The study described in this paper consists of amethod that applies gabor wavelet transform technique to extractrelevant features related to image of tomato leaf in conjunctionwith using Support Vector Machines (SVMs) with alternatekernel functions in order to detect and identify type of diseasethat infects tomato plant. Initially, we collected real samples ofdiseased tomato leaves, next we isolated each leaf in single image,wavelet based feature technique has been employed to identify anoptimal feature subset. Finally, a support vector machine classierwith different kernel functions including Cauchy kernel, InvmultKernel and Laplacian Kernel was employed to evaluate the abilityof this approach to detect and identify where tomato leaf infectedwith Powdery mildew or early blight.To evaluate the performance of presented approach, we presenttests on dataset consisted of 100 images for each type of tomatodiseases. Extensive experimental results demonstrate that theproposed approach provides excellent annotation with accuracy99.5 %. Efficient result obtained from the proposed approachcan lead to tighter connection between agriculture specialists andcomputer system, yielding more effective and reliable results.Index Terms—image processing, K-Mean clustering algorithm,Geometric features, support vector machine (SVM
Tomato leaves diseases detection approach based on Support Vector Machines. Available from: https://www.researchgate.net/publication/296706272_Tomato_leaves_diseases_detection_approach_based_on_Support_Vector_Machines [accessed Jun 2, 2016]. |
