Weeds identification using Evolutionary Artificial Intelligence Algorithm

In a world reached a population of six billion huma ns increasingly demand it for food, feed with a wat er shortage and the decline of agricultural land and t he deterioration of the climate needs 1.5 billion h ectares of agricultural land and in case of failure to combat pests needs about 4 billion hectares. Weeds represe nt 34% of the whole pests while insects, diseases and the deterioration of agricultural land present the rema ining percentage. Weeds Identification has been one of t he most interesting classification problems for Art ificial Intelligence (AI) and image processing. The most co mmon case is to identify weeds within the field as they reduce the productivity and harm the existing crops . Success in this area results in an increased prod uctivity, profitability and at the same time decreases the co st of operation. On the other hand, when AI algorit hms combined with appropriate imagery tools may present the right solution to the weed identification prob lem. In this study, we introduce an evolutionary artific ial neural network to minimize the time of classifi cation training and minimize the error through the optimiz ation of the neuron parameters by means of a geneti c algorithm. The genetic algorithm, with its global s earch capability, finds the optimum histogram vecto rs used for network training and target testing throug h a fitness measure that reflects the result accura cy and avoids the trial-and-error process of estimating th e network inputs according to the histogram data.