A Novel Plant Leaf Ailment Recognition Method using Image Processing Algorithms
In the 21st Century, agriculture still remains the major source of food for human beings and it has far shadowed other sources such as hunting, fishing and gathering. Since environmental conditions are beyond the scope of human control, plant illness identification is acting as a critical position in the agricultural field. This paper suggests a method to replace the traditional methods of identifying disease through the use of “image-processing” techniques. In this study, an image of the leaf of a diseased plant has been taken using a digital camera. Three segmentation algorithms namely Green Pixel Masking, “CIE L*a*b colour space” extraction and H element of HSV extraction have been used to split the image into diseased and healthy regions. The diseased region is then used to calculate 13 parameters which are utilized as inputs by a pre-trained neural network which utilizes “feed-forward back propagation algorithm” to determine the final output. The proposed methodology has achieved a maximum accuracy of 95.62% for Apple leaves, 91.62% for Grape leaves and 91.1% for Tomato leaves.
Artificial Neural Network, Classification, Image Processing, Plant disease detection
Full Text: PDF (downloaded 202 times)
- There are currently no refbacks.