CN114511777A - Peach tree flowering phase monitoring and evaluating method based on image recognition method - Google Patents

Peach tree flowering phase monitoring and evaluating method based on image recognition method Download PDF

Info

Publication number
CN114511777A
CN114511777A CN202210017857.9A CN202210017857A CN114511777A CN 114511777 A CN114511777 A CN 114511777A CN 202210017857 A CN202210017857 A CN 202210017857A CN 114511777 A CN114511777 A CN 114511777A
Authority
CN
China
Prior art keywords
peach
flowering phase
flowering
monitoring
peach tree
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
Application number
CN202210017857.9A
Other languages
Chinese (zh)
Inventor
朱怡航
张小斌
顾清
赵懿滢
郑可锋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Academy of Agricultural Sciences
Original Assignee
Zhejiang Academy of Agricultural Sciences
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Academy of Agricultural Sciences filed Critical Zhejiang Academy of Agricultural Sciences
Priority to CN202210017857.9A priority Critical patent/CN114511777A/en
Publication of CN114511777A publication Critical patent/CN114511777A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Image Analysis (AREA)

Abstract

The invention provides a peach flowering phase monitoring and evaluating method based on an image recognition method, which comprises the following steps of 1) arranging a crop growth recorder in a peach orchard, and shooting peach flowering phase images at the same visual angle at fixed time points by using an agricultural intelligent camera of the crop growth recorder; 2) arranging an agricultural meteorological monitor in the peach orchard, and recording the date that the daily average temperature is first greater than the biological zero degree of the peach trees, wherein the agricultural meteorological monitor is used for monitoring meteorological data such as air temperature, illumination, rainfall and the like of the peach orchard; 3) and combining meteorological data, and applying an image recognition method to construct a model for automatically monitoring and evaluating the flowering phase of the peach tree. The method for monitoring and evaluating the flowering phase of the peach tree has high accuracy, can effectively consider the influence of environmental conditions on the flowering phase of the peach flower, and can accurately judge the flowering phase of the peach flower. Meanwhile, the method can be combined with various environmental impact factors to continuously and automatically monitor the flowering phase change of the peach blossom, and can provide basic data for peach tree cultivation management and germplasm evaluation.

Description

Peach tree flowering phase monitoring and evaluating method based on image recognition method
Technical Field
The invention belongs to the technical field of peach blossom season monitoring and measurement, and particularly relates to a peach tree florescence detection and evaluation method based on an image recognition method.
Background
Peach is one of the tree species for cultivating fruit trees, and the cultivation area and the total yield are the first in the world. And with the optimization and adjustment of peach variety structures and the gradual expansion of the economic benefit scale of the peach industry, the peach industry is still continuously developed at present. The flowering season of peach is an important period influencing the peach yield and the peach tourism industry, so that the accurate monitoring of the flowering season of peach trees is of great significance to the peach industry.
However, at present, the phenological period of the peach tree is mainly measured and calculated by means of manual observation, certain errors exist, inconsistency also exists in the past year record, and meanwhile, external environmental factors such as rainfall, strong wind, cooling and the like can have certain influence on the peach blossom during the blooming period of the peach blossom. The traditional manual observation and measurement cannot accurately evaluate the flowering phase of the peach tree and the change condition of the peach blossom in the flowering phase, and a general fitting and predicting peach blossom phase algorithm cannot intuitively, real-timely and accurately predict the change of the peach tree and the peach blossom in the phenological period. Therefore, the method for reasonably and effectively monitoring and evaluating the flowering phase of the peach trees in real time has important significance on peach tree management, labor cost, fruit yield, peach tree travel industry and the like.
The high-throughput phenotypic technology is a technical method for rapidly acquiring phenotypic information of a large number of plants by using technologies such as imaging sensing, image processing and machine learning, and has remarkable promotion effects on improving the efficiency and accuracy of plant phenotypic acquisition and accelerating new species breeding. The various plant phenotype character information can be acquired by various imaging devices, such as a digital camera, a multispectral/hyperspectral camera, a laser radar device and the like, by taking the image as a carrier. At present, in plant high-throughput phenotypic acquisition technologies and platforms, visible light imaging remains a main approach for realizing phenotypic analysis and acquisition. The target area is identified from the image by utilizing the image processing technology, and the interesting characteristics such as quantity, distribution, length, area, color and the like are extracted, so that high-flux phenotype information acquisition is realized. Visible image processing techniques are used in the collection of diverse phenotypic information for a variety of crops, such as melon fruit color, potato size and appearance, grape fruit order size, pear fruit size, jujube appearance shape parameters, and the like. The visible light imaging has the main advantages of high imaging rate, low cost and simple operation, meets the aim of constructing a lightweight and low-cost phenotype acquisition system in the research, and can meet the requirement of extracting important phenotype shapes of the peach tree in the phenological period.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides a peach blossom period monitoring and evaluating method based on an image recognition method.
The invention adopts the following technical scheme:
a peach tree flowering phase monitoring and evaluating method based on an image recognition method comprises the following steps:
1) and arranging a crop growth recorder in the peach orchard after the peach blossom is started for monitoring the flowering phase of the peach blossom. The instrument consists of a power supply solar cell panel and an agricultural intelligent camera, can shoot peach blossom images at the same position and the same visual angle in the field according to a set time interval, and uploads the peach blossom images to a specified network position through a wireless network. And 2 or more monitoring points of the crop growth recorder are arranged in the same peach orchard.
2) An agricultural meteorological monitor is arranged in the peach orchard before the peach orchard locally enters winter, and is used for monitoring meteorological data such as air temperature, illumination, rainfall and the like of the peach orchard. The instrument is powered by a power supply solar panel, can ensure normal work in continuous rainy days, and only needs to be arranged at one place in the same peach garden.
3) Fixing the angle of the camera, and setting a certain time point of 10-14 points every day for the crop growth recorder to shoot the peach blossom image, wherein the specific time point can be selected according to the field illumination condition. Meanwhile, peach blossom is observed every day, and the current flowering phase of the peach blossom is determined according to the flowering phase standard of the peach tree, wherein the flowering phase comprises a flower bud phase, an initial flowering phase, a full flowering phase and a withering phase. And after the peach blossom is completely withered after the whole flowering period is finished, acquiring all peach blossom images and meteorological data to carry out next modeling work.
4) According to literature data, taking the biological zero degree of the phenological period of the peach tree as 4 ℃; looking up meteorological data or calculating the date when the average temperature of the current day is greater than the biological zero degree of the peach tree for the first time according to the monitored meteorological data, and recording as D0. The distance D of each day in the flowering phase is calculated0The days elapsed are recorded as the growth duration of the peach trees.
5) A rectangular area is selected from the flowering phase shooting image, so that peach tree branches are contained as much as possible, and complex backgrounds are contained as little as possible. According to the characteristic of peach blossom color, an image recognition technology is applied to convert the image into an HSL (Hue, Saturation, brightness) color mode, S (Saturation) is more than 20, L (brightness) is more than 37, H (Hue) is respectively between 260-275, 276-290, 291-305, 306-320, 321-335 and 335-360 pixel areas in the rectangular area in the image of each day of the flowering period, and the ratio (%) of the pixel areas occupying the rectangular area is calculated and is respectively marked as H260, H275, H290, H305, H320 and H335 sequences.
6) And integrating the growth duration H260, H275, H290, H305, H320 and H335 ratio sequence (%) and air temperature (DEG C), illumination (lux) and rainfall (mm) data as input, taking the observation result of the peach flowering phase as output, constructing a peach flowering phase classification model by applying a machine learning method, and optimizing the model by distinguishing training and testing data and adjusting model parameters.
7) For other peach gardens or peach gardens in the next year, the model is applied to analyze images shot by the crop growth recorder to obtain corresponding peach flowering phase, and the peach flowering phase can be used for peach flowering monitoring and evaluation.
The invention has the beneficial effects that:
according to the method, the peach blossom images are collected in batches in a mode that a crop growth recorder continuously shoots the peach blossom period, a machine learning method is applied to construct a peach blossom period classification model, manual operation and random errors in peach blossom period monitoring and evaluating work are reduced, the accuracy and reliability of peach variety test data are improved, and meanwhile, the method for obtaining the peach blossom period phenotype information is provided.
Drawings
FIG. 1(a) is a flow chart of peach tree blooming period detection and evaluation based on an image recognition method;
FIG. 1(b) is a peach tree blooming period image shot by the crop growth recorder in FIG. 1 (a);
FIG. 1(c) is a rectangular area selected from FIG. 1 (a);
FIG. 1(d) is a schematic diagram showing the superposition of different pixel regions extracted according to different hues, saturations and luminances in FIG. 1(a), wherein a black region is an unextracted region;
FIG. 2(a) is an image taken at the flowering phase of peach blossom in the flowering bud phase;
FIG. 2(b) is an image taken at the flowering stage of a peach blossom at the initial flowering stage;
FIG. 2(c) is an image taken at the flowering stage of peach blossom at the initial flowering stage;
FIG. 2(d) is an image taken at the flowering stage of peach blossom in full-bloom stage;
fig. 2(e) is an image taken at the flowering stage of peach blossom in the withering and falling period.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention are described below clearly and completely, and it is obvious that the described embodiments are some, not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1(a) -1 (d), the peach blooming period monitoring and evaluating method based on the image recognition method of the present invention is a method for performing continuous image shooting of peach blooming period based on a peach orchard field crop growth recorder, selecting an area where a peach branch is located, extracting a pixel area-to-area ratio of HSLs in the selected area within a certain range by using an image processing method, and further constructing a peach blooming period classification model by using a machine learning method, thereby realizing automatic peach blooming period monitoring and evaluating.
The method specifically comprises the following steps: step 1, before the peach enters the flowering phase, a crop growth recorder is arranged at more than 2 places in the peach orchard, the shooting direction is aligned with the branches of the peach, and 14 shooting points per day are set for shooting images and uploading.
Step 2, before the peach orchard enters the winter locally, 1 agricultural meteorological monitor is distributed in the orchard and used for recording air temperature, illumination and rainfall data, recording the date that the average air temperature of the current year is higher than 4 ℃ for the first time and recording the date as D0。
Step 3, calculating the distance D every day0Recording the elapsed days as the growth duration, observing the flowering condition of the peach blossom, determining the current flowering phase of the peach blossom by referring to the standard reference of the flowering phase of the peach tree, wherein the flowering phase of the peach blossom can be divided into: the flower bud stage, the initial flowering stage, the full flowering stage and the falling regulating stage.
For the peach tree image shot by the crop growth recorder, an HSL color channel conversion image is adopted, firstly, a peach branch rectangular area is set, then, pixel areas with S (saturation) being more than 20, L (brightness) being more than 37 and H (hue) being respectively 260-.
The growth duration of each day, the air temperature (DEG C), the illumination (lux), the rainfall (mm) at the moment of image shooting, the ratio sequence (%) of the images H260, H275, H290, H305, H320 and H335 and the peach blossom period are organized into a data set. And randomly selecting 85% of data in the data set as a training set, and the rest 15% of the data in the data set as a testing set, constructing a peach tree flowering phase classification model on the training set by using a decision tree Random Forest (Random Forest) method, optimizing the model by adjusting parameters such as ntree (Forest scale) and mtry (sampling number), and testing the accuracy and classification error of the model on the testing set. Repeating the steps of selecting, training and testing until the accuracy and classification error of the model on the test set are not reduced.
The peach tree flowering phase classification model is applied to analyze other or next-year peach tree flowering phase images, and the peach tree flowering phase can be automatically and accurately monitored and evaluated.
Examples
And (3) using 108 peach tree shot images before and after the whole flowering phase in a decision tree random forest method to construct a peach tree flowering phase classification model, and adjusting according to a training set classification error to obtain optimal parameters (ntree is 1000 and mtry is 9). As shown in fig. 1(a) -fig. 1(d), the peach blossom period image shown in the figure is shot at 14:00 for 62 days, the selected rectangular area contains peach branches as much as possible and reduces complex background, the air temperature is 21.6 ℃, the illumination is 47001.6lux, the rainfall is 0mm, the ratio sequence (%) is H260-0.983, H275-1.482, H290-1.738, H305-1.734, H320-1.257 and H335-0.209, and the peach blossom period stage is judged to be full bloom period by the classification model, which is consistent with the actual situation.
As shown in fig. 2(a) -2 (e), the classification model judgment is performed on peach tree flowering images in different periods. Wherein, A is a period of 50 days and 14:00 shooting, the air temperature is 12.5 ℃, the illumination is 81838.0lux, the rainfall is 0mm, the sequence of the ratio is that (H260) ═ 0.072, (H275) ═ 0.032, (H290) ═ 0.023, (H305) ═ 0.083, (H320) ═ 0.423 and (H335) ═ 0.977, and the classification model judges that the flowering phase of the peach tree is a flowering bud phase; b, shooting at 14:00 days for 55 days, wherein the air temperature is 9.4 ℃, the illumination is 5253.1lux, the rainfall is 0mm, the ratio sequence (%) H260 is 0.090, H275 is 0.026, H290 is 0.035, H305 is 0.063, H320 is 0.289 and H335 is 0.122, and the classification model judges that the flowering phase of the peach tree is the initial flowering phase; c, shooting for 58 days at a rate of 14:00, wherein the air temperature is 20.9 ℃, the illumination is 49414.4lux, the rainfall is 0mm, the ratio sequence (%) is H260-0.149, H275-0.147, H290-0.174, H305-0.261, H320-0.879 and H335-1.069, and the classification model judges that the flowering phase of the peach tree is the beginning flowering phase; d, shooting at 14:00 for 62 days after growth, wherein the air temperature is 21.6 ℃, the illumination is 47001.6lux, the rainfall is 0mm, the ratio sequence (%) is H260 of 0.983, H275 of 1.482, H290 of 1.738, H305 of 1.734, H320 of 1.257 and H335 of 0.209, and the classification model judges that the flowering phase of the peach tree is full-bloom; e is shooting at 14:00 days for 77 days, the air temperature is 29.3 ℃, the illumination is 29998.1lux, the rainfall is 0mm, the ratio sequence (%) H260 is 0.202, H275 is 0.115, H290 is 0.093, H305 is 0.250, H320 is 0.458 and H335 is 0.373, and the classification model judges that the flowering phase of the peach tree is the withering phase. The judgment results of the classification models are all consistent with the actual situation.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (6)

1. A peach tree florescence monitoring and evaluating method based on an image recognition method is characterized by comprising the following steps:
step 1, laying a crop growth recorder in a peach orchard, and shooting a peach tree blooming period image at the same visual angle at a fixed time point by using an agricultural intelligent camera of the crop growth recorder;
the method specifically comprises the following steps: selecting a rectangular area in the peach blossom image, so that the rectangular area contains more peach branches as much as possible and contains less complex background as much as possible, converting the image into an HSL color mode by applying an image identification method according to the characteristics of peach flower color, extracting pixel areas of 260-;
step 2, arranging an agricultural meteorological monitor in the peach orchard, and recording the date that the daily average temperature is first higher than zero of the biological temperature of the peach trees, wherein the agricultural meteorological monitor is used for monitoring meteorological data of temperature, illumination and rainfall of the peach orchard;
and 3, combining meteorological data, and applying a machine learning method to construct a model for automatically monitoring and evaluating the flowering phase of the peach tree.
2. The method for monitoring and evaluating the flowering phase of a peach tree based on an image recognition method as claimed in claim 1, wherein the step 1 further comprises observing the peach tree every day when the method is applied for the first time, and determining the current flowering phase of the peach tree according to the peach tree flowering phase standard until the peach flower is completely withered.
3. The method for monitoring and evaluating the flowering phase of the peach tree based on the image recognition method as claimed in claim 1, wherein an area which contains more peach tree branches and less complex background is selected from the photographed peach tree flowering phase image.
4. The method for monitoring and evaluating the flowering period of the peach trees based on the image recognition method as claimed in claim 1, wherein the step 2 further comprises calculating the number of days elapsed from the date that the average air temperature of the current day of the flowering period of the peach trees is firstly greater than zero degree of the biological temperature of the peach trees in the current day, and recording the number of days elapsed for the peach trees to grow.
5. The peach tree flowering monitoring and evaluation method based on the image recognition method as claimed in claim 4, wherein the ratio sequence of H260, H275, H290, H305, H320 and H335 and the data of air temperature, light and rainfall in step 2 are used as input in combination with the growth duration, and the observation result of the peach tree flowering stage in step 1 is used as output, a machine learning method is applied to construct a peach tree flowering stage classification model, and the model is optimized by distinguishing training and testing data and adjusting model parameters.
6. The method for monitoring and evaluating the flowering phase of the peach tree based on the image recognition method as claimed in claim 1, wherein for other or next-year peach tree images, the peach tree flowering phase classification model is applied to judge the flowering phase of each day, and the phenological state of the peach tree in the flowering phase is monitored and evaluated.
CN202210017857.9A 2022-01-07 2022-01-07 Peach tree flowering phase monitoring and evaluating method based on image recognition method Pending CN114511777A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210017857.9A CN114511777A (en) 2022-01-07 2022-01-07 Peach tree flowering phase monitoring and evaluating method based on image recognition method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210017857.9A CN114511777A (en) 2022-01-07 2022-01-07 Peach tree flowering phase monitoring and evaluating method based on image recognition method

Publications (1)

Publication Number Publication Date
CN114511777A true CN114511777A (en) 2022-05-17

Family

ID=81550092

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210017857.9A Pending CN114511777A (en) 2022-01-07 2022-01-07 Peach tree flowering phase monitoring and evaluating method based on image recognition method

Country Status (1)

Country Link
CN (1) CN114511777A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115620151A (en) * 2022-12-16 2023-01-17 中化现代农业有限公司 Method and device for identifying phenological period, electronic equipment and storage medium
CN117893914A (en) * 2024-03-15 2024-04-16 浙江茂源林业工程有限公司 Plant growth monitoring method and system based on image recognition

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115620151A (en) * 2022-12-16 2023-01-17 中化现代农业有限公司 Method and device for identifying phenological period, electronic equipment and storage medium
CN117893914A (en) * 2024-03-15 2024-04-16 浙江茂源林业工程有限公司 Plant growth monitoring method and system based on image recognition
CN117893914B (en) * 2024-03-15 2024-05-24 浙江茂源林业工程有限公司 Plant growth monitoring method and system based on image recognition

Similar Documents

Publication Publication Date Title
CN106372592B (en) A kind of winter wheat planting area calculation method based on winter wheat area index
EP3707990B1 (en) Information processing device, information processing method, and vegetation management system
CN109115776B (en) Color and depth information-based plug seedling growth nondestructive monitoring method and device
White et al. Measuring fractional cover and leaf area index in arid ecosystems: digital camera, radiation transmittance, and laser altimetry methods
Crimmins et al. Monitoring plant phenology using digital repeat photography
CN114511777A (en) Peach tree flowering phase monitoring and evaluating method based on image recognition method
CN109029588A (en) A kind of Grain Growth Situation prediction technique based on climatic effect
CN112931150B (en) Irrigation system and method based on spectral response of citrus canopy
CN111476149A (en) Plant cultivation control method and system
CN113705937B (en) Farmland yield estimation method combining machine vision and crop model
CN113221765A (en) Vegetation phenological period extraction method based on digital camera image effective pixels
CN102663396B (en) Method for automatically detecting rice milky ripe stage
Yang et al. A VI-based phenology adaptation approach for rice crop monitoring using UAV multispectral images
CN112819227A (en) County scale winter wheat yield per unit prediction method and system
CN115439402A (en) Sugar cane yield-per-unit remote sensing estimation method and device based on SAFY-Sugar model
CN117292267B (en) Method and system for estimating rice aboveground biomass in segments based on weather information
CN117197595A (en) Fruit tree growth period identification method, device and management platform based on edge calculation
Liang et al. CanopyCAM–an edge-computing sensing unit for continuous measurement of canopy cover percentage of dry edible beans
CN116071662A (en) Crop identification and interpretation method and system based on remote sensing image
Wang et al. Research on application of smart agriculture in cotton production management
CN114895616A (en) Intelligent remote monitoring diagnostic system for agricultural greenhouse
CN110751322B (en) Litchi shoot control and flower promotion management method based on big data analysis and prediction
Sadiq et al. A review on the imaging approaches in agriculture with crop and soil sensing methodologies
Pasinetti et al. How water affects growing plants: an experimental set-up based on 2D vision and on-the-edge sensing
Shajahan et al. Monitoring plant phenology using phenocam: A review

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