Soil moisture methods for home agricultural applications

Technology is manifested in all areas of our lives. Technology makes human life easier and more comfortable. Along with technological developments, there is also great progress in areas such as communication, transportation, health care and engineering. Every field production is accelerated, the time losses are getting smaller and thanks to this savings in various aspects are provided. Information Technology is widely used in different sectors. These include sectors where agricultural activities are carried out. It increases the use of technology and IT productivity in agricultural activities. Therefore, a study area was selected to contribute to the tenological developments in this area. As urbanization increases in cities, green areas decrease. Besides this, with the increase of the population, the food need is increasing. The need for organic food and green space has led people to new solutions. Household agriculture and home garden applications are becoming widespread. It is important to carry out household agriculture and home gardens activities in a conservative way. Domestic agriculture practices benefit from technology to increase yields. Increasingly smart household systems can include agricultural activities. In this study, which can be used in household agriculture applications, observations were made on new methods by examining many studies made on the estimation of the distribution of soil noodle on the surface. The data used in the study were taken instantaneously from the hardware circuit consisting of 1 Arduino and 6 sensors which can be used for home farming, and it is recorded on the computer. In order to receive the data via the ground, 6 sensor points were set in the area and equally spaced sensors were installed. The placement locations were determined by calculating the domain of the sensors. Interpolation studies were performed on the received data to estimate the values at the points outside the points where the sensors are located. It started with the Inverse Distance Weighted (IDW) method, which is commonly used for interpolation studies. One of the sensors was determined as a prediction point and tried to estimate the value of the sensor at the predicted point by the values obtained from the other sensor points. Errors were measured by comparing the results with the actual data. There are different interpolation techniques. These methods are divided into various sub-methods according to the region where the interpolation is to be performed and the purpose of the interpolation. There are three main approaches to solve interpolation problems. These approaches are interpolating with point interpolation, interpolation with a single function covering the entire region, and locally defined piecewise functions. Inverse Distance Weighted Method is often used in interpolation operations. It is a local method for producing predictions based on neighboring spots. Estimation of unknown points is calculated by weighting according to the distance between points. Interpolation with polynomial surfaces is a method used when interpolation is required for a surface. Linear, quadratic, cubic surfaces are obtained in interpolation processes with orthogonal polynomials. Unknown binary bi-liners, bi-quadratic, bi-cubic surfaces are obtained in interpolation processes in non-nonspecific polynomials and in processes performed for Bi-linear, Bi-quadratic and Bi-cubic surfaces. If interpolation according to least squares method, many areas such as prediction, spectral analysis, functional analysis are used. Interpolation processes with Kriging methods are commonly used interpolation methods in geographical statistics. Geology, environment, mining, meteorology, such as the work done in these areas are available. Simple Kriging, Ordinary Kriging (OK), Universal Kriging, Block Kriging, Indicator Kriging, Disjunctive Kriging and Cokriging are used in different areas. In this work, the cubic interpolation applied to the data gives a unique set of cubic polynomials between the data points. Thus, smooth curves can be generated from the array points. The coefficient in cubic polynomials is the weight used for interpolation. Linear Interpolation, Spline Interpolation and Cubic Interpolation methods, which are different interpolation methods, have been applied to the data separately. According to the results, the moisture distribution on the surface was seen in Matlab. The measurement results for different time intervals are compared. Studies with the same data in different interpolation methods yielded similar moisture distribution results. These distributions are easily seen on the surfaces created in Matlab. The blue colored parts on the surface of the moisture show that the soil is damp and the yellow colored parts indicate that the soil is dry. Inverse Distance Weighting two algorithms have been proposed to improve the IDW method and to increase the accuracy of estimates. In the first algorithm, it is aimed to create the coverage area for the prediction points and calculate the estimation of the desired region so that the distant points are deducted from the estimation effect. The predicted points are estimated by the user at the specified distance interval. In the second algorithm, the sensors are detected suddenly and the effects are reduced so as not to reduce the prediction success of the sudden change in the values. As a result, the results obtained from the IDW algorithms are evaluated and the error rates are compared. In calculating error rates, the Mean Absolute Percentage Error MAPE formula is used. As a result of these comparisons, it was observed that the program performed a successful measurement based on the given range. Furthermore, it has been observed that the fact that suddenly changing sensors are detected and not included in the calculation, increases the accuracy at the predicted points. In future studies, the proposed algorithm can be improved. Soil moisture depends on many other influences such as ambient temperature, ambient humidity, and soils in the soil. For this reason, to increase the success of the measurements, these factors can also be included in the algorithm by being measured and controlled. The generated electronic circuitry and software can be integrated to make it smaller. In addition to the generated codes, this work can be added to the intelligent home system by becoming a manageable system from smart devices.