Multiple noise reduction using hybrid method for leaf recognition

Plant recognition is very demanding in biology and agriculture as new plant discovery and the computerization of the management of plant species become more popular. Developing an automatic plant recognition system becomes a very challenging task. Image pre-processing techniques, such as filtering for noise reduction and enhancement of the image is highly essential. The computerized plant recognition mainly suppresses noise in the input image leading to stable feature extraction of plants. By applying Gaussian noise, the leaf image appears blurred, due to which the edges of the leaf vein is not clearly visible. Speckle noise occurrence is often undesirable, and also very dangerously causes damage to the leaf shape and structure. So, the ultimate goal is to develop a system where both multiple Gaussian and speckle noise can be removed and restored so that the image becomes noise free and produces clear vein and shape of the leaf which is highly essential for further process. Hybrid filter method is developed to remove the noise, enhance the quality of image and thereby produces better results compared to other traditional filters. Different parametric metrics are used to evaluate the performance of the hybrid filter and proves by giving suitable results when compared to other traditional filters.