CN116188872A - Automatic forestry plant diseases and insect pests identification method and device - Google Patents
Automatic forestry plant diseases and insect pests identification method and device Download PDFInfo
- Publication number
- CN116188872A CN116188872A CN202310244263.6A CN202310244263A CN116188872A CN 116188872 A CN116188872 A CN 116188872A CN 202310244263 A CN202310244263 A CN 202310244263A CN 116188872 A CN116188872 A CN 116188872A
- Authority
- CN
- China
- Prior art keywords
- image
- pest
- insect pests
- forestry
- plant diseases
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a forestry plant diseases and insect pests automatic identification method and device, wherein the method comprises the following steps: acquiring a target image, wherein the target image comprises at least one image information of a plant disease and insect pest to be identified; inputting the target image into a pre-trained recognition model to obtain a disease and pest recognition result; the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample. The technical problems of high pest and disease damage identification cost and poor accuracy of identification results in the prior art are solved.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an automatic identification method and device for forestry diseases and insect pests.
Background
In the process of construction of the forestry ecological environment, plants can be subjected to double influences of natural disasters and plant diseases and insect pests, and the plant growth is affected more frequently by the plant diseases and insect pests, so that plant dysplasia and even death can be caused, the forest ecological system structure is affected, and great losses are caused to the forestry economy. In order to effectively control diseases and insect pests, the traditional forestry management inputs a great deal of manpower, material resources and financial resources for on-site investigation and collection of the diseases and insect pests, but the traditional method has higher manpower cost, and the identification result of the diseases and insect pests is greatly influenced by the professional knowledge and experience of personnel,
in order to reduce labor cost, the prior art provides a multi-spectrum-based disease and pest detection and identification method, which utilizes a multi-spectrum camera carried by an unmanned aerial vehicle to collect multi-spectrum images of vegetation in a detection area, then analyzes the multi-spectrum images to obtain vegetation indexes, and further analyzes vegetation health conditions. However, in the method in the prior art, the unmanned aerial vehicle is utilized to acquire and analyze the multispectral image to acquire the health condition of the vegetation, the spectrum is greatly affected by the surrounding environment, the analysis precision is not high, and the disease and pest analysis by the health condition of the vegetation is still greatly affected by personal expertise and experience, so that the accuracy of the identification result is poor.
Therefore, the automatic identification method and device for the forestry diseases and insect pests are provided, so that the technical problems of high disease and insect pest identification cost and poor accuracy of identification results in the prior art are solved, and the technical problems to be solved by the technicians in the field are solved urgently.
Disclosure of Invention
Therefore, the embodiment of the invention provides a forestry plant diseases and insect pests automatic identification method and device, which aim to at least solve the technical problems of higher plant diseases and insect pests identification cost and poorer accuracy of identification results in the prior art.
In order to achieve the above object, the embodiment of the present invention provides the following technical solutions:
the invention provides an automatic forestry plant diseases and insect pests identification method, which comprises the following steps:
acquiring a target image, wherein the target image comprises at least one image information of a plant disease and insect pest to be identified;
inputting the target image into a pre-trained recognition model to obtain a disease and pest recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
In some embodiments, training is performed by using pest sample data based on a pre-built deep learning network to obtain the identification model, which specifically includes:
collecting image information of a large number of pest and disease samples, and preprocessing all the image information to obtain the image samples;
respectively carrying out category labeling on the image samples to obtain plant diseases and insect pests sample data comprising the image samples and category labels corresponding to the image samples;
and inputting the plant disease and insect pest sample data into a pre-constructed deep learning network for training to obtain the identification model.
In some embodiments, collecting image information of a mass of pest samples, and preprocessing all the image information to obtain the image samples, specifically including:
performing multi-dimensional image acquisition on the captured living bodies of the plant diseases and insect pests, and constructing a forestry plant diseases and insect pests data set by utilizing the acquired images;
and carrying out image processing on all the images in the forestry plant diseases and insect pests data set to obtain the image sample.
In some embodiments, the image processing at least includes batch matting all the acquired images in the forestry pest data set, and automatically cropping a target area containing the pest in the images by using a projection algorithm.
In some embodiments, image processing is performed on all images in the forestry pest data set to obtain the image sample, and then further comprising:
scaling and randomly rotating the image samples according to different proportions, and fusing any combination of various types and a plurality of background template images formed by the image samples to ensure the randomness of the distribution of the plant diseases and insect pests in the fused images, the possibility of mutual overlapping shielding and the authenticity of the plant diseases and insect pests under a specific distance, thereby generating massive plant diseases and insect pests sample data.
In some embodiments, the network architecture of the pre-built deep learning network includes: backbone network BoTNet-101, feature fusion network FPN, and head detection network GFLHead in mmdetection;
wherein, the backbone network BotNet-101 is added with multi-head self-attention MHSA based on ResNet-101;
the feature fusion network FPN unifies the features of different layers and different channel numbers of the BotNet-101 into 256-channel number features;
the head detection network GFLHead integrates the quality focus loss function QFL to optimize the class-quality joint score continuous value labels and the distribution focus function DFL to optimize any probability distribution of the bounding box.
In some embodiments, training with pest sample data based on a pre-built deep learning network to obtain the identification model further comprises:
the recognition model is deployed in a system program to facilitate retrieval.
The invention also provides an automatic forestry plant diseases and insect pests identification device, which comprises:
the image acquisition unit is used for acquiring a target image, wherein the target image comprises at least one image information of the plant diseases and insect pests to be identified;
the image recognition unit is used for inputting the target image into a pre-trained recognition model so as to obtain a plant disease and insect pest recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method as described above when executing the program.
The invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method as described above.
According to the forestry plant diseases and insect pests automatic identification method provided by the invention, the target image is obtained, the target image contains at least one type of image information of plant diseases and insect pests to be identified, and the target image is input into a pre-trained identification model, so that a plant diseases and insect pests identification result can be obtained rapidly and accurately; the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample. The invention is based on AI technology and image recognition technology, uses a self-caught large number of different insect pests as a recognition model for automatically recognizing forestry insect pests, which is obtained by sample training, to perform on-line automatic recognition of forestry insect pests, can save a great deal of time and physical strength of garden staff, improves working efficiency, and has extremely important effect on pest control. Therefore, the technical problems of high pest and disease damage identification cost and poor accuracy of identification results in the prior art are solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
The structures, proportions, sizes, etc. shown in the present specification are shown only for the purposes of illustration and description, and are not intended to limit the scope of the invention, which is defined by the claims, so that any structural modifications, changes in proportions, or adjustments of sizes, which do not affect the efficacy or the achievement of the present invention, should fall within the ambit of the technical disclosure.
FIG. 1 is a schematic flow chart of an automatic identification method for forestry diseases and insect pests provided by the invention;
FIG. 2 is a second flow chart of the method for automatically identifying forestry plant diseases and insect pests provided by the invention;
FIG. 3 is a third flow chart of the method for automatically identifying forestry plant diseases and insect pests provided by the invention;
figure 4 is a schematic structural view of the forestry plant diseases and insect pests automatic identification device provided by the invention;
fig. 5 is a block diagram of a computer device according to the present invention.
Detailed Description
Other advantages and advantages of the present invention will become apparent to those skilled in the art from the following detailed description, which, by way of illustration, is to be read in connection with certain specific embodiments, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order to solve the problems in the prior art, the invention realizes the forestry plant diseases and insect pests automatic identification method by utilizing the self-captured mass plant diseases and insect pests sample data and AI image identification technology according to the actual application scene of the garden plant diseases and insect pests control work. According to the invention, a large number of living bodies of diseases and insect pests in different forestry management areas are captured by using a field black light lamp, the diseases and insect pests are identified and classified through priori experience, then research and development personnel take pictures of all angles of the diseases and insect pests in a laboratory environment to acquire images, then an image segmentation technology is used for acquiring high-precision disease and insect pest images, massive disease and insect pest sample data can be acquired by rotating the disease and insect pest images in multiple directions, finally the established disease and insect pest identification model is used for training and optimizing the disease and insect pest identification model, and the disease and insect pest identification model is deployed into a designed automatic identification device.
Referring to fig. 1, fig. 1 is a schematic flow chart of an automatic identifying method for forestry diseases and insect pests provided by the invention.
In a specific embodiment, the invention provides an automatic forestry plant diseases and insect pests identification method, which comprises the following steps:
s110: acquiring a target image, wherein the target image comprises at least one image information of a plant disease and insect pest to be identified;
s120: inputting the target image into a pre-trained recognition model to obtain a disease and pest recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
Specifically, as shown in fig. 2, training is performed by using plant disease and insect pest sample data based on a pre-constructed deep learning network to obtain the identification model, and the method specifically comprises the following steps:
s210: and collecting image information of a large number of pest samples, and preprocessing all the image information to obtain the image samples. When the pest samples are collected, the living bodies of the forestry pests can be induced in the field lamp, particularly the black light lamp can emit ultraviolet rays invisible to human eyes, the pest induction effect is very strong, and a large number of different types of living bodies of the pests can be collected in different forestry areas of Beijing from 3 to 10 months by utilizing the phototactic property of the forestry pests.
S220: and respectively carrying out category labeling on the image samples to obtain plant diseases and insect pests sample data comprising the image samples and category labels corresponding to the image samples. The prior experience can be used for identifying and classifying the plant diseases and insect pests, and particularly, in order to ensure the scientificity of a sample library, a plurality of experts for preventing and controlling the plant diseases and insect pests in forestry are invited to identify and classify a large number of collected plant diseases and insect pests. The pest samples can comprise 32 kinds of fall webworm, yang Er-tail armyworm, liu Due, pine moth, citrus fruit fly, chinese scholartree inchworm, longicorn beetle, meadow caterpillar, yellow thorn moth, chinese pine caterpillar and the like, and the body length intervals of various pests are recorded respectively and are used for later sample library production.
S230: and inputting the plant disease and insect pest sample data into a pre-constructed deep learning network for training to obtain the identification model.
In the step S210, as shown in fig. 3, image information of a large number of pest samples is collected, and all the image information is preprocessed to obtain the image samples, which specifically includes the following steps:
s310: and carrying out multi-dimensional image acquisition on the captured living bodies of the plant diseases and insect pests, and constructing a forestry plant diseases and insect pests data set by utilizing the acquired images. When the plant diseases and insect pests image acquisition is carried out, a 1200 ten thousand resolution camera can be used for carrying out image acquisition on all captured plant diseases and insect pests living bodies from a plurality of angles, different distances, different illumination, different living states and the like in a laboratory, so that a mass forestry plant diseases and insect pests data set is established.
S320: and carrying out image processing on all the images in the forestry plant diseases and insect pests data set to obtain the image sample. The image processing at least comprises batch matting of all the acquired images in the forestry plant diseases and insect pests data set, and automatic cutting of target areas containing plant diseases and insect pests in the images by utilizing a projection algorithm. In a specific use scene, when image processing is carried out, the high-precision image of the plant diseases and insect pests is segmented, namely, an improved image segmentation algorithm Background Matting V is utilized, a backbone network adopts BoTNet with a self-attention mechanism, the obtained high-precision plant diseases and insect pests images are rapidly and batched, the range of the plant diseases and insect pests is automatically cut by utilizing a projection algorithm, a standard plant diseases and insect pests atlas is obtained, and as the quality of a sample library is directly influenced by the image matting effect, edge details can be distinguished from each image matting.
S330: scaling and randomly rotating the image samples according to different proportions, and fusing any combination of various types and a plurality of background template images formed by the image samples to ensure the randomness of the distribution of the plant diseases and insect pests in the fused images, the possibility of mutual overlapping shielding and the authenticity of the plant diseases and insect pests under a specific distance, thereby generating massive plant diseases and insect pests sample data.
In other words, when the synthesis of the pest sample library is performed, in order to increase the robustness of the training sample set and the recognition model, each segmented pest image is scaled and randomly rotated according to different proportions, then various types and a plurality of background template images are combined randomly, meanwhile, the randomness of the distribution of the pest in the fused image, the possibility of overlapping and shielding each other and the authenticity of the pest size at a specific distance are ensured, and a mass sample library and corresponding labels are generated.
The network architecture of the pre-built deep learning network comprises the following components: backbone BoTNet-101, feature fusion network FPN, and GFLHead in mmdetection. The main network BotNet-101 is additionally provided with a multi-head self-attention MHSA on the basis of ResNet-101, the FPN integrates the characteristics of different layers and different channel numbers of the BotNet-101 into 256-channel number characteristics, and the GFLHead integrates a quality focus loss function QFL to optimize the classification-quality joint score continuous value label and a distribution focus function DFL to optimize any probability distribution of a boundary box.
When the pest detection model is trained and optimized based on the network architecture, a mass of pest detection training sample libraries and labels are firstly established, a main network BoTNet with a self-focusing mechanism is used for improving a target detection algorithm GFL, a main network ResNeXt-101 of the GFL is replaced by a BotNet-101, the network is trained, the levels and the nodes are continuously adjusted to obtain better fitting, then the pest detection sample libraries are used for detection, the model is adjusted and optimized according to the detection result, and the final model recognition rate reaches more than 90%.
Further, training by using plant disease and insect pest sample data based on a pre-built deep learning network to obtain the identification model, and then further comprising:
the recognition model is deployed in a system program to facilitate retrieval. When the plant diseases and insect pests detection model is deployed, for example, a trained forestry plant diseases and insect pests automatic identification model is deployed into a plant diseases and insect pests management system for use, 32 plant diseases and insect pests can be automatically identified through the extracted identification model, and the accuracy is more than 90%.
In the specific embodiment, the forestry plant diseases and insect pests automatic identification method provided by the invention can quickly and accurately obtain the plant diseases and insect pests identification result by acquiring the target image, wherein the target image contains at least one type of image information of plant diseases and insect pests to be identified, and inputting the target image into a pre-trained identification model; the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample. The invention is based on AI technology and image recognition technology, uses a self-caught large number of different insect pests as a recognition model for automatically recognizing forestry insect pests, which is obtained by sample training, to perform on-line automatic recognition of forestry insect pests, can save a great deal of time and physical strength of garden staff, improves working efficiency, and has extremely important effect on pest control. Therefore, the technical problems of high pest and disease damage identification cost and poor accuracy of identification results in the prior art are solved.
In addition to the method, the invention also provides an automatic forestry plant diseases and insect pests identification device, as shown in figure 4, which comprises:
an image acquisition unit 410, configured to acquire a target image, where the target image includes at least one image information of a pest to be identified;
an image recognition unit 420, configured to input the target image into a pre-trained recognition model, so as to obtain a pest and disease damage recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
In some embodiments, training is performed by using pest sample data based on a pre-built deep learning network to obtain the identification model, which specifically includes:
collecting image information of a large number of pest and disease samples, and preprocessing all the image information to obtain the image samples;
respectively carrying out category labeling on the image samples to obtain plant diseases and insect pests sample data comprising the image samples and category labels corresponding to the image samples;
and inputting the plant disease and insect pest sample data into a pre-constructed deep learning network for training to obtain the identification model.
In some embodiments, collecting image information of a mass of pest samples, and preprocessing all the image information to obtain the image samples, specifically including:
performing multi-dimensional image acquisition on the captured living bodies of the plant diseases and insect pests, and constructing a forestry plant diseases and insect pests data set by utilizing the acquired images;
and carrying out image processing on all the images in the forestry plant diseases and insect pests data set to obtain the image sample.
In some embodiments, the image processing at least includes batch matting all the acquired images in the forestry pest data set, and automatically cropping a target area containing the pest in the images by using a projection algorithm.
In some embodiments, image processing is performed on all images in the forestry pest data set to obtain the image sample, and then further comprising:
scaling and randomly rotating the image samples according to different proportions, and fusing any combination of various types and a plurality of background template images formed by the image samples to ensure the randomness of the distribution of the plant diseases and insect pests in the fused images, the possibility of mutual overlapping shielding and the authenticity of the plant diseases and insect pests under a specific distance, thereby generating massive plant diseases and insect pests sample data.
In some embodiments, the network architecture of the pre-built deep learning network includes: backbone BoTNet-101, feature fusion network FPN, GFLHead in mmdetection. The main network BotNet-101 is additionally provided with a multi-head self-attention MHSA on the basis of ResNet-101, the FPN integrates the characteristics of different layers and different channel numbers of the BotNet-101 into 256-channel number characteristics, and the GFLHead integrates a quality focus loss function QFL to optimize the classification-quality joint score continuous value label and a distribution focus function DFL to optimize any probability distribution of a boundary box.
In some embodiments, training with pest sample data based on a pre-built deep learning network to obtain the identification model further comprises:
the recognition model is deployed in a system program to facilitate retrieval.
In the specific embodiment, the forestry plant diseases and insect pests automatic identification device provided by the invention can quickly and accurately obtain the plant diseases and insect pests identification result by acquiring the target image, wherein the target image contains at least one type of image information of plant diseases and insect pests to be identified, and inputting the target image into a pre-trained identification model; the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample. The invention is based on AI technology and image recognition technology, uses a self-caught large number of different insect pests as a recognition model for automatically recognizing forestry insect pests, which is obtained by sample training, to perform on-line automatic recognition of forestry insect pests, can save a great deal of time and physical strength of garden staff, improves working efficiency, and has extremely important effect on pest control. Therefore, the technical problems of high pest and disease damage identification cost and poor accuracy of identification results in the prior art are solved.
In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 5. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and model predictions. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The model predictions of the computer device are used to store static information and dynamic information data. The network interface of the computer device is used for communicating with an external terminal through a network connection. Which computer program, when being executed by a processor, carries out the steps of the above-mentioned method embodiments.
It will be appreciated by those skilled in the art that the structure shown in FIG. 5 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
Corresponding to the above embodiments, the present invention further provides a computer storage medium, which contains one or more program instructions. Wherein the one or more program instructions are for being executed with the method as described above.
The present invention also provides a computer program product comprising a computer program storable on a non-transitory computer readable storage medium, the computer program being capable of performing the above method when being executed by a processor.
In the embodiment of the invention, the processor may be an integrated circuit chip with signal processing capability. The processor may be a general purpose processor, a digital signal processor (Digital Signal Processor, DSP for short), an application specific integrated circuit (Application Specific f ntegrated Circuit ASIC for short), a field programmable gate array (FieldProgrammable Gate Array FPGA for short), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components.
The disclosed methods, steps, and logic blocks in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be embodied directly in the execution of a hardware decoding processor, or in the execution of a combination of hardware and software modules in a decoding processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The processor reads the information in the storage medium and, in combination with its hardware, performs the steps of the above method.
The storage medium may be memory, for example, may be volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory.
The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable ROM (Electrically EPROM, EEPROM), or a flash Memory.
The volatile memory may be a random access memory (Random Access Memory, RAM for short) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (Double Data RateSDRAM), enhanced SDRAM (ESDRAM), synchronous DRAM (SLDRAM), and direct memory bus RAM (directracram, DRRAM).
The storage media described in embodiments of the present invention are intended to comprise, without being limited to, these and any other suitable types of memory.
Those skilled in the art will appreciate that in one or more of the examples described above, the functions described in the present invention may be implemented in a combination of hardware and software. When the software is applied, the corresponding functions may be stored in a computer-readable medium or transmitted as one or more instructions or code on the computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.
The foregoing detailed description of the invention has been presented for purposes of illustration and description, and it should be understood that the foregoing is by way of illustration and description only, and is not intended to limit the scope of the invention.
Claims (10)
1. An automatic identification method for forestry diseases and insect pests, which is characterized by comprising the following steps:
acquiring a target image, wherein the target image comprises at least one image information of a plant disease and insect pest to be identified;
inputting the target image into a pre-trained recognition model to obtain a disease and pest recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
2. An automatic identification method for forestry diseases and insect pests according to claim 1, wherein training is performed by using disease and insect pest sample data based on a pre-constructed deep learning network to obtain the identification model, specifically comprising:
collecting image information of a large number of pest and disease samples, and preprocessing all the image information to obtain the image samples;
respectively carrying out category labeling on the image samples to obtain plant diseases and insect pests sample data comprising the image samples and category labels corresponding to the image samples;
and inputting the plant disease and insect pest sample data into a pre-constructed deep learning network for training to obtain the identification model.
3. The automatic identification method of forestry plant diseases and insect pests according to claim 2, wherein collecting image information of a large number of plant diseases and insect pests samples, and preprocessing all the image information to obtain the image samples, comprises the following steps:
performing multi-dimensional image acquisition on the captured living bodies of the plant diseases and insect pests, and constructing a forestry plant diseases and insect pests data set by utilizing the acquired images;
and carrying out image processing on all the images in the forestry plant diseases and insect pests data set to obtain the image sample.
4. A method for automatically identifying a forestry pest according to claim 3, wherein the image processing comprises at least batch matting of all images in the obtained forestry pest data set, and automatically cropping a target area containing the pest in the images by using a projection algorithm.
5. A method of automatically identifying forestry pests as defined in claim 3, wherein all images in the forestry pest data set are image-processed to obtain the image sample, and further comprising:
scaling and randomly rotating the image samples according to different proportions, and fusing any combination of various types and a plurality of background template images formed by the image samples to ensure the randomness of the distribution of the plant diseases and insect pests in the fused images, the possibility of mutual overlapping shielding and the authenticity of the plant diseases and insect pests under a specific distance, thereby generating massive plant diseases and insect pests sample data.
6. An automatic identification method of forestry diseases and insect pests according to claim 2, wherein the network architecture of the pre-constructed deep learning network comprises: a backbone network BoTNet-101, a feature fusion network FPN and a head detection network GFLHead in mmdetection;
wherein, the backbone network BotNet-101 is added with multi-head self-attention MHSA based on ResNet-101;
the feature fusion network FPN unifies the features of different layers and different channel numbers of the BotNet-101 into 256-channel number features;
the head detection network GFLHead integrates the quality focus loss function QFL to optimize the class-quality joint score continuous value labels and the distribution focus function DFL to optimize any probability distribution of the bounding box.
7. An automatic identification method of forestry pests according to claim 2, wherein training is performed by using pest sample data based on a pre-constructed deep learning network to obtain the identification model, and further comprising:
the recognition model is deployed in a system program to facilitate retrieval.
8. An automatic identification device for forestry diseases and insect pests, characterized in that the device comprises:
the image acquisition unit is used for acquiring a target image, wherein the target image comprises at least one image information of the plant diseases and insect pests to be identified;
the image recognition unit is used for inputting the target image into a pre-trained recognition model so as to obtain a plant disease and insect pest recognition result;
the recognition model is obtained by training by using pest sample data based on a pre-built deep learning network, wherein the pest sample data comprises an image sample and a category label corresponding to the image sample.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 7 when the program is executed.
10. A non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310244263.6A CN116188872A (en) | 2023-03-09 | 2023-03-09 | Automatic forestry plant diseases and insect pests identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310244263.6A CN116188872A (en) | 2023-03-09 | 2023-03-09 | Automatic forestry plant diseases and insect pests identification method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116188872A true CN116188872A (en) | 2023-05-30 |
Family
ID=86440416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310244263.6A Pending CN116188872A (en) | 2023-03-09 | 2023-03-09 | Automatic forestry plant diseases and insect pests identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116188872A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117253122A (en) * | 2023-11-17 | 2023-12-19 | 云南大学 | Corn seed approximate variety screening method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112215293A (en) * | 2020-10-20 | 2021-01-12 | 平安国际智慧城市科技股份有限公司 | Plant disease and insect pest identification method and device and computer equipment |
| US20210150283A1 (en) * | 2019-11-15 | 2021-05-20 | Salesforce.Com, Inc. | Noise-Resistant Object Detection with Noisy Annotations |
| CN112989969A (en) * | 2021-02-26 | 2021-06-18 | 清远市智慧农业农村研究院 | Crop pest and disease identification method and device |
| CN113469177A (en) * | 2021-06-30 | 2021-10-01 | 河海大学 | Drainage pipeline defect detection method and system based on deep learning |
| US20220058420A1 (en) * | 2020-08-22 | 2022-02-24 | Tsinghua University | Scene text detection method and system based on sequential deformation |
| CN114639087A (en) * | 2022-03-14 | 2022-06-17 | 武汉理工大学 | Traffic sign detection method and device |
-
2023
- 2023-03-09 CN CN202310244263.6A patent/CN116188872A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210150283A1 (en) * | 2019-11-15 | 2021-05-20 | Salesforce.Com, Inc. | Noise-Resistant Object Detection with Noisy Annotations |
| US20220058420A1 (en) * | 2020-08-22 | 2022-02-24 | Tsinghua University | Scene text detection method and system based on sequential deformation |
| CN112215293A (en) * | 2020-10-20 | 2021-01-12 | 平安国际智慧城市科技股份有限公司 | Plant disease and insect pest identification method and device and computer equipment |
| CN112989969A (en) * | 2021-02-26 | 2021-06-18 | 清远市智慧农业农村研究院 | Crop pest and disease identification method and device |
| CN113469177A (en) * | 2021-06-30 | 2021-10-01 | 河海大学 | Drainage pipeline defect detection method and system based on deep learning |
| CN114639087A (en) * | 2022-03-14 | 2022-06-17 | 武汉理工大学 | Traffic sign detection method and device |
Non-Patent Citations (2)
| Title |
|---|
| 田宏: "基于深度学习的航拍图像小目标检测技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 2, pages 138 - 994 * |
| 醉居楼: "mmdetection(搬砖)", pages 1 - 22, Retrieved from the Internet <URL:《https://zhuanlan.zhihu.com/p/478580526》> * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117253122A (en) * | 2023-11-17 | 2023-12-19 | 云南大学 | Corn seed approximate variety screening method, device, equipment and storage medium |
| CN117253122B (en) * | 2023-11-17 | 2024-01-23 | 云南大学 | Corn seed approximate variety screening method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Weinstein | Scene‐specific convolutional neural networks for video‐based biodiversity detection | |
| US11636701B2 (en) | Method for calculating deviation relations of a population | |
| Kumar | Plant disease detection using CNN | |
| Bhagat et al. | WheatNet-lite: A novel light weight network for wheat head detection | |
| Latif et al. | Deep learning based intelligence cognitive vision drone for automatic plant diseases identification and spraying | |
| Kumar K et al. | Detection of rice plant disease using AdaBoostSVM classifier | |
| Zhang et al. | Deep learning based automatic grape downy mildew detection | |
| CN116188872A (en) | Automatic forestry plant diseases and insect pests identification method and device | |
| Kirola et al. | Plants diseases prediction framework: a image-based system using deep learning | |
| Kumar et al. | An identification of crop disease using image segmentation | |
| Rai et al. | Classification of diseased cotton leaves and plants using improved deep convolutional neural network | |
| Miranda et al. | Pest identification using image processing techniques in detecting image pattern through neural network | |
| Araujo et al. | Disease identification in chilli leaves using machine learning techniques | |
| CN112231535B (en) | Method for making multi-modal data set in field of agricultural diseases and insect pests, processing device and storage medium | |
| Montellano | Butterfly, Larvae and Pupae Defects Detection Using Convolutional Neural Network and Apriori Algorithm | |
| Yi et al. | Lesion region segmentation via weakly supervised learning | |
| Nan et al. | A novel method for maize leaf disease classification using the RGB-D post-segmentation image data | |
| Dahiya et al. | An Effective Detection of Litchi Disease using Deep Learning | |
| Safari et al. | A Review on Automated Detection and Assessment of Fruit Damage Using Machine Learning | |
| Sri et al. | Disease Detection using Region-Based Convolutional Neural Network and ResNet | |
| Mayuri et al. | Artificial Neural Network (ANN) With Chan-Vese (CV) Algorithm-Based Plant Disease Detection And Classification | |
| Afonso et al. | Automatic trait estimation in floriculture using computer vision and deep learning | |
| Kalaiarasi et al. | Transforming Agriculture with AI: Automated Crop Insect Identification Using InceptionV3 Transfer Learning | |
| CN118115815A (en) | Crop pest identification method based on deep learning | |
| Virdi | Aristotle Philosophy on Governance (A Study) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |