1 Experimental background

Chlorophyll concentration is often a good indicator parameter for plant nutritional stress, photosynthetic capacity, and various stages of development and aging. The detection of chlorophyll concentration using ground spectral and remote sensing technologies often receives great attention in research directions such as agriculture and natural plant communities. Most studies focus on the relationship between spectral reflectance and chlorophyll a or total chlorophyll concentration, while others investigate chlorophyll b or the ratio of chlorophyll a to chlorophyll b concentration. Other studies mainly focus on the concentration of carotenoids or the ratio of carotenoids to chlorophyll.

Many studies have shown that the "red edge" of vegetation spectra is sensitive to the growth status of vegetation. The inflection point of the vegetation red edge is the point with a large slope in the plant reflectance spectrum curve, located between wavelengths of 690-720nm. Laboratory experiments have confirmed a positive correlation between the wavelength position of the red edge inflection point and the chlorophyll concentration of leaf samples.

This article further investigates the correlation between rice leaf reflectance and chlorophyll concentration, as well as the relationship between the spectral red edge wavelength position of rice leaves and their chlorophyll concentration, by measuring and calculating the reflectance of rice leaves with different chlorophyll concentrations and the red edge of the first-order differential curve of reflectance.

Basic working principle of ground spectrometers

In addition, ground spectrometers have a wide range of applications, including mineral identification and resource estimation in mineral exploration, as well as surface classification and disaster monitoring in remote sensing applications. For example, in mineral exploration, different minerals have unique spectral characteristics, and spectral analysis can identify mineral types and distribution. In remote sensing applications, surface classification is carried out through spectrometers carried by satellites or drones to classify and monitor the surface, such as forests, farmland, cities, etc. Disaster monitoring quickly assesses the scope and severity of natural disasters through spectral analysis.

3 Laboratory Design

preliminary preparation

In this experiment, the hybrid stem rice variety Changyou 1 (Changyou 99-1) was selected as the research object. Fertilization treatments were carried out at different points in 9 regions where the rice was planted, and there were significant differences in its growth quality.

Ten representative functional leaves were selected from nine different fertilization treatment areas of Changyou rice, and the chlorophyll concentration of each leaf was measured using a chlorophyll meter. The chlorophyll concentration of each leaf was measured four times at different leaf positions, and then the leaves were removed. The 10 leaves from each area were grouped together and measured for spectral reflectance under sunlight using an ATP9100F handheld ground spectrometer. Ten blades were stacked in different orders, and spectral measurements were quickly taken on each blade using a 25 ° field of view probe at a height of approximately 5cm from the blade. The spectral measurement range was 300-1100nm, and the spectral resolution was within 2nm.

Data analysis and processing

Between 400~500nm and 600~700nm, due to the light radiation in the blue and red light bands, chlorophyll is fully absorbed for photosynthesis, forming two bottom reflection zones with low reflectance ratios, generally below 10%; The absorption of leaves decreases from 500nm onwards, generally low at 550nm, which is the green band where a small reflection peak is formed, with a reflectivity of about 20%. This is also the reason why plant leaves appear green; Then the reflectivity suddenly increases at 700nm, and after entering the near-infrared band, the curve becomes flat, becoming a high reflection plateau area, generally reaching 50%~60%, and some even higher. The transition region where the reflectance changes from the low point of the red absorption zone of chlorophyll to the high reflectance due to leaf scattering in the near-infrared is called the "red edge" zone, generally located between wavelengths of 690-720nm. The "red edge" is important in the first-order differential curve of reflectance, and some researchers have confirmed the relationship between leaf red edge and chlorophyll concentration; And the "red edge" is the point in the plant reflectance spectrum curve with a large slope in the "red edge" area, that is, the point where the reflectance changes quickly.

Single band correlation analysis

Perform single correlation analysis on the spectral reflectance at different wavelengths and chlorophyll concentration of rice leaves (9 samples) with different levels of nitrogen application. Figure 5 shows the correlation coefficient between the two, and it is evident from Figure 5 that:

① From 420nm to 720nm, there is a negative correlation between the two. As the wavelength increases, the negative correlation coefficient sharply increases. At 460nm, the correlation coefficient reaches -0.6, at 560nm, the correlation coefficient reaches -0.8, and then at 686nm, the correlation coefficient reaches as high as -0.95;

② From 686nm to 726nm, the maximum negative correlation between the two decreases to zero, and then continues to increase to the positive correlation at 745nm, with a correlation coefficient of approximately 0.92;

③ From 726nm to 1000nm, except for a slight decrease in correlation coefficient near 940nm, the correlation coefficient between the two has maintained a good positive correlation, with a correlation coefficient of about 0.9.

The above changes in the correlation between the two are mainly due to the significant influence of chlorophyll on plant reflectance spectra in the visible light range. During the growth process of rice leaves, as the chlorophyll content increases, the number of intercellular spaces in the leaf flesh increases, resulting in a decrease in visible light reflectance and an increase in near-infrared reflectance, causing a strong negative correlation between the two in the visible light range and a strong positive correlation in the near-infrared band. Between 686nm and 745nm, the correlation between the two changes from a large negative value to a large positive value, and the correlation coefficient is close to zero at 726nm. This is mainly because the chlorophyll in rice leaves has a significant impact on visible spectral reflectance, while the structure of rice leaves has a significant impact on near-infrared spectra. The correlation between the two is very small at the junction.

The wavelength position of the red edge inflection point on rice leaves

Related to its chlorophyll concentration

Due to the fact that the ATP9100F handheld ground spectrometer collects discrete data, the spectral data can be approximately calculated for the first derivative using formula (1).

Among them: R (λ)_i）’For wavelength i (λ)_i）The first-order differential reflectance spectrum at, λ_i+1, λ_iλ_i-1 is the adjacent wavelength.

Using formula (1), take the first derivative of the reflectance spectra of rice leaves with selected non ventilated chlorophyll concentrations in 9 regions, and obtain the wavelength position of the red edge inflection point

Perform linear regression analysis with chlorophyll concentration as the independent variable x and the wavelength position of the red edge inflection point of rice leaves as the dependent variable y

There is a good correlation between the red edge inflection point of rice leaf reflection and chlorophyll concentration, with a complex correlation coefficient (R²）Reaching 0.88 is consistent with the theory that the wavelength position of the red edge inflection point and the chlorophyll concentration of leaf samples are positively correlated, which has been confirmed by laboratory experiments. Therefore, it can be seen that it is feasible to quantitatively estimate and analyze the chlorophyll concentration of rice leaves using field spectrometer data.

4 Conclusion

Based on the analysis of the reflectance spectral characteristics of rice leaves, this experiment studied and analyzed the correlation between their spectral reflectance and the position of the red edge of reflectance with leaf chlorophyll concentration. The research results indicate that:

(1) The correlation between rice leaf reflectance and chlorophyll concentration is closely related to the wavelength spectrum, showing a negative correlation between 420-726nm and a positive correlation between 726-1000nm; A high correlation is maintained between 450-686nm and 750-770nm, and a high correlation is reached at 686nm with a correlation coefficient greater than 0.95, while the correlation is 0 at 726nm. Therefore, the results of this experiment demonstrate the detectability of rice chlorophyll concentration using ground hyperspectral technology and the necessity of selecting the optimal wavelength band.

(2) There is a good linear correlation between the chlorophyll content of rice leaves and the position of the red edge inflection point in their reflectance spectra, providing a technical method and approach for quantitatively estimating the chlorophyll concentration of rice leaves using hyperspectral imaging.