Silage corn pest and disease identification method and system based on image analysis

Abstract

The present invention belongs to the field of image analysis technology, and specifically relates to a method and system for identifying pests and diseases of silage corn based on image analysis, which obtains growth characteristic data and pest and disease characteristics of silage corn, and establishes a general database and a sub-database. The silage corn is imaged and the current growth stage is determined; the sub-database is called for primary matching, and if the match is successful, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, the general database is called for intermediate matching, and if it is matched, the corresponding matching result is sent to the management terminal. If it is not matched, the ultimate matching is performed based on big data, and if it is matched, the corresponding matching result is sent to the management terminal, and if it is still not matched, the corresponding pest and disease image data and the extracted pest and disease characteristics are sent to the management terminal, and the unmatched result is notified. Pest and disease management of silage corn throughout its life cycle is achieved, effectively improving economic benefits.

Description

Technical Field

[0001] The present invention belongs to the technical field of image analysis, and in particular relates to a method and system for identifying pests and diseases of silage corn based on image analysis. Background Art

[0002] Silage corn is corn that is harvested during the appropriate harvest period. All green plants on the ground, including the cobs, are harvested, chopped, processed, and fermented to make silage feed. It is used to feed cattle, sheep and other livestock in winter. It is very common in northern pastoral areas. Silage corn is one of the three types of corn (grain corn, silage corn, and fresh corn) divided by harvest and use.

[0003] The growth stages of silage corn are mainly divided into the seedling stage, jointing stage, trumpet stage, tasseling and silking stage, filling stage and maturity stage. The seedling stage is a period from emergence to jointing, mainly centered on root growth, with fast root growth and slow growth of the above-ground part. The management goal of the seedling stage is to ensure full seedlings, seedling yield and strong seedlings. The standards for strong seedlings include developed root system, deep roots and green leaves, wide and short leaves, flat seedling base, short and thick internodes, uniform colony and vigorous growth. Common diseases and pests in the seedling stage mainly include corn root rot, seedling blight, white grubs and thrips.

[0004] The jointing stage is an important turning point in the growth of silage corn, when the plant begins to grow rapidly in height and thickness. The main pests and diseases faced during the jointing stage include corn borers, locusts, corn aphids, top rot, brown spot disease, and bacterial stem rot.

[0005] The trumpet-mouth stage is the period when the growth rate of corn plants reaches its peak and the vegetative growth is vigorous. The pests and diseases faced during the trumpet-mouth stage mainly include fall armyworm, armyworm, thrips, corn borer, two-spotted firefly leaf beetle, corn leaf spot disease and rust.

[0006] The tasseling and silking period is the reproductive growth stage, when corn begins to tassel and silk. This is a critical period in the process of corn grain formation. The tasseling and silking period includes two stages. The first stage is when the silk of the female ear begins to emerge from the bracts, which is the silking period of corn. The second stage means that the volume of the grain in the middle of the ear is basically completed, the endosperm is in a clear pulp, and the moisture content of the grain is about 85%. The pests and diseases faced during this period include corn borer, southern rust, large leaf spot, and small leaf spot.

[0007] The grain filling period is the main period for grain filling, during which the volume and weight of corn kernels increase rapidly. The pests and diseases during this period are southern rust, large spot disease, small spot disease, sheath blight, and corn potassium deficiency.

[0008] The maturity stage is when the stems and leaves of silage corn plants stop growing and enter the reproductive growth stage centered on flowering, pollination, and fruiting and maturation. Proper management at this time can prevent premature aging of stems and leaves, promote flowering, pollination, and grain filling, and increase the number and weight of fruiting grains. The main pests and diseases during this period are southern rust, large spot disease, and small spot disease.

[0009] The above growth stages are the key periods for corn from germination to maturity, and each stage has its own specific growth characteristics and management points. Since the growth characteristics of silage corn are different at each growth stage, the types of diseases and pests at each growth stage will be different. Therefore, each growth stage of silage corn needs to be managed according to its growth characteristics and corresponding pest characteristics, so as to achieve good growth management of silage corn to the greatest extent, maximize the yield of silage corn, and achieve the highest economic benefits.

[0010] Therefore, how to efficiently identify pests and diseases at each growth stage of silage corn based on the characteristics of each growth stage and the characteristics of pests and diseases of silage corn, and then realize specific pest and disease management for the entire life cycle of silage corn based on the pest and disease identification results, is a technical problem that needs to be solved urgently. Summary of the invention

[0011] The purpose of the present invention is to maximize the final yield of silage corn and improve the economic benefits by identifying pests and diseases at each growth stage of silage corn and then managing pests and diseases specifically throughout the entire life cycle.

[0012] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0013] In a first aspect, a method for identifying pests and diseases of silage corn based on image analysis is provided, comprising the following steps:

[0014] S1: Based on big data, the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage are obtained to establish a general database; according to the relationship between each growth stage of silage corn and the corresponding pest and disease characteristic data, a sub-database of each growth stage of silage corn is established;

[0015] S2: collecting images of the silage corn plants through an image collection module, and analyzing the collected images to determine the current growth stage of the silage corn plants;

[0016] S3: extracting corresponding pest and disease features according to the collected image data of the silage corn plant, calling the sub-database of the corresponding growth stage, and performing primary matching between the corresponding pest and disease features and the pest and disease feature data of the corresponding growth stage. If the corresponding pest and disease type is matched, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, executing step S4;

[0017] S4: calling the general database, performing intermediate matching of the extracted pest and disease characteristics with the pest and disease characteristic data corresponding to other growth stages in the general database, and if the corresponding pest and disease type is matched, sending the corresponding matching result to the management terminal, and if the corresponding pest and disease type is not matched, executing step S5;

[0018] S5: Based on big data, the extracted pest and disease features are ultimately matched with other pest and disease feature data. If the corresponding pest and disease type is matched, the pest and disease location information and pest and disease distribution characteristics of the silage corn image data with the pest and disease features are further extracted, and step S6 is executed; if the corresponding pest and disease type is still not matched, the corresponding pest and disease image data and the extracted pest and disease features are sent to the management terminal, and the unmatched result is notified;

[0019] S6: analyzing the severity, development trend and control measures of pests and diseases according to the further extracted distribution characteristics of pests and diseases, and sending the location information, severity, development trend and control measures of pests and diseases to the management terminal;

[0020] S7: The silage corn planting manager carries out corresponding control work according to the received information on the location of pests and diseases, the severity of pests and diseases, the development trend of pests and diseases and control measures.

[0021] Preferably, the data in the total database in step S1 include plant growth characteristic data, growth image data of all growth stages of silage corn, and disease and pest characteristic data and disease and pest image data of all silage corn; the number of sub-databases is the same as the number of growth stages of silage corn, and the data in each sub-database include plant growth characteristic data, growth image data of each growth stage, and disease and pest characteristic data and disease and pest image data of each growth stage.

[0022] Preferably, the image acquisition module in step S2 acquires images of silage corn by taking aerial photos of silage corn plants using a drone; and the normalized cross-correlation matching algorithm is used to analyze the acquired images and determine the current growth stage of the silage corn.

[0023] Preferably, step S2 includes the following specific process:

[0024] S21: The image acquisition module acquires images of the silage corn to obtain image data of the current silage corn plant;

[0025] S22: performing data preprocessing on the acquired image data of the current silage corn plant;

[0026] S23: extracting plant growth features from the image data after data preprocessing, calling the overall database, matching the extracted plant growth features with the growth features of each stage in the overall database, and determining the current growth stage of the silage corn.

[0027] Preferably, the specific process of data preprocessing in step S22 is as follows:

[0028] S221: performing data cleaning on the acquired image data of the current silage corn plant, and deleting invalid data and duplicate data in the image data;

[0029] S222: performing image correction and multi-spectral image registration and adjustment on the image data after data cleaning;

[0030] S223: performing image enhancement processing on the image correction, multi-spectral image registration and adjusted image data;

[0031] S224: performing image fusion processing on the image after image enhancement processing.

[0032] Preferably, in step S3, the process of extracting corresponding pest and disease features from the collected image data of silage corn and matching the images with those in the sub-database is implemented by a robot image feature matching algorithm.

[0033] Preferably, the specific process of image feature extraction and image matching by the robot image feature matching algorithm in step S3 is as follows:

[0034] S31: constructing the scale space of silage corn images;

[0035] S32: Find extreme points in the scale space of the constructed silage corn image;

[0036] S33: testing the extreme points found in the scale space of the silage corn image to eliminate mismatching points and obtain accurate feature points;

[0037] S34: Calculate a direction for each of the precise feature points, perform further calculations based on the direction, and use the gradient direction distribution characteristics of the pixels in the neighborhood of the key point to specify a direction parameter for each key point, so that the operator has rotation invariance;

[0038] S35: Generate 128-dimensional descriptors of the matched silage corn image and the silage corn image in the sub-database, match the descriptors of each part of the two images, and when the 128-dimensional descriptors are matched, the two feature points are successfully matched.

[0039] In the second aspect, a silage corn pest and disease identification system based on image analysis includes a data acquisition module, a database establishment module, an image acquisition module, an image preprocessing module, a feature extraction module, an image matching module, a data call module, a data transmission module, and a management terminal, wherein the data acquisition module is connected to the database establishment module, the image acquisition module is connected to the image preprocessing module, the image preprocessing module is connected to the feature extraction module, the feature extraction module is connected to the image matching module, the image matching module is connected to the data call module, the image matching module is connected to the data transmission module, and the data transmission module is connected to the management terminal;

[0040] The data acquisition module is used to acquire growth characteristic data of silage corn at each growth stage and corresponding pest and disease characteristic data at each growth stage based on big data;

[0041] The database establishment module is used to establish a general database based on the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage; and to establish a sub-database of each growth stage of silage corn based on the relationship between each growth stage of silage corn and the corresponding pest and disease characteristic data;

[0042] The image acquisition module is used to acquire images of the current silage corn plants;

[0043] The image preprocessing module is used to perform data preprocessing on the acquired image data of the current silage corn plant, including data cleaning, image correction, registration and adjustment of multi-spectral images, image enhancement processing and image fusion processing;

[0044] The feature extraction module is used to extract features from the pre-processed image of the silage corn plant;

[0045] The image matching module is used to perform image matching based on the extracted image features;

[0046] The data calling module is used to call the pest and disease characteristic data in the main database or sub-database or big data;

[0047] The data transmission module is used to send the matching result to the management terminal;

[0048] The management terminal is used for the management personnel to obtain the corresponding matching results, so that the management personnel can perform corresponding pest management based on the matching results.

[0049] Preferably, the main database and sub-database also include treatment suggestions corresponding to each type of pest and disease. The data calling module will call the corresponding treatment suggestions while calling the pest and disease feature data in the main database, sub-database or big data. When the data transmission module sends the matching results to the management terminal, it will also send the matched pest and disease data and the corresponding treatment suggestions to the management terminal.

[0050] The beneficial effects of the present invention include:

[0051] The silage corn pest and disease identification method and system based on image analysis provided by the present invention establish a general database and a sub-database by acquiring the growth characteristic data and pest and disease characteristics of silage corn. The silage corn is imaged and the current growth stage is determined; the sub-database is called for primary matching, and if the match is successful, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, the general database is called for intermediate matching, and if it is matched, the corresponding matching result is sent to the management terminal. If it is not matched, the ultimate matching is performed based on big data, and if it is matched, the corresponding matching result is sent to the management terminal, and if it is still not matched, the corresponding pest and disease image data and the extracted pest and disease characteristics are sent to the management terminal, and the unmatched result is notified.

[0052] On the one hand, the image data of the silage corn plants collected in real time can be used to obtain the corresponding silage corn growth stage. The corresponding growth stage can be judged based on the real-time silage corn image data, and whether the growth and development of the silage corn is normal can be judged based on the growth time, so as to know whether the irrigation or nutrition of the silage corn is sufficient.

[0053] On the other hand, based on the real-time collected silage corn image data, the sub-database is called for initial matching, and efficient pest and disease identification is performed on the real-time image data based on the characteristics of each growth stage and the characteristics of pests and diseases of the silage corn. Based on the pest and disease identification results, it is judged whether there are pests and diseases in the current growth stage. If pests and diseases exist, the corresponding pest and disease treatment is carried out in time, thereby realizing specific pest and disease management for the entire life cycle of silage corn.

[0054] On the other hand, after matching the corresponding pest and disease information, the pest and disease location information and distribution characteristics of the corresponding silage corn image data are further extracted to analyze the severity of the pests and diseases, the development trend of the pests and diseases, and the control measures. The data is sent to the management terminal, and the administrator uses the data to carry out subsequent pest and disease control work, thus achieving accurate and efficient pest and disease control. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The figure is a schematic flow chart of the silage corn pest and disease identification method based on image analysis of the present invention.

[0056] Figure 2 The schematic diagram of the architecture of the silage corn pest and disease identification system based on image analysis of the present invention.

[0057] Figure 3 It is a schematic diagram of the process of image feature extraction and image matching of the present invention. DETAILED DESCRIPTION

[0058] The following is combined with Figure 1-Figure 3 The present invention is further described in detail:

[0059] Example 1

[0060] Silage corn pest and disease identification method based on image analysis, see Figure 1 , including the following steps:

[0061] S1: Based on big data, the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage are obtained to establish a general database; according to the relationship between the growth stage of silage corn and the corresponding pest and disease characteristic data, a sub-database of each growth stage of silage corn is established as the data basis for the subsequent growth stage analysis and pest and disease analysis of the current silage corn;

[0062] S2: The image acquisition module is used to acquire images of the silage corn plants, and the acquired images are analyzed to determine the current growth stage of the silage corn plants. The current growth stage is compared with the growth stage that should be in the actual growth time to determine whether the current growth and development of the silage corn is normal. At the same time, it serves as the basis for subsequent pest and disease identification, so as to facilitate matching the corresponding pest and disease categories according to the corresponding growth stage. The reason for not identifying pests and diseases in all stages in the database is to improve the efficiency of data comparison, that is, the efficiency of pest and disease identification;

[0063] S3: extracting corresponding pest and disease features according to the collected image data of the silage corn plant, calling the sub-database of the corresponding growth stage, and performing primary matching between the corresponding pest and disease features and the pest and disease feature data of the corresponding growth stage. If the corresponding pest and disease type is matched, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, executing step S4;

[0064] S4: calling the general database, performing intermediate matching of the extracted pest and disease characteristics with the pest and disease characteristic data corresponding to other growth stages in the general database, and if the corresponding pest and disease type is matched, sending the corresponding matching result to the management terminal, and if the corresponding pest and disease type is not matched, executing step S5;

[0065] S5: Based on big data, the extracted pest and disease features are ultimately matched with other pest and disease feature data. If the corresponding pest and disease type is matched, the pest and disease location information and pest and disease distribution characteristics of the silage corn image data with the pest and disease features are further extracted, and step S6 is executed; if the corresponding pest and disease type is still not matched, the corresponding pest and disease image data and the extracted pest and disease features are sent to the management terminal, and the unmatched result is notified;

[0066] S6: analyzing the severity, development trend and control measures of pests and diseases according to the further extracted distribution characteristics of pests and diseases, and sending the location information, severity, development trend and control measures of pests and diseases to the management terminal;

[0067] S7: The silage corn planting manager carries out corresponding control work according to the received information on the location of pests and diseases, the severity of pests and diseases, the development trend of pests and diseases and control measures.

[0068] At the same time as step S3, the current growth stage of the silage corn plant determined by analysis is combined with the planting time to determine whether the current growth and development of the silage corn plant is normal, and the determination result of whether the growth and development is normal is sent to the management terminal. The management personnel obtain the determination result of whether the current growth and development of the silage corn is normal through the management terminal. If it is found that the growth and development of the silage corn is abnormal, the irrigation amount of the silage corn and the supplementary amount of nutrients are adjusted according to the current actual growth and development situation.

[0069] Since there are several important growth stages for silage corn, mainly including seedling stage, jointing stage, large trumpet stage, tasseling and silking stage, filling stage until maturity. There are corresponding growth characteristics in these growth stages, and there may be corresponding pests and diseases in each growth stage, that is to say, the pest and disease type of silage corn is closely related to its specific growth stage, and the pest and disease type existing in different growth stages is different. Therefore, the present application expects to identify pests and diseases for silage corn in a targeted manner at different stages of silage corn, on the one hand, it can improve the pest and disease identification efficiency, on the other hand, it is also expected to judge the growth stage of silage corn before pest and disease identification, and then judge whether the growth and development of silage corn in the current growth stage is normal according to the current growth stage and planting cycle.

[0070] The data in the total database in step S1 include plant growth characteristic data, growth image data, and pest and disease characteristic data and pest and disease image data of all silage corn at all growth stages; the number of sub-databases is the same as the number of growth stages of silage corn, and the data in each sub-database include plant growth characteristic data, growth image data, and pest and disease characteristic data and pest and disease image data of each growth stage. Step S1 is the data basis of the entire pest and disease identification method. The established total database and sub-database contain plant growth characteristic data, growth image data, and pest and disease characteristic data and pest and disease image data of all silage corn at all growth stages, which serve as the data comparison basis for subsequent identification of the current growth stage of silage corn and pest and disease identification.

[0071] After matching the corresponding pest information from the silage corn plant image data, the pest location information and pest distribution characteristics of the silage corn image data with pest characteristics are further extracted, and the pest severity, pest development trend and control measures are analyzed according to the pest distribution characteristics, and the pest location information, pest severity, pest development trend and control measures are sent to the management terminal; the silage corn planting manager performs corresponding control work according to the received pest location information, pest severity, pest development trend and control measures. When pest information exists, it is necessary to analyze the current pest situation of silage corn in order to carry out subsequent pest control work.

[0072] If the corresponding pest type is still not matched, the corresponding pest image data, location information and extracted pest features will be sent to the management terminal, and the unmatched result will be notified. Through manual on-site inspection, the control plan will be further determined.

[0073] The image acquisition module in step S2 acquires images of silage corn by taking aerial photos of silage corn plants using a drone. The normalized cross-correlation matching algorithm is used to analyze and determine the current growth stage of the silage corn.

[0074] Example 2

[0075] Based on Example 1, step S2 includes the following specific process:

[0076] S21: The image acquisition module acquires images of the silage corn to obtain image data of the current silage corn plant;

[0077] S22: performing data preprocessing on the acquired image data of the current silage corn plant to avoid abnormal data in the acquired image data of the current silage corn plant, which may affect the accuracy of the recognition of the subsequent growth stage and the recognition result of pests and diseases;

[0078] S23: extracting plant growth features from the image data after data preprocessing, calling the overall database, matching the extracted plant growth features with the growth features of each stage in the overall database, and determining the current growth stage of the silage corn.

[0079] The specific process of data preprocessing in step S22 is as follows:

[0080] S221: Cleaning the acquired image data of the current silage corn plant to delete invalid data and duplicate data in the image data.

[0081] S222: Image correction and multi-spectral image registration and adjustment are performed on the image data after data cleaning. The purpose of image correction is to eliminate the influence of scattering caused by the sun angle and atmosphere, nonlinear and inconsistent distortion of remote sensors, nonlinearity of film development, etc., and various calibration and calibration processes are performed. The registration and adjustment of multi-spectral images is to ensure that images acquired in different bands or at different times can be accurately aligned for subsequent image analysis.

[0082] S223: Image enhancement processing is performed on the image correction, multi-spectral image registration and adjusted image data. Image enhancement processing includes image inversion, contrast stretching, dynamic range compression, grayscale slicing, etc., which are used to improve the visual effect of the image and highlight the features of interest.

[0083] S224: performing image fusion processing on the image after the image enhancement processing, fusing multiple images into one image, or mosaicking the images to cover a larger area, and cropping the images to remove unnecessary parts as needed.

[0084] In another embodiment, the image processing process also includes cloud and shadow removal processing to remove clouds and shadows in the image so as to more accurately analyze the surface features, and atmospheric correction processing to remove the influence of the atmosphere on the image, such as atmospheric scattering and absorption, so as to improve the authenticity and accuracy of the image.

[0085] Example 3

[0086] Based on Example 1 or Example 2, the process of extracting the corresponding pest and disease features from the collected image data of silage corn and matching the images with those in the sub-database in step S3 is implemented by a robot image feature matching algorithm. Figure 3 ,The specific process of image feature extraction and image matching is as follows:

[0087] S31: constructing the scale space of silage corn images;

[0088] S32: Find extreme points in the scale space of the constructed silage corn image;

[0089] S33: testing the extreme points found in the scale space of the silage corn image to eliminate mismatching points and obtain accurate feature points;

[0090] S34: Calculate a direction for each of the precise feature points, perform further calculations based on the direction, and use the gradient direction distribution characteristics of the pixels in the neighborhood of the key point to specify a direction parameter for each key point, so that the operator has rotation invariance;

[0091] S35: Generate 128-dimensional descriptors of the matched silage corn image and the silage corn image in the sub-database, match the descriptors of each part of the two images, and when the 128-dimensional descriptors are matched, the two feature points are successfully matched.

[0092] In the above scheme, the process of constructing the scale space of the silage corn image in step S31 is an initialization operation process to simulate the multi-scale characteristics of the image data and generate the scale space of the silage corn image by convolving the Gaussian difference kernels of different scales with the image.

[0093] Step S32 searches for extreme points in the scale space of the constructed silage corn image. Each sampling point is compared with all its neighboring points to see whether it is larger or smaller than its neighboring points in the image domain and scale domain. The actual process is: compare the middle detection point with its 8 neighboring points of the same scale and 9×2 points corresponding to the upper and lower adjacent scales, a total of 26 points, to ensure that extreme points are detected in both the scale space and the two-dimensional image space. If a point is the maximum or minimum value in the 26 fields of the current layer and the upper and lower layers of the DOG scale space, it is considered to be a feature point of the image at that scale.

[0094] In step S33, the extreme points found in the scale space of the silage corn image are tested to eliminate mismatched points and obtain accurate feature points. Since the DOG value is sensitive to noise and edges, the local extreme points detected in the above DOG scale space need to be further tested before they can be accurately located as feature points. In order to improve the stability of key points, it is necessary to perform curve fitting on the scale space DOG function, and among the detected feature points, low-contrast feature points and unstable edge response points should be removed.

[0095] In step S34, a direction is calculated for each of the precise feature points, and further calculations are performed based on the direction. The gradient direction distribution characteristics of the key point neighborhood pixels are used to specify direction parameters for each key point, so that the operator has rotation invariance. In the actual operation process, sampling is performed in the neighborhood window centered on the key point, and the gradient direction of the neighborhood pixels is counted using a histogram. The range of the gradient histogram is 0 to 360 degrees, with a column every 45 degrees, a total of 8 columns, or a column every 10 degrees, a total of 36 columns. The histogram should be smoothed using a Gaussian function to reduce the impact of mutations. The peak of the histogram represents the main direction of the neighborhood gradient at the key point, that is, the direction of the key point.

[0096] In step S35, a 128-dimensional descriptor of the matched silage corn image and the silage corn image in the sub-database is generated, and the descriptors of each part of the two images are matched. When the matching reaches 128 dimensions, the two feature points are successfully matched. In the actual calculation process, in order to enhance the robustness of the matching, each key point is described using 4×4, a total of 16 seed points, so that 128 data can be generated for one key point, that is, a 128-dimensional robot image feature matching feature vector is finally formed. At this time, the robot image feature matching feature vector has removed the influence of geometric deformation factors such as scale change and rotation. If the length of the feature vector is normalized, the influence of illumination change can be further removed. After the robot image feature matching feature vectors of the two images are generated, the next step is to use the Euclidean distance of the key point feature vector as the similarity determination metric of the key points in the two images. Take a key point in the matched silage corn image and find the first two key points with the closest Euclidean distance to the matched silage corn image. If the closest distance divided by the second closest distance is less than a certain ratio threshold, then accept this pair of matching points. Lowering this ratio threshold will reduce the number of matching points for robot image feature matching, but it will be more stable.

[0097] Silage corn pest and disease identification system based on image analysis, see Figure 2 , including a data acquisition module, a database establishment module, an image acquisition module, an image preprocessing module, a feature extraction module, an image matching module, a data calling module, a data transmission module, and a management terminal, wherein the data acquisition module is connected to the database establishment module, the image acquisition module is connected to the image preprocessing module, the image preprocessing module is connected to the feature extraction module, the feature extraction module is connected to the image matching module, the image matching module is connected to the data calling module, the image matching module is connected to the data transmission module, and the data transmission module is connected to the management terminal.

[0098] The data acquisition module is used to acquire the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage based on big data. The database establishment module is used to establish a general database based on the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage. According to the relationship between each growth stage of silage corn and the corresponding pest and disease characteristic data, a sub-database of each growth stage of silage corn is established. The image acquisition module is used to acquire images of the current silage corn plant.

[0099] The image preprocessing module is used to perform data preprocessing on the acquired image data of the current silage corn plant, including data cleaning, image correction, registration and adjustment of multi-spectral images, image enhancement processing and image fusion processing. The feature extraction module is used to extract features from the preprocessed image of the silage corn plant. The image matching module is used to perform image matching based on the extracted image features. The data calling module is used to call the pest and disease feature data in the main database or sub-database or big data. The data transmission module is used to send the matching results to the management terminal. The management terminal is used for the management personnel to obtain the corresponding matching results, so that the management personnel can perform corresponding pest and disease management based on the matching results.

[0100] The general database and sub-database also include treatment suggestions corresponding to each type of pest and disease. The data calling module will call the corresponding treatment suggestions while calling the pest and disease feature data in the general database, sub-database or big data. When the data transmission module sends the matching results to the management terminal, it will also send the matched pest and disease data and the corresponding treatment suggestions to the management terminal.

[0101] In summary, the silage corn pest and disease identification method and system based on image analysis provided by the present invention establishes a general database and a sub-database by acquiring the growth characteristic data and pest and disease characteristics of silage corn. The silage corn is imaged and the current growth stage is determined; the sub-database is called for primary matching, and if the match is successful, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, the general database is called for intermediate matching, and if it is matched, the corresponding matching result is sent to the management terminal. If it is not matched, the ultimate matching is performed based on big data, and if it is matched, the corresponding matching result is sent to the management terminal, and if it is still not matched, the corresponding pest and disease image data and the extracted pest and disease characteristics are sent to the management terminal, and the unmatched result is notified.

[0102] The corresponding silage corn growth stage is obtained through the real-time collected image data of silage corn plants. And according to the growth time, it is judged whether the growth and development of silage corn is normal, so as to know whether the irrigation or nutrition of silage corn is sufficient. On the other hand, according to the real-time collected silage corn image data, the sub-database is called for initial matching, and efficient pest and disease identification is performed. According to the pest and disease identification results, it is judged whether there are pests and diseases in the current growth stage. By extracting the location information and distribution characteristics of the pests and diseases of the corresponding silage corn image data, the severity of the pests and diseases, the development trend of the pests and diseases and the control measures are analyzed, and sent to the management terminal. The administrator uses the data obtained to carry out subsequent pest and disease control work, realizing accurate and efficient pest and disease control.

Claims

1. A method for identifying pests and diseases of silage corn based on image analysis, characterized in that: The following steps are involved: S1: Based on big data, the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage are obtained to establish a general database; according to the relationship between each growth stage of silage corn and the corresponding pest and disease characteristic data, a sub-database of each growth stage of silage corn is established; S2: collecting images of the silage corn plants through an image collection module, and analyzing the collected images to determine the current growth stage of the silage corn plants; S3: extracting corresponding pest and disease features according to the collected image data of the silage corn plant, calling the sub-database of the corresponding growth stage, and performing primary matching between the corresponding pest and disease features and the pest and disease feature data of the corresponding growth stage. If the corresponding pest and disease type is matched, the corresponding matching result is sent to the management terminal. If the corresponding pest and disease type is not matched, executing step S4; S4: calling the general database, performing intermediate matching of the extracted pest and disease characteristics with the pest and disease characteristic data corresponding to other growth stages in the general database, and if the corresponding pest and disease type is matched, sending the corresponding matching result to the management terminal, and if the corresponding pest and disease type is not matched, executing step S5; S5: Based on big data, the extracted pest and disease features are ultimately matched with other pest and disease feature data. If the corresponding pest and disease type is matched, the pest and disease location information and pest and disease distribution characteristics of the silage corn image data with the pest and disease features are further extracted, and step S6 is executed; if the corresponding pest and disease type is still not matched, the corresponding pest and disease image data and the extracted pest and disease features are sent to the management terminal, and the unmatched result is notified; S6: analyzing the severity, development trend and control measures of pests and diseases according to the further extracted distribution characteristics of pests and diseases, and sending the location information, severity, development trend and control measures of pests and diseases to the management terminal; S7: The silage corn planting manager carries out corresponding control work according to the received information on the location of pests and diseases, the severity of pests and diseases, the development trend of pests and diseases and control measures.

2. The method for identifying pests and diseases of silage corn based on image analysis according to claim 1, characterized in that: The data in the total database in step S1 include plant growth characteristic data, growth image data of all growth stages of silage corn, and disease and insect pest characteristic data and disease and insect pest image data of all silage corn; the number of sub-databases is the same as the number of growth stages of silage corn, and the data in each sub-database include plant growth characteristic data, growth image data of each growth stage, and disease and insect pest characteristic data and disease and insect pest image data of each growth stage.

3. The method for identifying pests and diseases of silage corn based on image analysis according to claim 1, characterized in that: The image acquisition module in step S2 acquires images of silage corn by taking aerial photos of silage corn plants using a drone; and the normalized cross-correlation matching algorithm is used to analyze the acquired images and determine the current growth stage of the silage corn.

4. The method for identifying pests and diseases of silage corn based on image analysis according to claim 1, characterized in that: Step S2 includes the following specific processes: S21: The image acquisition module acquires images of the silage corn to obtain image data of the current silage corn plant; S22: performing data preprocessing on the acquired image data of the current silage corn plant; S23: extracting plant growth features from the image data after data preprocessing, calling the overall database, matching the extracted plant growth features with the growth features of each stage in the overall database, and determining the current growth stage of the silage corn.

5. The method for identifying pests and diseases of silage corn based on image analysis according to claim 4, characterized in that: The specific process of data preprocessing in step S22 is as follows: S221: performing data cleaning on the acquired image data of the current silage corn plant, and deleting invalid data and duplicate data in the image data; S222: performing image correction and multi-spectral image registration and adjustment on the image data after data cleaning; S223: performing image enhancement processing on the image correction, multi-spectral image registration and adjusted image data; S224: performing image fusion processing on the image after image enhancement processing.

6. The method for identifying pests and diseases of silage corn based on image analysis according to claim 1, characterized in that: In step S3, the process of extracting the corresponding pest and disease features from the collected image data of the silage corn and matching the images with those in the sub-database is implemented by a robot image feature matching algorithm.

7. The method for identifying pests and diseases of silage corn based on image analysis according to claim 6, characterized in that: The specific process of image feature extraction and image matching in step S3 is as follows: S31: constructing the scale space of silage corn images; S32: Find extreme points in the scale space of the constructed silage corn image; S33: testing the extreme points found in the scale space of the silage corn image to eliminate mismatching points and obtain accurate feature points; S34: Calculate a direction for each of the precise feature points, perform further calculations based on the direction, and use the gradient direction distribution characteristics of the pixels in the neighborhood of the key point to specify a direction parameter for each key point, so that the operator has rotation invariance; S35: Generate 128-dimensional descriptors of the matched silage corn image and the silage corn image in the sub-database, match the descriptors of each part of the two images, and when the 128-dimensional descriptors are matched, the two feature points are successfully matched.

8. The method for identifying pests and diseases of silage corn based on image analysis according to claim 1, characterized in that: Step S6 includes the following specific processes: S61: analyzing the severity of pests and diseases according to the size of the distribution range and the density of pests and diseases in the extracted pest and disease distribution characteristics, wherein the severity of pests and diseases is divided into three levels, including severe, moderate and mild; S62: Analyze the development trend of pests and diseases according to the distribution of edges in the extracted pest and disease distribution characteristics; S63: Formulate corresponding control measures based on the analyzed severity and development trend of the pests and diseases, and send the pest and disease location information, pest and disease severity, pest and disease development trend and control measures to the management terminal.

9. A silage corn pest identification system based on image analysis, used to implement the silage corn pest identification method based on image analysis according to any one of claims 1 to 8, characterized in that: It includes a data acquisition module, a database establishment module, an image acquisition module, an image preprocessing module, a feature extraction module, an image matching module, a data call module, a data transmission module, and a management terminal. The data acquisition module is connected to the database establishment module, the image acquisition module is connected to the image preprocessing module, the image preprocessing module is connected to the feature extraction module, the feature extraction module is connected to the image matching module, the image matching module is connected to the data call module, the image matching module is connected to the data transmission module, and the data transmission module is connected to the management terminal; The data acquisition module is used to acquire growth characteristic data of silage corn at each growth stage and corresponding pest and disease characteristic data at each growth stage based on big data; The database establishment module is used to establish a general database based on the growth characteristic data of silage corn at each growth stage and the corresponding pest and disease characteristic data at each growth stage; and to establish a sub-database of each growth stage of silage corn based on the relationship between each growth stage of silage corn and the corresponding pest and disease characteristic data; The image acquisition module is used to acquire images of the current silage corn plants; The image preprocessing module is used to perform data preprocessing on the acquired image data of the current silage corn plant, including data cleaning, image correction, registration and adjustment of multi-spectral images, image enhancement processing and image fusion processing; The feature extraction module is used to extract features from the pre-processed image of the silage corn plant; The image matching module is used to perform image matching based on the extracted image features; The data calling module is used to call the pest and disease characteristic data in the main database or sub-database or big data; The data transmission module is used to send the matching result to the management terminal; The management terminal is used for the management personnel to obtain the corresponding matching results, so that the management personnel can perform corresponding pest management based on the matching results.

10. The silage corn pest identification system based on image analysis according to claim 9, characterized in that: The general database and sub-database also include treatment suggestions corresponding to each type of pest and disease. The data calling module will call the corresponding treatment suggestions while calling the pest and disease feature data in the general database, sub-database or big data. When the data transmission module sends the matching results to the management terminal, it will also send the matched pest and disease data and the corresponding treatment suggestions to the management terminal.

Images