Soil pH parameter Estimation Using Image Processing and Regression Analysis

Abstract

Soil pH is an important parameter which affects nutrients in soil and is helpful in deciding crop as well as fertilizer management. The classical methods are proven methods and have their own limitations like sensor and calibration of meters. The alternate solutions such as image processing based estimation give reliable and consistent results than traditional methods. In proposed paper, color images of fined soil are used to develop the correlation of color components or index derived o using color components. It has been observed that the proposed index asv pH index value gives better correlation coefficient of 0.76 with actual pH values. The regression equation obtained by curve fitting gives R-square of 0.6. The soil classification as acidic, neutral or basic is done effectively and found to be accurate. The overall test sample error is 3.99 percent with lies between the desired span.