Paddy Leaf Disease Detection and Recognition using Image Processing

Rice crops have been recognized as one of the most powerful energy sources for resource production over the last few decades. Rice plant diseases are regarded as a major cause of crop failure, economic and communal loss in the agricultural field's future development. Since the last ten years, researchers have been keenly interested in the diagnosis of plant disease approaches to image processing techniques. The primary goal of this research is to create an image processing system capable of identifying and classifying bacterial blight disease. A set of infected rice plant images from a rice field are captured using a digital camera and empirically evaluated using background removal and segmentation techniques. Image segmentation, feature extraction, feature selection, and classification are used to compare them. This paper also makes recommendations for future research on the diagnosis of bacterial blight disease.

Genetic Divergence of Rice Genotypes Revealed by Bacterial Blight Disease and Morphological Traits

Bacterial blight is a perilous impediment for rice production. Resistant variety is a sustainable approach to fend off the loss of rice due to bacterial blight disease. In this study, 94 genotypes were screened against bacterial blight disease and its morphological diversity was assessed to find out the resistant donor with desirable morphological characters. Bacterial blight pathogen was inoculated following leaf clipping method for disease scoring. Out of 94 genotypes, 12 showed a resistant reaction, 13 showed moderately resistant reaction and 69 genotypes showed a susceptible reaction. Positive correlation was recorded between yield and most of the morphological characters. Yield hill─1 was significantly correlated with the number of tiller hill─1 (0.503**), number of effective tiller hill─1 (0.538**), total number of spikelets panicle─1 (0.595**), number of filled grain panicle─1 (0.595**), number of unfilled spikelet panicle─1 (0.239*) and 1000 grain weight hill─1 (0.843**). Eleven quantitative characters grouped 94 rice genotypes in 16 clusters at coefficient 3.38 and it indicated the presence of great amount diversity among the genotypes. Principal component analysis (PCA) supported the cluster analysis and the first four principal components explained around 70.99% of total divergence for all morphological characters. Principal coordinate analysis (PCoA) demonstrated that the genotypes BR8862-29-1-5-1-3, SVIN301, SVIN321, BR9207-45-2-2, SVIN018, lRBB5, SVIN038, BRRI dhan28 and BRRI dhan29 were placed in distant position from the centroid and it indicated that they were more diverse than the genotypes near the centroid. However, based on disease reaction and genetic diversity analysis crossing could be made between, resistant genotypes such as SVIN317, SVIN017, SVIN316, SVIN313, SVIN315, SVIN314, SVIN038, SVIN307, SVIN302, SVIN304 with the susceptible variety more specifically with BRRI dhan28, BRRI dhan29, BRRI dhan50, BRRI dhan58, BRRI dhan63, BRRI dhan74, BRRI dhan81 and BRRI dhan84 to develop bacterial blight resistant variety. Bangladesh Rice J. 24 (1): 73-84, 2020