Banana_Leaf_Disease_Dataset_Bangladesh

The Banana leaf disease dataset dataset comprises 424 images of banana leaves collected in Noakhali, Rangamati and Bandarban,Chittagong, Bangladesh. The dataset’s primary directory is named "Banana_Leaf_Disease_Dataset_Bangladesh" across four classes Cordana, healthy (fresh), sigatoka, and pestalotiopsis. The raw images were initially captured in different formats (JPG, PNG, JPEG) and varied resolutions but were resized to a uniform 320x427 pixels at 72 dpi for model compatibility. The dataset was designed to support deep learning-based classification and detection models. Images were captured using a Xiaomi M2006C3MG mobile camera under natural lighting conditions to replicate real-world agricultural settings. The photographs vary in format, including JPG, PNG, and JPEG, with original dimensions of 320x500, 750x1000, and 2448x3264 pixels at a horizontal and vertical resolution of 96 dpi. To ensure consistency and model compatibility, all images were resized to 320x427 pixels at 72 dpi. Each image was checked for clarity, and samples were selected to represent variations in disease severity.