FIELD STRESS DETECTION USING REMOTE SENSING IN AGRICULTURE

Abstract

One of the global challenges facing the world today is the question of how to produce enough food for the population which grows year to year. Considering the limitations of arable land, the solution represents the improvement of the process of production and growing plants. The expansion of “precision agriculture”, whose basic element is remote sensing, as a method of collecting information from the ground without physical contact with it, has enabled a significant improvement in the process of plants growing in the field of crop monitoring. By analyzing digital images, certain information is collected from the sensor that contains a part of the electromagnetic spectrum. By studying various changes in the value of the electromagnetic spectrum, it is possible to make certain conclusions about the condition of the crops on the arable surface. This paper presents an algorithm for the detection of crop stress in the field, based on the supervised classification that requires the selection of segments of the picture that represent a healthy plant. The algorithm is based on the extraction of the vegetation index values (VI’s) from the selected segments of the digital image, and then the creation of appropriate vegetation index histograms of a healthy plant. The
resulting histograms are then compared with the histograms of all other segments of the image, where differences in the values of vegetation index manifest themselves as stress in the field.