Abstract:Maize is widely planted in China and even in the world. Nitrogen is an essential nutrient for the growth and development of maize, which has a significant impact on maize yield. In order to provide a non-destructive and rapid detection method for nitrogen content of maize leaves, the relationship between the color characteristics and nitrogen content of maize leaves was analyzed, and a nitrogen content detection software for maize leaves was developed based on Android platform. The image containing the measured maize leaf and the calibration color block group (red, green, blue, white, black and grey) were obtained. In order to reduce the distortion caused by the external illumination and other factors, the image color was corrected by the calibration color block. After the image segmentation, image smoothing, and color feature information extraction, the relationship between the color features and the nitrogen content of the maize leaves was analyzed. The experimental results showed that the linear relationship between the green standard value and the nitrogen content was the best. Besides, Java programming language and OpenCV were applied to realize image acquisition, image processing and results viewing based on Android platform. The validation results indicated that the absolute error of the method for the nitrogen content of maize leaves was between -0.40% and 0.35%, and the root mean square error was 0.20%. The time from image collection to giving results was less than 10s. The proposed nitrogen detection method had the advantages of rapidity, economy and portability, and can be used for real-time detection on nitrogen content of maize leaves.
