A novel method for the estimation of soybean chlorophyll content using a smartphone and image analysis

Abstract

The development of smartphones, specifically their cameras, and imaging technologies has enabled their use as sensors/measurement tools. Here we aimed to evaluate the applicability of a fast and noninvasive method for the estimation of total chlorophyll (Chl), Chl a, Chl b, and carotenoids (Car) content of soybean plants using a smartphone camera. Single leaf disc images were obtained using a smartphone camera. Subsequently, for the same leaf discs, a Chl meter was used to obtain the relative index of Chl and the photosynthetic pigments were then determined using a classic method. The RGB, HSB and CIELab color models were extracted from the smartphone images and correlated to Chl values obtained using a Chl meter and by a standard laboratory protocol. The smartphone camera was sensitive enough to capture successfully a broad range of Chl and Car contents seen in soybean leaves. Although there was a variation between color models, some of the proposed regressions (e.g., the S and b index from HSB and Lab color models and NRI [RGB model]) were very close to the Chl meter values. Based on our findings, smartphones can be used for rapid and accurate estimation of soybean and Car contents in soybean leaves.

Mathematical models for the estimation of leaf chlorophyll content based on RGB colours of contact imaging with smartphones: A pomegranate example

“…The coefficient of determination (R 2 ) for the same colour indices in current work (when red LED light passed through leaf) was found to be as 0.3047, 0.8355 and 0.797. The coefficient of determination (R 2 ) of the present study for the linear regression among the R, G and B values with leaf Chl content are also higher than the results of Dey et al (2016), Rigon et al (2016) and Vesali et al (2017). When comparing the Chl a and Chl b in the present study, the estimation of the Chl a was higher than the Chl b.…”

“…This is why the laboratory extraction techniques use 663 and 645 nm wavelengths at the spectrophotometer to identify the Chl contents (Sudhakar et al, 2016). According to the authors' knowledge, many studies (Karcher and Richardson, 2003;Graeff et al, 2008;Li et al, 2010;Rorie et al, 2011;Lee and Lee, 2013;Tewari et al, 2013;Wang et al, 2013;Saberioon et al, 2014;Dey et al, 2016;Rigon et al, 2016) were conducted before for the estimation of the leaf Chl contents by using image scanning techniques, but none of them tested red LED light sources. Therefore, the results of the present study provided a better correlation among the RGB values and leaf Chl content (mg • MW −1 ) than the previous studies.…”

“…The results also showed that the B value also has a very strong positive correlation with the leaf Chl content (when red LED passed). Furthermore, it was observed that in some cases, the determination of Chl b was inaccurate (Rigon et al, 2016). This resulted from the fact that the absorption peak of Chl a was similar to the wavelength emitted by the device, making it more difficult to separate (Neves et al 2005).…”

“…Moreover, Graeff et al (2008) reported that the CIE L*a*b values have a correlation with the nitrogen content of broccoli plants. Low-cost digital cameras were previously reported to be useful for the estimation of the leaf chlorophyll and nitrogen content of field crops, that is, rice (Li et al, 2010;Saberioon et al, 2014;Rigon et al, 2016). To determine the leaf chlorophyll contents of any plants with the use of any of the above-mentioned independent indices or similar, mathematical models are required.…”

Desenvolvimento de software de processamento digital de imagens para análise do >i<status>/i< nutricional de nitrogênio em >i<Brachiaria brizantha>/i< cv. Xaraés utilizando rede neural artificial

“…variação dentre as três classes definidas pode ser observada na Figura 17 com um exemplo de folhas em cada classe. Uma variação semelhante no padrão dos componentes RGB de acordo com o teor de N contido nas folhas neste experimento foram descritos para as clorofilas 'a' e 'b' e os carotenóides porRigon et al (2016) em folhas de soja (Glycine max L.) e porHu et al(2013) em arroz (Oryza sativa L.). Exemplo de folhas após classificação do status e respectivos valores RGB após processamento.…”