Prediction of Color Level and Chlorophyll Content of Corn (Zea mays L.) Leaves by Using Mobile Phone Cameras

1st International Conference on Mechanical Electronic and Biosystem Engineering

Astika, I.W.^1 ; Khayati, N.N.^2

Department of Mechanical and Biosystem Engineering, Bogor Agricultural University, Bogor, Indonesia^1

Alumnus of Department of Mechanical and Biosystem Engineering, Bogor Agricultural University, Bogor, Indonesia^2

关键词: Chlorophyll contents;    Chlorophyll meters;    Color component;    Corn (Zea mays L.);    Correlation coefficient;    Manual measurements;    Mobile phone cameras;    National ID cards;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/557/1/012029/pdf
DOI  :  10.1088/1757-899X/557/1/012029

来源: IOP

【 摘 要 】

This study aimed at designing a method to determine color level of corn leave and to estimate chlorophyll content using mobile phone cameras. The image of the leafs are captured with hand phone camera, and then the R, G, B color components are extracted. Four levels of color levels were used as those of leaf color card issued by IRRI. As for the natural light compensation, the colors of the Indonesian national inhabitant ID card were used as color references. The leafs were placed on the ID card, captured together, and then the leaf color level is determined by the variation of the grey levels of the leaves and the color references. The chlorophyll content were measured with SPAD chlorophyll meter. The average accuracy of leaf color prediction were 81.48% for sweet corn and 76.82% for hybrid corn under various field luminance, while the accuracy were 94.45% under a fixed 1500 lux luminance. The accuracy of chlorophyll prediction was very low with and R2 = 0.4762 for sweet corn and R2 = 0.5284 with national ID card references for hybrid corn. This low results were relatively contrast with the prediction from manual measurement of leaf color level, where the correlation coefficients were R2 = 0.8466 for sweet corn and R2 = 0.8506 for hybrid corn.