Weed Detection Using AlexNet Architecture In The Farming Fields

Authors:

Abstract
Farming is one of the beginnings of human food in this world. These days because of rapid growth in population we need the more prominent useful capacity of farming to satisfy the needs. There are a few problems like weeds which are keeping us from getting good yield. For this, humans are opted to depend upon destructive toxins such as herbicides and weedicides. Because of using these harmful toxins which are sprayed in the entire field are affecting not only weeds but also the crops, which will bring an uncertainty up in our food on this lovely earth. In this paper, carried out certain strategies to detect the presence of weed in a certain area of the field. In the previous methods weeds are recognized from the images, making utilization of image handling strategy by shape, shading and size included utilizing Support Vector Machine(SVM) or Artificial Neural Network(ANN) by using these methods established a model to classify weeds and crops. These established methods do not have efficient feature extraction and classification method. In this paper AlexNet architecture in convolutional neural networks is utilized for better results. The weedcrop, deepweed, plantseedlings-v2 data sets are used to work on the model. The pictures are resized accordingly to give them as input to the model. The proposed method shows an average accuracy of 96%.

This publication has 2 references indexed in Scilit: