Rei Sonobe's research while affiliated with Shizuoka University and other places

Quantifying carotenoid content in agriculture is essential for assessing crop nutritional value, improving crop quality, promoting human health, understanding plant stress responses, and facilitating breeding and genetic improvement efforts. Hyperspectral reflectance imaging is a nondestructive and rapid tool for estimating the carotenoid content. In spectrometer reflectance measurements, there are various sources of noise that can compromise the accuracy of carotenoid content estimations. Recently, various machine learning algorithms have been identified as robust against various types of noise, eliminating the need for denoising processes. Specifically, Cubist and the one-dimensional convolutional neural network (1D-CNN) have been used in evaluating vegetation properties based on reflectance data. We used regression models based on Cubist and 1D-CNN to estimate carotenoid content from reflectance data (the spectral resolution was resampled in 5 nm bands across the entire wavelength domain from 400 to 850 nm) with various degrees of Gaussian and spike noise added. The Cubist-based model was the most robust for this purpose: it achieved a ratio of performance to deviation of 1.41, a root mean square error of 1.11 µg/cm2, and a coefficient of determination (R2) of 0.496 when applied to reflectance data with a combination of Gaussian (mean: 0; variance: 0.04) and spike noise (density: 0.05; amplitude: 0.05).

Tea is second only to water as the world’s most popular drink and it is consumed in various forms, such as black and green teas. A range of cultivars has therefore been developed in response to customer preferences. In Japan, farmers may grow several cultivars to produce different types of tea. Leaf chlorophyll content is affected by disease, nutrition, and environmental factors. It also affects the color of the dried tea leaves: a higher chlorophyll content improves their appearance. The ability to quantify chlorophyll content would therefore facilitate improved tea tree management. Here, we measured the hyperspectral reflectance of 38 cultivars using a compact spectrometer. We also compared various combinations of preprocessing techniques and 14 variable selection methods. According to the ratio of performance to deviation (RPD), detrending was effective at reducing the influence of additive interference of scattered light from particles and then regression coefficients was the best variable selection method for estimating the chlorophyll content of tea leaves, achieving an RPD of 2.60 and a root mean square error of 3.21 μg cm−2.

Leaf chlorophyll content is used as a major indicator of plant stress and growth, and hyperspectral remote sensing is frequently used to monitor the chlorophyll content. Hyperspectral reflectance has been used to evaluate vegetation properties such as pigment content, plant structure and physiological features using portable spectroradiometers. However, the prices of these devices have not yet decreased to consumer-affordable levels, which prevents widespread use. In this study, a system based on a cost-effective fingertip-sized spectrometer (Colorcompass-LF, a total price for the proposed solution was approximately 1600 USD) was evaluated for its ability to estimate the chlorophyll contents of radish and wasabi leaves and was compared with the Analytical Spectral Devices FieldSpec4. The chlorophyll contents per leaf area (cm2) of radish were generally higher than those of wasabi and ranged from 42.20 to 94.39 μg/cm2 and 11.39 to 40.40 μg/cm2 for radish and wasabi, respectively. The chlorophyll content was estimated using regression models based on a one-dimensional convolutional neural network (1D-CNN) that was generated after the original reflectance from the spectrometer measurements was de-noised. The results from an independent validation dataset confirmed the good performance of the Colorcompass-LF after spectral correction using a second-degree polynomial, and very similar estimation accuracies were obtained for the measurements from the FieldSpec4. The coefficients of determination of the regression models based on 1D-CNN were almost same (with R2 = 0.94) and the ratios of performance to deviation based on reflectance after spectral correction using a second-degree polynomial for the Colorcompass-LF and the FieldSpec4 were 4.31 and 4.33, respectively

Dimensionality reduction Zizania latifolia Hyperspectral Machine learning The amount of chlorophyll in a plant useful to indicate its physiological activity and then changes in chlorophyll content have been used as a good indicator of disease as well as nutritional and environmental stresses on plants. Chlorophyll content estimation is one of the most applications of hyperspectral remote sensing data. The aim of this study is to evaluate dimensionality reduction for estimating chlorophyll contents from hyperspectral reflectance. Random Forest (RF) has been applied to assess biochemical properties such as chlorophyll content from remote sensing data; however, an approach integrating with dimensionality reduction techniques has not been fully evaluated. A total of 200 Zizania latifolia leaves with 5 treatments from Shizuoka University field were measured for reflectance and chlorophyll content. then, the regression models were generated based on RF with three dimensionality reduction methods including principal component analysis, kernel principal component analysis and independent component analysis. This research clarified that PCA is the best method for dimensionality reduction for estimating chlorophyll content in Zizania Latifolia with a RMSE value of 5.65 ± 0.58 μg cm-2 .

Chlorophyll content has been used as an indicator for assessing photosynthetic ability, health and defense against a variety of degenerative diseases. Hyperspectral remote sensing offers some non-destructive methods and has played an important role in evaluating vegetation characteristics. However, the prices of traditional field portable spectroradiometers, such as Ocean Optics Hyperspectral Vis-NIR spectroradiometers and Analytical Spectral Devices FieldSpec series, have not yet decreased to consumer levels, which prevents much practical use. Recently, fingertip-sized spectrometers have been developed and then they could be powerful tools for evaluating vegetation characteristics. In the present study, a compact spectrometer (C12880MA-10, Hamamatsu Photonics) was used to evaluate chlorophyll content in tea leaves. Incorporating pre-processing techniques was effective for obtaining estimated values with high accuracy and then de-trending was the best pre-processing technique in this study, achieving an RPD of 2.03 and an RMSE of 3.07 μg cm－2. The proposed method is cost effective, practical for consumers to apply and will enable effective crop management.

The Normalized Difference Vegetation Index (NDVI) has been used for evaluating various vegetation properties and then it is also effective for improving classification accuracies. However, optical remote sensing imagery is limited by cloud contamination. In this study, NDVI images were simulated using the image-to-image translation methods including CycleGAN, pix2pix and pix2pixHD and then they were evaluated for classifying crop types. A significant improvement was confirmed by adding NDVI images generated by pix2pix or pix2pixHD on Sentinel-1 C-SAR VH/VV polarization data and resulted in overall accuracies of 68.0%.

The Normalized Difference Vegetation Index (NDVI) is effective for expressing vegetation status and quantified vegetation attributes. However, optical remote sensing imagery is limited by cloud contamination. On the other hand, synthetic aperture radar (SAR) can work under all weather conditions and overcome this disadvantage of optical remote sensing while it is difficult to recognize the land cover types visually due to the mechanisms of SAR imaging and the speckle noise. In this study, the image-to-image translation methods (pix2pix and CycleGAN) were used to convert Sentinel-1 C-SAR images into Sentinel-2 NDVI images. The results show that the combination of CycleGAN and VH polarization data works well during the growing season of beetroots and the simulated NDVI values were significantly correlated with the real NDVI values.

Chlorophyll content is effective indicator of photosynthesis and then it can be indicative of plant physiological activity. Nowadays, Hyperspectral remote sensing data are having been used for evaluating chlorophyll content. Especially, it has been shown that combining hyperspectral data and machine learning algorithms could be more effective to evaluate vegetation properties. The wasabi (Eutrema japonicum) plants were cultivated individually in Wagner plot (1/5000 a) containing 3L of tap water adjusted to a pH of 6.0 using HCl and NaOH. The 7 different treatments were applied including control, 0N, 2N, 0P, 2P, 0K, 2K, 0.5S, and 2S. A total of 100 wasabi leaves were sampled from plant tops among expanding leaves and reflectance was measured using Fieldspec4. The regression models were generated base on Random Forest (RF) machine learning algorithm using Original Reflectance (OR) and 5 preprocessing methods including First Derivative Reflectance (FDR), Continuum Removed (CR), Detrending (DT), Standard Normal Variate (SNV), and Multiplicative Scatter Correction (MSC) for estimating chlorophyll contents from reflectance. The objectives of this study were to examine the potential hyperspectral remote sensing approaches and to identify an optimized preprocessing technique for estimating chlorophyll content. The results indicate that all preprocessing technique were effective for improving estimation accuracies. Based on the ratio of performance of deviation (RPD), SNV was the best preprocessing technique with the RPD value of 2.48.