A method for predicting rice chalkiness and a method for screening rice with low chalkiness

Abstract

The application discloses a method for predicting rice chalkiness and a method for screening rice with low chalkiness, comprising the following steps: taking a photo of a rice plant from a top view at the jointing stage of the rice; performing a convex hull drawing on the image of the rice plant in the taken photo, performing an elliptical drawing according to the convex hull shape, and then calculating the eccentricity of the ellipse; and predicting the chalkiness of the rice after maturation and harvesting based on the obtained correlation between the eccentricity of the ellipse in the image of the rice plant and the chalkiness of the rice after maturation and harvesting. The correlation between the eccentricity of the ellipse in the image of the rice plant and the chalkiness of the rice after maturation and harvesting is as follows: ebd=0.41+26.42*ECCENTRICITY. The application can predict the rice chalkiness and screen rice with low chalkiness at the jointing stage of the rice, has high prediction accuracy, and can improve the breeding efficiency and speed up the breeding process.

Description

Technical Field

[0001] This invention relates to the field of rice quality testing and breeding, and in particular to a method for predicting the chalkiness of rice and a method for screening rice with low chalkiness. Background Technology

[0002] Over the past two decades, with rising living standards, my country's rice market has seen a surge in demand for high-quality rice, leading to increased imports. To meet this growing market demand, rice breeding programs have placed increasingly stringent requirements on rice quality. Chalk whiteness, a key parameter for measuring rice quality, shows a significant negative correlation between lower chalk whiteness and improved quality and increased commercial value. In other words, lower chalk whiteness equates to higher rice quality and economic value, consequently commanding a higher market price. Consequently, over the past decade, rice processing enterprises have implemented increasingly stringent standards for chalk whiteness, particularly favoring rice varieties with lower chalkiness to optimize product quality and economic benefits.

[0003] Traditional methods for detecting chalkiness in rice require the rice to complete its entire growth cycle of approximately 120 days, meaning that chalkiness can only be measured after the rice has matured, been harvested, and hulled. This implies that chalkiness-related breeding selection decisions can only be made at the end of the entire growth period, resulting in a certain degree of lag.

[0004] It is evident that the existing technologies described above still have inconveniences and shortcomings, and urgently need further improvement. The current industry urgently needs to develop a method for predicting rice chalkiness and screening for low-chalk rice varieties, enabling the prediction and screening of low-chalk rice before harvest to improve breeding efficiency and accelerate the breeding process. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for predicting the chalkiness of rice and a method for screening low chalkiness rice, so that the chalkiness of rice can be predicted and low chalkiness rice can be screened before the rice matures and is harvested, thereby improving breeding efficiency and accelerating the breeding process.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] On one hand, the present invention provides a method for predicting the chalkiness of rice, comprising:

[0008] Step (1): Take a photo of the rice plant from an overhead angle during the jointing stage of the rice plant;

[0009] Step (2): Perform convex hull plotting on the rice plant image in the photograph, draw an ellipse based on the shape of the convex hull, and then calculate the eccentricity of the ellipse;

[0010] Step (3): Based on the calculated ellipse eccentricity, predict the chalkiness of rice after harvest according to the correlation between the ellipse eccentricity in the obtained rice plant images and the chalkiness of rice after harvest.

[0011] As a further improvement of the present invention, in step (1), a photograph of rice plants is taken from a top-down angle using an RGB imaging system.

[0012] Furthermore, before step (2), the process includes: processing the photograph to remove the background and leave only the rice plant part; in step (2), convex hull drawing is performed on the rice plant image with the background removed from the photograph.

[0013] Furthermore, in step (2), the convex hull drawing is to draw the shape of a polygon obtained by drawing the points surrounding the outer perimeter of the rice plant image by stretching rubber bands.

[0014] Furthermore, the elliptical drawing based on the convex hull shape is performed such that the boundary of the ellipse coincides exactly with the boundary of the polygon obtained from the convex hull drawing.

[0015] Furthermore, in step (2), for elliptical drawing, the semi-focal length and semi-major axis of the ellipse are first calculated, and then the eccentricity of the ellipse is calculated using the following equation:

[0016] ECCENTRI CI TY = Half focal length / Half major axis;

[0017] Where ECCENTRI CI TY represents the eccentricity of the ellipse.

[0018] Furthermore, in step (3), the correlation between the ellipse eccentricity in the obtained rice plant image and the chalkiness of the rice after harvest is expressed by the following equation:

[0019] ebd=0.41+26.42*ECCENTRI CI TY

[0020] Wherein, ebd represents the chalkiness of rice, and the unit of chalkiness is percentage;

[0021] ECCENTRI CI TY represents the eccentricity of the ellipse; where 0.41 is the intercept of the equation and 26.42 is the coefficient of the eccentricity of the ellipse.

[0022] Furthermore, in step (3), the correlation between the eccentricity of the ellipse in the rice plant image and the chalkiness of the rice after harvest is obtained as follows:

[0023] The aforementioned relationships are formed based on historical datasets;

[0024] The historical dataset is a collection of data using the elliptic eccentricity values ​​of rice plants calculated at the jointing stage of rice and the corresponding measured chalkiness data of rice.

[0025] Furthermore, the correlation is expressed by the following equation:

[0026] ebd=A+B*ECCENTRI CI TY

[0027] Wherein, ebd represents the chalkiness of rice, and the unit of chalkiness is percentage;

[0028] ECCENTRI CI TY represents the eccentricity of the ellipse; where A is the intercept of the equation and B is the coefficient of the eccentricity of the ellipse.

[0029] Secondly, the present invention also provides a method for screening low chalkiness rice, which screens low chalkiness rice based on the aforementioned rice chalkiness prediction results.

[0030] The present invention provides a method for predicting the chalkiness of rice and a method for screening rice with low chalkiness. This method enables early, large-scale, and rapid prediction and screening of rice chalkiness during the jointing stage. Compared with traditional breeding selection techniques, this invention achieves a significant improvement in breeding efficiency. It eliminates the need to wait for the rice to mature and be harvested, allowing for the prediction of chalkiness and timely breeding selection decisions during the jointing stage, thereby greatly improving the efficiency of breeding work. Attached Figure Description

[0031] The above is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] Figure 1 This is a schematic diagram of a rice plant captured by the RGB imaging system of an indoor high-throughput phenotypic imaging analysis platform according to an embodiment of the present invention.

[0033] Figure 2 yes Figure 1 A schematic diagram of RGB images of rice plants taken by the RGB imaging system of the high-throughput phenotypic imaging analysis platform in the middle room;

[0034] Figure 3 yes Figure 2 A schematic diagram of the image obtained after removing the background from an RGB photograph of a rice plant.

[0035] Figure 4 This is a schematic diagram of convex hull drawing on an image of a rice plant after removing the background;

[0036] Figure 5This is a schematic diagram of drawing an ellipse based on the shape of the convex hull;

[0037] Figure 6 It is the value of the elliptic eccentricity of rice plants automatically calculated by the Platscreen DataAnalyzer and MorphoAnalyzer analysis software equipped in the RGB imaging system of the indoor high-throughput phenotypic imaging analysis platform.

[0038] Figure 7 It is the value of the elliptic eccentricity of another rice plant automatically calculated by the Platscreen DataAna yzer and MorphoAna yzer analysis software equipped in the RGB imaging system of the indoor high-throughput phenotypic imaging analysis platform. Detailed Implementation

[0039] This invention provides a method for predicting the chalkiness of rice, comprising: step (1): taking a photograph of the rice plant from a top-down angle during the rice jointing stage; step (2): performing convex hull plotting on the rice plant image in the photograph, drawing an ellipse based on the shape of the convex hull, and then calculating the ellipse eccentricity; step (3): based on the calculated ellipse eccentricity, predicting the chalkiness of the rice after harvest according to the correlation between the ellipse eccentricity in the obtained rice plant image and the chalkiness of the rice after maturity. Based on the prediction results of the method for predicting the chalkiness of rice, low-chalkiness rice can be screened out.

[0040] The prediction method of this invention breaks through the traditional method of obtaining the chalkiness of rice only after harvest, and is an accidental discovery. While studying other RGB imaging experiments, the inventors obtained a large number of ellipse eccentricity values ​​of rice plants at the jointing stage. Later, after rice harvest, they discovered a strong correlation between the chalkiness of rice and the ellipse eccentricity values ​​of rice plants at the jointing stage. Based on this, experimental verification was conducted, further demonstrating that the method has high prediction accuracy.

[0041] The above method will be described in detail below through examples:

[0042] 1. This example uses 91 rice varieties, the names of which are listed in Table 1. These 91 rice varieties were purchased from domestic seed companies. Pot cultivation was conducted on all 91 rice varieties. The method was as follows: On May 22, 2023, the 91 rice varieties were sown at the experimental farm of the China National Rice Research Institute in Fuyang District, Hangzhou City, Zhejiang Province. Ten days after sowing, the rice materials were transplanted into plastic buckets (18cm high and 20cm in diameter). Each rice variety was planted in 3 buckets (i.e., 3 replicates per variety), with 2 seedlings planted in each bucket, resulting in 6 individual plants per variety. The potting soil used was paddy field clay loam, with approximately 4 kg of dry soil in each bucket.

[0043] 2. Seventy days after sowing, that is, on July 31, 2023, when most rice varieties were in the jointing stage, RGB imaging of the rice plants was performed. The imaging system used for RGB imaging was the Photon Systems Instruments (PSI) PlantScreen from the Czech Republic. TM The indoor high-throughput phenotypic imaging analysis platform has an RGB imaging system located on the ceiling of the imaging chamber (see...). Figure 1 This allows for the capture of RGB images of rice plants from a top-down view, looking towards the roots. It enables RGB imaging of rice varieties from a top-down perspective. Figure 2 The computer software attached to the imaging system was used to remove the background, resulting in an image that retained only the rice plants. Figure 3 This optimized approach avoids interference and facilitates subsequent drawing. Convex hull drawing is then performed sequentially on the rice plant image after background removal. Figure 4 ) and elliptical drawing ( Figure 5 ), and automatically calculate the eccentricity of the ellipse based on the results of the ellipse drawing. Figure 6 ).

[0044] In the above embodiments, convex hull drawing, elliptical drawing, and automatic calculation of elliptical eccentricity based on the results of elliptical drawing are all performed using the PlantscreenData Analyzer and MorphoAnalyzer analysis software provided by PSI (Photon Systems Instruments) in the Czech Republic. Specifically, this includes: performing convex hull drawing on the rice image after removing the background, and drawing the shape of the polygon obtained by surrounding each point of the rice image with stretched rubber bands. Figure 4 See the original image used. Figure 2 and Figure 3 An ellipse is drawn based on the shape of the convex hull, and the boundary of the drawn ellipse is exactly the same as... Figure 4The polygon boundaries obtained from the convex hull drawing coincide, and the semi-focal length and semi-major axis of the ellipse are automatically calculated. The eccentricity of the ellipse is calculated according to the equation "ECCENTRI CI TY = semi-focal length / semi-major axis", where ECCENTRI CI TY represents the eccentricity of the ellipse, and the semi-focal length is as follows: Figure 5 As shown in b, the semi-major axis is as follows Figure 5 As shown in Figure a, the semi-minor axis is as follows Figure 5 As shown in c.

[0045] Figure 6 The PlantscreenData Analyzer and MorphoAnalyzer analysis software, which are provided by the Czech company PSI (Photon Systems Instruments), automatically calculates the value of the ellipse eccentricity. The eccentricity value of the ellipse in the rice plant image is 0.7283. The closer the eccentricity value is to 1, the farther the top view image of the rice is from a circle, that is, the closer the shape is to an ellipse.

[0046] Figure 7 The PlantscreenData Analyzer and MorphoAnalyzer analysis software, which are included with the Czech company PSI (Photon Systems Instruments), automatically calculates the value of the eccentricity of the ellipse. The eccentricity value of the other rice plant shown in the figure is 0.1605. The closer the eccentricity value is to 0, the closer the top view image of the rice is to a circle, that is, the closer the shape is to a circle.

[0047] 3. The equation “ebd=0.41+26.42*ECCENTRI CI TY” was used to calculate and predict the chalkiness of rice after harvest. Here, ebd represents the chalkiness of the rice as a percentage; ECCENTRI CI TY represents the eccentricity of an ellipse obtained by removing the background from a top-down image of the rice plant taken at the jointing stage and performing convex hull and elliptic plotting; 0.41 is the intercept of the equation, and 26.42 is the coefficient of the elliptic eccentricity. The predicted chalkiness values ​​calculated using this equation for 91 rice varieties are shown in Table 1. Varieties with lower predicted chalkiness values ​​were selected.

[0048] 4. To test the accuracy of chalkiness prediction, the chalkiness of all 91 rice varieties was measured after harvest. The measured chalkiness values ​​were compared with the predicted chalkiness values ​​calculated using the equation "ebd=0.41+26.42*ECCENTRI CI TY" to analyze the accuracy of the equation's chalkiness prediction. Table 1 shows that for 36 of the 91 rice varieties, the difference between the predicted and measured chalkiness values ​​was less than 3%, meaning that 39.6% of the varieties had a difference of <±3%. Table 1 also shows that for 57 of the 91 rice varieties, the difference between the predicted and measured chalkiness values ​​was less than 4%, meaning that 62.6% of the varieties had a difference of <±4%. Furthermore, Table 1 shows that for 74.7% of the varieties, the difference between the predicted and measured values ​​was <±5%. This analysis shows that the equation "ebd=0.41+26.42*" can be used to...

[0049] The "ECENTRI CI TY" method demonstrates high accuracy in predicting chalkiness. For 74.7% of varieties, the predicted chalkiness difference from the measured value is <±5%; for 62.6% of varieties, the difference is <±4%; and for 39.6% of varieties, the difference is <±3%. Table 1 shows that only 12 varieties have a difference greater than ±7% between predicted and measured chalkiness, meaning only 13.2% of varieties have a significant difference. For 86.8% of varieties, the difference is <±7%. This accuracy has significant practical value for large-scale screening of low-chalkiness varieties. Using this method, combined with RGB imaging results obtained through simple operation, rapid early screening of chalkiness can be achieved at the rice jointing stage, enabling early selection of chalkiness as a breeding target.

[0050] Table 1 shows RGB imaging of 91 rice varieties at the jointing stage. The chalkiness of rice was predicted using the equation "ebd=0.41+26.42*ECCENTRICI TY". The predicted chalkiness was compared with the actual chalkiness measured after the rice matured and was harvested. In the equation, ebd represents the chalkiness of rice, expressed as a percentage. ECCENTRICI TY represents the eccentricity of the ellipse obtained after removing the background from the top-view image of the rice plant taken at the jointing stage and performing convex hull plotting and elliptic plotting. 0.41 is the intercept of the equation, and 26.42 is the coefficient of the elliptic eccentricity. The eccentricity data is the average of three replicates (n=3). The chalkiness data is the average of three replicates (each replicate measures 30 grams of rice).

[0051]

[0052]

[0053]

[0054]

[0055] In the above embodiment, the rice was sown on May 22 in Fuyang District, Hangzhou City, Zhejiang Province. 70 days after sowing, that is, on July 31, most rice varieties were in the jointing stage, which is a suitable time for RGB imaging. However, since the number of days required from sowing to entering the jointing stage is not always 70 days when rice is sown in different seasons and locations, the number of days required from sowing to entering the jointing stage is adjusted according to different seasons and planting locations, with the observation of rice entering the jointing stage at the rice planting location as the standard.

[0056] In the above embodiments, the selected RGB imaging system is the RGB imaging system of the PlatScreen™ indoor high-throughput phenotypic imaging analysis platform from PSI (Photon Systems Instruments) in the Czech Republic; however, it is not limited to the above RGB imaging system. The RGB imaging system can be an imaging system from different manufacturers and brands, or it can be obtained by taking pictures with a mobile phone camera, digital camera, etc.

[0057] It should also be noted that the correlation equation in the above embodiment, ebd = 0.41 + 26.42 * ECCENTRICI TY, is obtained by fitting the correlation between the eccentricity of the rice plant ellipse and the chalkiness of the rice after harvest, automatically calculated by the software built into the PlantScreen™ indoor high-throughput phenotypic imaging analysis platform of PSI (Photon Systems Instruments) in the Czech Republic. However, in practice, different computer software and algorithms can be used to remove the background from the RGB image in the above embodiment. Convex hull drawing is performed on the rice image after background removal, and then an ellipse is drawn based on the shape of the convex hull. The semi-focal length and semi-major axis of the ellipse are calculated, as well as the eccentricity of the ellipse. Different manufacturers or brands of computer software and algorithms can be used, or self-made computer software and algorithms can be employed.

[0058] If the computer software or algorithm is changed, the correlation between the eccentricity of the ellipse in the rice plant image and the chalkiness of the rice after harvest can be re-fitted using the following method:

[0059] The aforementioned association is formed based on historical datasets; wherein, the historical dataset is a data set using the ellipse eccentricity value calculated at the rice jointing stage and the corresponding measured chalkiness data of rice.

[0060] The correlation relationship is represented by the following equation:

[0061] ebd=A+B*ECCENTRI CI TY

[0062] Where ebd represents the chalkiness of rice, and the unit of chalkiness is percentage; ECCENTRI CI TY represents the eccentricity of the ellipse; where A is the intercept of the equation and B is the coefficient of the eccentricity of the ellipse.

[0063] By substituting historical data into the above equation, the specific values ​​of the equation intercept A and the coefficient B of the ellipse eccentricity can be calculated, ultimately forming a usable correlation equation.

[0064] In summary, this invention enables the prediction of chalkiness in harvested rice before harvest, overcoming the limitations of existing technologies that require visual inspection and processing of the rice grains for chalkiness testing. It predicts chalkiness without requiring measurement. Furthermore, it can predict the chalkiness of mature rice at an early stage of growth, specifically during the jointing stage, approximately 70 days after sowing. This early prediction allows for rapid screening of low-chalk rice materials, improving breeding efficiency and accelerating the breeding process, which is of great significance for cultivating high-quality rice.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, or alterations made by those skilled in the art using the disclosed technical content shall fall within the protection scope of the present invention.

Claims

1. A method for predicting the chalkiness of rice, characterized in that, include: Step (1): Take a photo of the rice plant from an overhead angle during the jointing stage of the rice plant; Step (2): Perform convex hull plotting on the rice plant image in the photograph, and draw an ellipse based on the shape of the convex hull. First, calculate the semi-focal length and semi-major axis of the ellipse, and then calculate the eccentricity of the ellipse using the following equation: ECCENTRICITY = Half focal length / Half major axis; Where ECCENTRICITY represents the eccentricity of the ellipse. Step (3): Based on the calculated ellipse eccentricity, predict the chalkiness of rice after harvest according to the correlation between the ellipse eccentricity in the obtained rice plant images and the chalkiness of rice after harvest. The correlation between the ellipse eccentricity in the obtained rice plant images and the chalkiness of the rice after harvest is expressed by the following equation: ebd = 0.41 + 26.42 * ECCENTRICITY; Where ebd represents the chalkiness of rice, and the unit of chalkiness is percentage; ECCENTRICITY represents the eccentricity of the ellipse; 0.41 is the intercept of the equation; 26.42 is the coefficient of the eccentricity of the ellipse.

2. The method for predicting the chalkiness of rice according to claim 1, characterized in that, In step (1), a photograph of rice plants is taken from a top-down angle using an RGB imaging system.

3. The method for predicting the chalkiness of rice according to claim 1, characterized in that, Before step (2), the process also includes: processing the photograph to remove the background and leave only the rice plant part; in step (2), convex hull drawing is performed on the rice plant image with the background removed in the photograph.

4. The method for predicting the chalkiness of rice according to claim 1, characterized in that, In step (2), the convex hull drawing is to draw the shape of a polygon obtained by drawing the points around the outer perimeter of the rice plant image surrounded by stretched rubber bands.

5. The method for predicting the chalkiness of rice according to claim 4, characterized in that, The elliptical drawing based on the convex hull shape is performed so that the boundary of the ellipse coincides exactly with the boundary of the polygon obtained from the convex hull drawing.

6. A method for screening rice with low chalkiness, characterized in that, Based on the prediction results of the method for predicting the chalkiness of rice according to any one of claims 1-5, rice with low chalkiness is screened.

Images