, a first hanger rail which is horizontally arranged, and a second hanger rail which is vertically arranged. The first hanger rail is slidably mounted on an arch frame, and the upper end of the second hanger rail is slidably mounted on a cantilever beam; one end of the first telescopic rod is slidably mounted on the first hanger rail, and the other end is slidably mounted on the second hanger rail; the upper end of the second telescopic rod is slidably installed on the first telescopic rod, and the lower end is rotatably provided with a camera facing a cylinder. According to the greenhouse crop three-dimensional directed sensing and fine-grained automatic acquisition device provided by the invention, the target crop is abstracted into a slender cylinder in a three-dimensional space, a directed sensing model conforming to full-target coverage in a real environment is established, and real coverage of a monitoring target can be realized.

The invention provides a multi-unmanned-aerial-vehicle cooperative farmland spraying method based on a deep learning algorithm. The method comprises the steps: uniformly dividing a farmland into N small regions with the same sizes and collecting a crop image in each small region through a camera on an exploration machine; inputting the crop images into a neural network model in an intelligent controller on the exploration machine and outputting the category of the crop in each small region; classifying the small regions corresponding to the same categories into one group by the exploration machine, planning an optimal path by using a genetic algorithm according to the coordinates of all groups of small regions, and sending the optimal path to target machines; and then executing the operation corresponding to the categories by the target machines based on the optimal path, making mutual cooperation between the target machines, and returning coordinate information to the exploration machine in real time based on a communication protocol. According to the invention, the pesticide spraying is performed by using the coordinated control method of the unmanned aerial vehicle; the orderliness and efficiency are improved; and resources are saved.

The invention provides a fertilization method. The fertilization method comprises the following steps: acquiring a current crop image; comparing the crop image with a plurality of images in a pre-stored image database to obtain a plurality of matching degrees, wherein each image in the pre-stored image database corresponds to a fertilization node; determining an image corresponding to the maximummatching degree in the plurality of matching degrees as a target image; determining a fertilization node corresponding to the crop according to the corresponding relationship between the image and thefertilization node and the target image, wherein the fertilization node is before a certain growth period of the crop; fertilization information is obtained according to the fertilization nodes, andthe crops are fertilized according to the fertilization information. The invention further provides a computer readable storage medium. According to the fertilization method provided by the invention,the fertilization nodes are matched through the crop images, and the crops are fertilized according to the fertilization nodes, so that the method is very intelligent and effective.

The invention relates to the field of agriculture, in particular to an intelligent agricultural internet-of-things system, and aims to solve the problem that an existing agricultural internet-of-things system cannot well learn nutrients needing to be supplemented for crops under current soil conditions and growth stages according to different crop growth stages and different soil conditions. The intelligent agricultural internet-of-things system provided by the invention comprises a cloud server and an agricultural greenhouse, and further comprises a database, a crop image acquisition device and a soil detection device, wherein the cloud server is wirelessly connected with the database, the crop image acquisition device and the soil detection device; and the cloud server comprises a crop image analysis module and a soil analysis module.

The invention discloses an unmanned water chestnut harvester based on visual navigation, and relates to the technical field of water chestnut harvesting; the unmanned water chestnut harvester comprises: a navigation mechanism that comprises a base and a tractor, wherein the top of the base is fixedly provided with a shell and an adjusting assembly, one side of the shell is provided with a telemetering assembly and a microcontroller, and the base is installed at the top end of the tractor; the driving mechanism that comprises a first driving shaft connected with the output end of the tractor and a transmission assembly driven by the first driving shaft; and the screening mechanism that is installed at the tail of the tractor and comprises a supporting frame of a frame type structure and tires for supporting the supporting frame to move, the supporting frame is used for supporting the transmission assembly and the screening assembly, and the screening assembly is connected with the transmission assembly. By arranging the microcontroller, the position coordinates and speed information of the tractor are obtained through the automatic navigator arranged on the microcontroller, the tractor walks according to a set path, images of cultivated crops are observed in real time through the telemetering camera and the monitoring camera, and the navigation precision is improved.

The invention discloses a real-time identification system for crops based on an unmanned aerial vehicle. The system comprises an unmanned aerial vehicle platform, a route planning module, a crop identification module, a server and a mobile terminal, wherein the unmanned aerial vehicle platform is provided with an image acquisition device and the crop identification module, and the route planning module is arranged in the server. The invention further discloses a real-time identification method for crops based on the unmanned aerial vehicle. According to the method, after the server plans the route of the unmanned aerial vehicle, the route is uploaded to the mobile terminal, and the unmanned aerial vehicle can execute crop image collection and real-time crop identification tasks after receiving an instruction, and upload the obtained crop image and a final identification result to an agricultural condition database. The crop identification module is realized based on a crop identification model trained by YOLOv3.

The invention relates to an automatic harvesting method for lodging crops. The automatic harvesting method comprises the following steps of S100, acquiring an on-site crop image; S200, analyzing the on-site crop image, and judging whether the crops fall down or not; S300, generating an optimal control instruction according to a recognition result and transmitting the optimal control instruction toan execution component; S400, automatically adjusting corresponding parameters according to the control instruction and executing harvesting operation by the execution component; and S500, repeatingthe steps S100-S400 until the harvesting operation is finished. The invention further discloses an automatic harvester for the lodging crops. According to the automatic harvesting method and the automatic harvester, a lodging crop identification visual model is established by using a deep learning technology, so that the lodging condition of the crops on site can be accurately and automatically identified in real time, the computing power requirement is low, analysis can be completed on common terminal computing equipment, implementation is simple, only a camera device and a vehicle-mounted computing terminal need to be additionally arranged outside an agricultural machine to form an external module, and the agricultural machine is not changed at all.

The invention relates to the technical field of cereal crop disease and insect pest recognition, in particular to a deep learning-based cereal crop disease and insect pest detection method, which comprises the following steps of: 1, establishing a disease and insect pest project database and a normal cereal crop model database; step 2, acquiring a to-be-identified field coarse cereal crop image based on a camera; step 3, preprocessing the obtained on-site coarse cereal crop image; 4, establishing a deep learning recognition model; 5, recognizing the preprocessed coarse cereal crop image through a deep learning recognition model; and step 6, the human-computer interaction interface displays an identification result. Calculation and analysis are carried out through the deep learning recognition model to obtain the disease and pest types of the coarse cereals; therefore, the problem that artificial sensory judgment is easily influenced by factors such as emotion, health and fatigue is effectively avoided.

The invention discloses an intelligent agricultural insecticidal system based on image recognition. The intelligent agricultural insecticidal system comprises an unmanned aerial vehicle terminal, a cloud terminal server and a client; the unmanned aerial vehicle terminal comprises a first image acquisition unit for acquiring crop images of crops and sending the crop images to a cloud end, a positioning unit for acquiring position data of the unmanned aerial vehicle terminal and sending the position data to the cloud end, a pesticide spraying unit, a flight control unit and an instruction receiving unit for receiving a decision instruction sent by the client to control the pesticide spraying unit and the flight control unit; the cloud terminal server comprises a cloud navigation unit for generating an initial flight control instruction and sending the initial flight control instruction to the unmanned aerial vehicle terminal to control the flight state of the unmanned aerial vehicle terminal, an auxiliary decision making unit for generating a cloud instruction suggestion according to the crop images and sending the cloud instruction suggestion to the client remotely connected with the cloud terminal server, a cloud storage unit and a cloud alarm unit. The intelligent agricultural insecticidal system provided by the technical scheme of the invention has the beneficial effects that through image acquisition by the unmanned aerial vehicle terminal and real-time analysis and processing by the cloud terminal server, the spraying operation of the unmanned aerial vehicle terminal can be rapidly adjusted by the client, and the problem that the timeliness and pertinence of pesticide spraying by an existing unmanned aerial vehicle are low is solved.

The invention discloses a method for crop mapping, and particularly relates to a method for crop mapping by using GaoFen-2 optical imaging and GaoFen-3 polarized synthetic aperture radar (POLSAR) databased on field combination. The advantages of multi-source remote sensing data are exploited fully, the problems of imprecise plot space and inaccurate crop mapping of remote sensing in crop mappingare solved, according to the study, a set of ''SAR-optics'' data collaborative crop planting structure remote sensing extraction models is developed by taking the space and time collaboration of SAR and optical remote sensing data as the point of penetration, multi-level collaboration is carried out on spatial information of optical images and characteristic information of SAR data to explore an automatic identification method of crop planting structure information based on multi-source remote sensing data supported by plot elements. The method mainly includes four steps of multi-scale segmentation of high spatial resolution images, SAR image characteristic extraction, optimal classification subset acquisition and object-oriented classification using SVM. According to the method, multi-temporal SAR (GaoFen-3) and optical imaging (GaoFen-2) are used as data sources, and collaboration processing is carried out by using "map" information of a cropland plot structure provided by GaoFen-2 images and polarization scattering, texture and other information of ground object characteristics provided by GaoFen-3 images so as to realize accurate identification of crop types and extraction of planting areas.

The embodiment of the invention provides a method and device for adjusting operation parameters of an agricultural machine, a processor, the agricultural machine and a storage medium, the method is applied to the agricultural machine, and the method comprises the steps that a to-be-detected crop image is obtained, and the to-be-detected crop image is obtained by shooting crops collected by the agricultural machine; inputting the to-be-detected crop image into a machine learning model, and analyzing the to-be-detected crop image through the machine learning model; obtaining a detection result output by the machine learning model, and determining an agricultural machinery evaluation index value according to the detection result; and adjusting the operation parameters of the agricultural machinery according to the evaluation index value of the agricultural machinery, so that the operation parameters of the agricultural machinery can be adjusted in time to improve the harvesting quality of crops.

The invention provides a facility tomato farming decision-making auxiliary method and device. The method comprises the following steps: acquiring a crop image in a facility through intelligent glasses; inputting the crop image into an improved SSD network model pre-stored in the intelligent glasses, outputting a part needing farming decision, and displaying the part in the intelligent glasses, wherein the improved SSD network model is obtained by performing training on a feature extraction network structure based on a MobileNet V3 network instead of an SSD network and based on a sample image with a determined farming decision part as a label, and the farming decision comprises any one or more of pruning, flower thinning, fruit thinning, leaf picking and ripe fruit picking. According to the method, images of parts such as fruits, flowers, branches and leaves of tomatoes in different periods are obtained in a lossless image mode, and recognition is performed through a lightweight classification model suitable for a mobile environment, so that the influence of equipment factors such as processing speed and storage scale of small mobile equipment on a recognition result is greatly reduced.

The invention provides a pest and disease identification and model training method and device, electronic equipment and a storage medium. The pest and disease identification method comprises the following steps: pre-training a to-be-trained feature extraction network by adopting a first crop image without a label to obtain a trained feature extraction network; training a to-be-trained disease and insect pest recognition model by adopting the second crop image with the label to obtain a trained disease and insect pest model; the to-be-trained pest recognition model comprises the trained feature extraction network and a to-be-trained multi-classifier. And recognizing the to-be-recognized image through the trained disease and pest recognition model to obtain a recognition result output by the disease and pest recognition model. According to the pest recognition method, pest recognition is performed on the to-be-recognized image through the trained pest recognition model, and the pest control effect is improved through the systematic pest recognition scheme.

The invention discloses a pasturing area drought identification method and system and a management platform. According to the pasturing area drought identification method, remote sensing data and image data shot by an unmanned aerial vehicle are combined, firstly, the drought condition is preliminarily judged through a drought index calculated through the remote sensing data, then crop images are shot through the unmanned aerial vehicle, and after a series of filtering processing, a model is input to obtain a drought grade result. According to the method, the system and the management platform provided by the invention, the labor cost can be saved, and the drought identification result with higher accuracy can be obtained.

The invention discloses a greenhouse crop disease control method based on real-time feedback, which specifically comprises the following steps of: acquiring image information of greenhouse crops by using an image recognition technology, and relates to the technical field of greenhouse crop disease control. According to the greenhouse crop disease prevention and control method based on real-time feedback, image information of greenhouse crops is subjected to real-time data acquisition by utilizing an image recognition technology, and a central computer compares, analyzes and judges the acquiredcrop image information with standard disease data information in a disease database; the disease types and the disease levels of the greenhouse crops are obtained; the environmental parameters in thegreenhouse are regulated and controlled according to the disease types and the disease levels of the greenhouse crops; and the diseases are treated in time by adopting biological and physical methods. According to the greenhouse crop disease control method, the timeliness of crop disease control can be greatly improved, the disease judgment accuracy is high, and the labor cost consumption is effectively reduced.

The invention discloses an intelligent fine management control system applied to intelligent agriculture. The system comprises a data acquisition module, a local control system, a local database, a local UI terminal, a cloud server, a cloud database, a remote UI terminal and a machine vision algorithm module. The local control system regularly acquires the environmental data and the crop image data acquired by the data acquisition module, stores the environmental data and the crop image data in the local database, and pushes the environmental data and the crop image data to the local UI terminal for display and query; at the UI terminal, a manager sets a management rule by himself / herself as required, and the system automatically controls the control equipment in the planting environment according to the management rule; meanwhile, the collected data is uploaded to a cloud end at regular time; at the remote UI terminal, a remote manager combines a control instruction and a recognitionresult of a machine vision algorithm module on the crop health condition, synchronously returns the recognition result to the local control system, and automatically controls crops or guides the manager to perform corresponding operation guidance through control equipment.

The invention relates to an image data quality control method and system in crop live-action observation, and the method comprises the following steps: generating a gray value of an image missing graytone and an image incomplete rate of a corresponding image missing gray tone according to a historical image; Identifying an image missing amount of the to-be-detected image and an image missing ratecorresponding to the image missing amount according to the gray value; And comparing the image missing rate with the image incomplete rate. According to the technical scheme, the incomplete image inthe to-be-detected image is identified and eliminated by utilizing the color characteristic parameters of the crop image, namely the grey tone when the historical image is lost, so that the complete image is provided for subsequent calculation of the crop coverage and the leaf area index, and the calculation accuracy is improved.

Methods and systems are disclosed configured to control the planting, application of pesticides, and harvesting of greenhouse crops, such as herbs. The greenhouse may include a variety of sensors, such as moisture sensors, ph sensors, and / or CO2 sensors. Unmanned vehicles may be utilized to capture crop images, and a learning engine may be used to determine the size of greenhouse crops. Such sensor data may be used to predict crop availability. A predication engine may be utilized to predict demand for greenhouse crops using current and historical orders for greenhouse crops. Greenhouse crop production instructions may be generated and transmitted to a greenhouse computer system to cause crops to be sown or harvested. Pallet loading instructions may be generated regarding the loading of specified quantities of crop packs on respective pallets for shipment to a destination.