Intersection rotational invariance regression based plant disease detection

Emerged Technologies include imparted potentiality to generate adequate food to converge ultimatum of society. However, aspects like, climate change, plant disease, refuse at pollinators as well as others are demanding to farmers. The presence of diseases pulls up the maturation of their corresponding species. Plant disease recognition using image processing and machine learning (ML) with data collected from Internet of Thing (IoT) sensors has been receiving greater amount of interest in recent years, however timely disease detection remains a demanding issues. Despite early disease detection though reduces the risk of destruction to plants nevertheless the noise present in the sensor networks (i.e. Internet of Things) while collecting internet scrapped images compromise overall precision and accuracy. To address on this research gap, i.e., addressing noise with timely disease detection, in this work ML based plant disease detection method called, Intersection Histogram and Rotational Invariant Principal Component Regression (IH-RIPCR) is introduced. IH-RIPCR technique is dividing as pre-processing and feature extraction. First with the raw plant dataset obtained as input from Corn or Maize Leaf Disease Dataset, pre-processing is done employing Intersection Histogram based Contrast Enhancement model. Second with the obtained pre-processed contrast enhanced images is subjected as input to ML-based Rotational Invariance Principal Component Regression feature extraction model to extract relevant features pertaining to corn or maize leaf images for disease detection in an accurate and precise manner. Experiments are conducted with corn or maize plant leaf dataset with different existing methods to verify hypothesis potentiality of IH-RIPCR technique. Significance of technique is evaluated based on numerous parameters in terms of PSNR, processing time, precision, recall and accuracy respectively.