A Method for Measurement of Fruit Tree Canopy Volume Based on Image Analysis

Abstract

The invention discloses an image analysis-based fruit tree crown volume measurement method. The method includes the following steps that: the images of five fruit trees are acquired, and tree crown volume is measured manually; tree crown area is obtained by using an image processing technology; based on the least squares method, a relationship model between the logarithm LnV of the tree crown volume and area S can be obtained through fitting; and when measurement is carried out, only the fruit tree images are required to be acquired, and the tree crown area can be obtained through image processing, and the tree crown volume can be obtained according to the relationship model. With the image analysis-based fruit tree crown volume measurement method of the invention adopted, the defects of low efficiency and high strength of manual measurement and the defects of complex operation and large data processing amount of an existing electronic measurement method can be eliminated; the relationship model can be calibrated at any time with the changes of seasons and environments, so that measurement accuracy can be differentiated. The image analysis-based fruit tree crown volume measurement method is simple in operation steps and is suitable for being used by ordinary farmers and has a bright market prospect.

Description

Technical field

[0001] The invention belongs to the field of fruit tree planting, in particular to a method for measuring the crown volume of fruit trees based on image analysis. Background technique

[0002] The crown is an important factor affecting the production capacity of fruit trees, and its growth status directly affects the yield and quality of the fruit. Therefore, the volume of the crown is an important indicator for evaluating tree structure and yield measurement. It is used in fruit tree breeding, yield measurement and fine management of the crown. Extremely important reference value, such as: fruit tree yield estimation, pesticide variable spraying, precise fertilization, fruit tree biomass calculation, leaf density calculation, etc. It can be said that measuring the crown volume of fruit trees is a prerequisite for healthy fruit tree growth and reasonable management.

[0003] At present, the commonly used methods of fruit tree volume measurement include traditional m...

Examples

Embodiment 1

[0038] Example 1 Collect fruit tree crown images and obtain fruit tree crown area S

[0039] Experiment time: 2015.9.20 7:30-11:40AM (cloudy day, illumination below 20,000 lux)

[0040] Experiment location: Lishu Park, Jiangpu Farm, Nanjing

[0041] In the first step, 30 pear trees in the park are selected as samples to collect images, and the crown shape should include all the shapes in the pear tree park as much as possible, and from these 30 samples, 5 pear trees are selected as the five-point calibration method The selection method of these 5 pear trees is: taking the crown diameter D at the center of the crown of the fruit tree as the selection criterion, first select the fruit tree with the smallest D value in the orchard as sample point 1, and then increase the D value by 20-30 % For sampling.

[0042] In this embodiment, a Canon EOS 350D (REBEL XT) camera is used to collect images in artificial intelligence autofocus mode. The horizontal and vertical resolutions are 72dpi res...

Embodiment 2

[0063] Example 2 Establishing a model of the correlation between crown volume and area universality based on parameter calibration method

[0064] The actual canopy area S and the manually measured canopy volume V of the 30 samples obtained in Example 1 are incorporated into the table, and the relationship model between the logarithm of the canopy volume LnV and the canopy area S is constructed based on the least square method, and the results are as follows Image 6 As shown, the correlation equation is obtained: LnV = 0.6654S-0.3538, R 2 =0.9444; It can be found that the correlation coefficient is above 0.94, showing a clear positive correlation.

[0065] At the same time, establish the relationship model between the logarithm of the canopy volume LnV and the canopy area S with the 5 samples selected in the embodiment, and the results are as follows Figure 7 As shown, the correlation equation is obtained: LnV = 0.6349S-0.3826, R 2 = 0.9866; compare Image 6 with Figure 7 It can b...

Embodiment 3 5

[0067] Example 3 Five-point parameter calibration method model volume prediction accuracy evaluation

[0068]

[0069] In formula (3), E is the measurement error, V 1 Is the predicted value of the canopy volume model (unit m 3 ), V 2 Is the artificial measurement value of canopy volume (unit m 3 );

[0070] According to formula (3), 30 pear tree samples are selected as the research object, and the accuracy of the model volume prediction built by two methods (30 samples and 5 samples) is evaluated and analyzed. In order to verify the validity of the five-point parameter calibration method, Figure 8 To compare the errors of the two prediction models, Figure 8 The classical model method refers to the model built by 30 pear trees, and the five-point calibration model refers to the model built by 5 samples. Table 1 shows the probability statistics of the sample size in the total sample of the two volume prediction models in different prediction error intervals:

[0071] Table 1 Error p...

Technical field

[0001] The invention belongs to the field of fruit tree planting, in particular to a method for measuring the crown volume of fruit trees based on image analysis. Background technique

[0002] The crown is an important factor affecting the production capacity of fruit trees, and its growth status directly affects the yield and quality of the fruit. Therefore, the volume of the crown is an important indicator for evaluating tree structure and yield measurement. It is used in fruit tree breeding, yield measurement and fine management of the crown. Extremely important reference value, such as: fruit tree yield estimation, pesticide variable spraying, precise fertilization, fruit tree biomass calculation, leaf density calculation, etc. It can be said that measuring the crown volume of fruit trees is a prerequisite for healthy fruit tree growth and reasonable management.

[0003] At present, the commonly used methods of fruit tree volume measurement include traditional m...

Examples

Embodiment 1

[0038] Example 1 Collect fruit tree crown images and obtain fruit tree crown area S

[0039] Experiment time: 2015.9.20 7:30-11:40AM (cloudy day, illumination below 20,000 lux)

[0040] Experiment location: Lishu Park, Jiangpu Farm, Nanjing

[0041] In the first step, 30 pear trees in the park are selected as samples to collect images, and the crown shape should include all the shapes in the pear tree park as much as possible, and from these 30 samples, 5 pear trees are selected as the five-point calibration method The selection method of these 5 pear trees is: taking the crown diameter D at the center of the crown of the fruit tree as the selection criterion, first select the fruit tree with the smallest D value in the orchard as sample point 1, and then increase the D value by 20-30 % For sampling.

[0042] In this embodiment, a Canon EOS 350D (REBEL XT) camera is used to collect images in artificial intelligence autofocus mode. The horizontal and vertical resolutions are 72dpi res...

Embodiment 2

[0063] Example 2 Establishing a model of the correlation between crown volume and area universality based on parameter calibration method

[0064] The actual canopy area S and the manually measured canopy volume V of the 30 samples obtained in Example 1 are incorporated into the table, and the relationship model between the logarithm of the canopy volume LnV and the canopy area S is constructed based on the least square method, and the results are as follows Image 6 As shown, the correlation equation is obtained: LnV = 0.6654S-0.3538, R 2 =0.9444; It can be found that the correlation coefficient is above 0.94, showing a clear positive correlation.

[0065] At the same time, establish the relationship model between the logarithm of the canopy volume LnV and the canopy area S with the 5 samples selected in the embodiment, and the results are as follows Figure 7 As shown, the correlation equation is obtained: LnV = 0.6349S-0.3826, R 2 = 0.9866; compare Image 6 with Figure 7 It can b...

Embodiment 3 5

[0067] Example 3 Five-point parameter calibration method model volume prediction accuracy evaluation

[0068]

[0069] In formula (3), E is the measurement error, V 1 Is the predicted value of the canopy volume model (unit m 3 ), V 2 Is the artificial measurement value of canopy volume (unit m 3 );

[0070] According to formula (3), 30 pear tree samples are selected as the research object, and the accuracy of the model volume prediction built by two methods (30 samples and 5 samples) is evaluated and analyzed. In order to verify the validity of the five-point parameter calibration method, Figure 8 To compare the errors of the two prediction models, Figure 8 The classical model method refers to the model built by 30 pear trees, and the five-point calibration model refers to the model built by 5 samples. Table 1 shows the probability statistics of the sample size in the total sample of the two volume prediction models in different prediction error intervals:

[0071] Table 1 Error p...

Images