CN105574898A - Method and system for monitoring plant lodging situation based on image detection - Google Patents

Abstract

The invention discloses a method for monitoring plant lodging based on image detection, comprising the following steps: S1, acquiring an original image of a plant in a monitoring area, and converting the original image into a grayscale image; S2, preset a Gaussian blur function, and Select multiple relaxation parameters σ _n ; S3. Convolute the grayscale image with the Gaussian blur function corresponding to each relaxation parameter σ _n to obtain multiple blurred images g _n ; S4. Preset the boundary threshold; S5. According to The demarcation threshold performs binarization on the fuzzy image g _n to obtain a plurality of binarized images gb _n ; S6, substituting the plurality of binarized images gb _n into the preset first calculation model for comprehensive calculation to obtain binarized lodging Image R(x,y); S7. Calculate the lodging rate R _rite according to the preset second calculation model and the lodging image R(x,y). The present invention avoids the need for staff to visit the site in person by remotely monitoring the plants, is beneficial to reduce labor intensity and labor requirements, and can realize plant monitoring in a wide range.

Description

A method and system for monitoring plant lodging based on image detection

technical field

The invention relates to the technical field of plant growth monitoring, in particular to a method and system for monitoring plant lodging based on image detection.

Background technique

Wheat lodging is one of the common disasters in agricultural production. Once wheat lodging occurs, the water, nutrient operation and photosynthesis of the plant will be reduced, and various diseases and insect pests will be induced, which will seriously affect the grain filling process, and ultimately affect the formation of wheat yield and grain quality. In severe lodging, the yield loss can reach 27%. . In addition, because lodging is not conducive to mechanical harvesting, the increase in labor harvesting costs will also aggravate the loss of farmland income. Large-scale and rapid monitoring of wheat lodging is the key to mastering disasters, timely prevention and control, and assessing losses. It is of great value for the agricultural sector to obtain timely information on farmland wheat growth.

Contents of the invention

Based on the technical problems existing in the background technology, the present invention proposes a method and system for monitoring plant lodging based on image detection.

A kind of plant lodging situation monitoring method based on image detection that the present invention proposes, comprises the following steps:

S1. Obtain the original image of the plants in the monitoring area, and convert the original image into a grayscale image;

S2. Preset a Gaussian blur function, and select multiple relaxation parameters σ _n ;

S3. Convolving the grayscale image with the Gaussian blur function corresponding to each relaxation parameter σ _n to obtain a plurality of blurred images g _n ;

S4. Preset demarcation threshold;

S5. Binarize the blurred image g _n according to the demarcation threshold to obtain a plurality of binarized images gb _n ;

S6. Substituting a plurality of binarized images gb _n into a preset first calculation model to perform comprehensive calculations to obtain a binarized lodging image R(x,y);

S7. Calculate the lodging rate R _rite according to the preset second calculation model and the lodging image R(x,y).

Preferably, in step S2, the relaxation parameter σ _n is selected according to the magnitude of the imaging magnification and the actual image.

Preferably, in step S2, multiple relaxation parameters _σn satisfy the following relationship:

σ _n =2σ _n-1 =2 ² σ _n-2 =...=2 ⁿ σ ₀ , where σ ₀ is a constant.

Preferably, it is characterized in that σ ₀ =1.

Preferably, in step S4, the number of demarcation thresholds is multiple, and corresponds to a plurality of blurred images g _n one by one; step S5 is specifically: binarize the blurred image g _n according to the corresponding demarcation thresholds, and obtain multiple Binarize the image gb _n .

Preferably, in step S6, the first calculation model is:

$R(x,\mathrm{the\; y})=\underset{\mathrm{no}}{\Σ}({\mathrm{\λ}}_{\mathrm{no}}\×{\mathrm{gb}}_{\mathrm{no}}),{\mathrm{\λ}}_{\mathrm{no}}$ is a preset constant.

Preferably, in step S7, the second calculation model is:

$R_{rite}=\underset{\mathrm{the\; y}}{\Σ}{k}_{\mathrm{the\; y}}\underset{x}{\Σ}R(x,\mathrm{the\; y}),$ k _y is a proportionality factor.

A plant lodging monitoring system based on image detection, including: ash processing module, Gaussian blur module, binarization processing module and lodging calculation module; wherein,

The ashing processing module is used to receive the original image sent by the camera device in the monitoring area, and convert the original image into a grayscale image;

A series of Gaussian functions are preset in the Gaussian blur module, and the relaxation parameters σ of any two Gaussian functions are different; the Gaussian blur module is connected to the ashing processing module, which receives the grayscale image and compares the grayscale image with each relaxation parameter The Gaussian blur function corresponding to σ _n performs convolution budget to obtain multiple blurred images g _n ;

The binarization processing module is connected with the Gaussian blur module, which is preset with a plurality of boundary thresholds corresponding to the blurred image g _n one by one, and is used to perform binarization on the blurred image g _n according to the corresponding boundary thresholds to obtain Multiple binarized images gb _n ;

The lodging calculation module is connected with the binarization processing module, and the first calculation model and the second calculation model are preset inside; the lodging calculation module substitutes a plurality of binarization results gb _n into the first calculation model for comprehensive operation to obtain the binary value The lodging image R(x,y) is optimized, and then the lodging rate R _rite is calculated according to the second calculation model and the lodging image R(x,y).

Preferably, the first calculation model is: λ _n is a preset constant.

Preferably, the second calculation model is: k _y is a proportionality factor.

In the method and system for monitoring plant lodging based on image detection provided by the present invention, the plant image in the monitoring area is obtained remotely as the original image, and the original image is sequentially ashed, Gaussian blurred, and binarized, and then according to the two The lodging image is calculated from the valued processing results, and the lodging rate is further calculated. The present invention can form a set of high-throughput breeding software through an automated image analysis method, which greatly improves the speed and accuracy of information acquisition, and is conducive to the development of subsequent related research work.

Moreover, the present invention avoids the need for staff to visit the site in person by remotely monitoring the plants, which is beneficial to reduce labor intensity and labor requirements, and can realize plant monitoring in a wide range.

Description of drawings

Fig. 1 is a flow chart of a plant lodging situation monitoring method based on image detection proposed by the present invention;

Fig. 2 is a structural diagram of a plant lodging monitoring system based on image detection proposed by the present invention.

detailed description

With reference to Fig. 1, a kind of plant lodging situation monitoring method based on image detection that the present invention proposes, comprises the following steps:

S1. Acquire the original image of the plants in the monitoring area, and convert the original image into a grayscale image.

In this embodiment, monitoring the growth of wheat is taken as an example. The method for monitoring plant lodging based on image detection provided in this embodiment is a remote monitoring method that does not require staff to be present in person. Therefore, remote acquisition of the original image of wheat is the key. The acquisition of the original image can be realized by presetting the camera device in the monitoring area, and the camera device automatically captures the original image and sends it to the subsequent equipment.

S2. Preset a Gaussian blur function, and select multiple relaxation parameters σ _n .

Gaussian function Is a commonly used image processing function, the relaxation parameter σ determines the shape of the two-dimensional Gaussian function. In this embodiment, the selection of multiple relaxation parameters _σn satisfies the following principles:

σ _n =2σ _n-1 =2 ² σ _n-2 =...=2 ⁿ σ ₀ , where σ ₀ is a constant.

In this embodiment, firstly, the relaxation parameter σ ₀ is selected according to the imaging magnification and the actual image, and then σ ₁ , σ ₂ , σ ₃ ... σ _n-1 , σ _n are sequentially obtained according to σ ⁰ and the above formula, and the σ ₁ , σ ₂ , σ ₃ ... σ _n-1 , σ _n are substituted into σ in the Gaussian function to obtain a series of Gaussian functions. During specific implementation, σ ₀ =1 may also be selected.

S3. Convolving the grayscale image with the Gaussian blur function corresponding to each relaxation parameter σ _n to obtain a plurality of blurred images g _n .

In this step, a series of Gaussian functions obtained in the previous step are used to perform Gaussian blur on the grayscale image, that is, using the Gaussian functions corresponding to σ ₀ , σ ₁ , σ ₂ , σ ₃ ... σ _n-1 , σ _n The function performs Gaussian blurring on the grayscale image to obtain a series of blurred images g ₀ , g ₁ , g ₂ , g ₃ ... g _n-1 , g _n .

S4. Presetting the demarcation threshold. In this embodiment, there are multiple boundary thresholds, which correspond to multiple blurred images g _n one by one.

_S5 . _Binarize the _fuzzy images _{g 0} , g ₁ , g ₂ , g ₃ . ₂ , gb ₃ ...gb _n-1 , gb _n . Specifically, when performing binarization, all point color values in the blurred image whose color values are lower than the boundary threshold are converted to 0, and point color values in the blurred image whose color values are higher than the boundary threshold are all converted to 255.

S6. Substituting multiple binarized results gb _n into the preset first calculation model to perform comprehensive calculations to obtain a binarized lodging image R(x, y).

The first calculation model is:

$R(x,\mathrm{the\; y})=\underset{\mathrm{no}}{\Σ}({\mathrm{\λ}}_{\mathrm{no}}\×{\mathrm{gb}}_{\mathrm{no}}),$ λ _n is a preset constant.

During specific implementation, λ _n can be selected as a constant constant, that is, λ _n =λ _n-1 =λ _n-2 =...=λ ₀ ;

λ _n can also be valued according to the formula: λ _n =ω ₁ λ _n-1 =ω ₂ λ _n-2 =...=ω _n λ ₀ , where ω ₁ , ω ₂ , ω _n and λ ₀ is a constant.

S7. Calculate the lodging rate R _rite according to the preset second calculation model and the lodging image R(x,y).

The second calculation model is:

$R_{rite}=\underset{\mathrm{the\; y}}{\Σ}{k}_{\mathrm{the\; y}}\underset{x}{\Σ}R(x,\mathrm{the\; y}),$ k _y is a proportionality factor.

The above method will be further explained below in conjunction with an image detection-based plant lodging monitoring system.

Referring to Fig. 2, the system for monitoring plant lodging based on image detection includes: an ashing processing module, a Gaussian blur module, a binarization processing module and a lodging calculation module.

The ashing processing module is used to receive the original image sent by the camera device in the monitoring area, and convert the original image into a grayscale image.

A series of Gaussian functions are preset in the Gaussian blur module The relaxation parameter σ of any two Gaussian functions is different, and the selection of the relaxation parameter σ satisfies the following formula:

σ _n =2σ _n-1 =2 ² σ _n-2 =...=2 ⁿ σ ₀ , where σ ₀ is a constant and is determined by the imaging magnification and the actual image.

The Gaussian blur module is connected with the graying processing module, which receives the grayscale image, and performs convolution budget on the grayscale image with the Gaussian blur function corresponding to each relaxation parameter σ _n to obtain multiple blurred images g _n .

The binarization processing module is connected with the Gaussian blur module, which is preset with a plurality of boundary thresholds corresponding to the blurred image g _n one by one, and is used to perform binarization on the blurred image g _n according to the corresponding boundary thresholds to obtain Multiple binarized images gb _n .

The first calculation model and the second calculation model are preset in the lodging calculation module.

The first calculation model is:

$R(x,\mathrm{the\; y})=\underset{\mathrm{no}}{\Σ}({\mathrm{\λ}}_{\mathrm{no}}\×{\mathrm{gb}}_{\mathrm{no}}),$ λ _n is a preset constant.

The second calculation model is:

k _y is a proportionality factor.

The lodging calculation module is connected with the binarization processing module, which substitutes a plurality of binarization results gb _n into the first calculation model for comprehensive operation to obtain a binarized lodging image R(x, y), and then according to the second calculation model and the lodging image R(x,y) to calculate the lodging rate R _rite .

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto, any person familiar with the technical field within the technical scope disclosed in the present invention, according to the technical solution of the present invention Any equivalent replacement or change of the inventive concepts thereof shall fall within the protection scope of the present invention.

Claims (10)

1. A plant lodging condition monitoring method based on image detection is characterized by comprising the following steps:

s1, acquiring an original image of the plant in the monitoring area, and converting the original image into a gray image;

s2, presetting a Gaussian fuzzy function, and selecting a plurality of relaxation parameters sigma_n；

S3, respectively connecting the gray-scale image with each relaxation parameter sigma_nCarrying out convolution budget on the corresponding Gaussian blur function to obtain a plurality of blurred images g_n；

S4, presetting a demarcation threshold;

s5, blurring the image g according to the boundary threshold value_nBinarizing to obtain multiple binary images gb_n；

S6, converting the plurality of binary images gb_nSubstituting the binary image into a preset first calculation model to perform comprehensive operation to obtain a binary lodging image R (x, y);

s7, calculating the lodging rate R according to the preset second calculation model and the lodging image R (x, y)_rite。

2. The method for monitoring plant lodging conditions based on image detection as claimed in claim 1, wherein in step S2, a relaxation parameter σ is selected according to the magnitude of imaging magnification and the actual image_n。

3. The method for monitoring plant lodging based on image detection as claimed in claim 2, wherein in step S2, a plurality of relaxation parameters σ_nThe following relationship is satisfied:

σ_n＝2σ_n-1＝2²σ_n-2＝......＝2ⁿσ₀where σ is₀Is a constant.

4. The image detection-based plant lodging condition monitoring method of claim 3, wherein σ₀＝1。

5. The image detection-based plant lodging monitoring method of claim 1, wherein in step S4, the number of demarcation thresholds is multiple and is associated with a plurality of blurred images g_nOne-to-one correspondence is realized; step S5 specifically includes: the blurred image g is subjected to a corresponding demarcation threshold_nBinarizing to obtain multiple binary images gb_n。

6. The method for monitoring plant lodging based on image detection as claimed in claim 1, wherein in step S6, the first calculation model is:

$R(x,y)=\underset{n}{\Σ}({\mathrm{\λ}}_n\×{\mathrm{gb}}_n),$ λ_nis a preset constant.

7. The method for monitoring plant lodging based on image detection as claimed in claim 1, wherein in step S7, the second calculation model is:

$R_{rite}=\underset{y}{\Σ}{k}_y\underset{x}{\Σ}R(x,y),$ k_yis a scaling factor.

8. A plant lodging condition monitoring system based on image detection is characterized by comprising: the device comprises an ashing processing module, a Gaussian blur module, a binarization processing module and a lodging calculation module; wherein,

the ashing processing module is used for receiving an original image sent by the monitoring area camera device and converting the original image into a gray image;

a series of Gaussian functions are preset in the Gaussian blur moduleThe relaxation parameters sigma of any two Gaussian functions are different; the Gaussian blur module is connected with the ashing processing module, receives the gray level image and respectively connects the gray level image with each relaxation parameter sigma_nCarrying out convolution budget on the corresponding Gaussian blur function to obtain a plurality of blurred images g_n；

The binarization processing module is connected with the Gaussian blur module, and a plurality of blurred images g are preset in the binarization processing module_nA boundary threshold value corresponding to each other and used for respectively aligning the blurred images g according to the corresponding boundary threshold values_nPerforming binarization processing to obtain multiple binarized images gb_n；

The lodging calculation module is connected with the binarization processing module, and a first calculation model and a second calculation model are preset in the lodging calculation module; the lodging calculation module converts a plurality of binarization results gb_nSubstituting the first calculation model to perform comprehensive operation to obtain a binary lodging image R (x, y), and calculating lodging rate R according to the second calculation model and the lodging image R (x, y)_rite。

9. The image detection-based plant lodging monitoring system of claim 8, wherein the first computational model is:λ_nis a preset constant.

10. The image detection-based plant lodging monitoring system of claim 8, wherein the second computational model is:k_yis a scaling factor.