The maize leaf is normally affected by Cerpospora, Common Rust, Northern Leaf blight diseases. Shadows in maize leaf disturb the disease detection. This Study aims at removing shadows in disease affected maize plant. The mean shift segmentation is used to detect shadows by applying Y channel information in YcbCrcolor space and HSI color space information. The detected Shadows are removed by processing in HSV color space and area is restored by histogram matching Techniques. No training model is needed to be applied in this study. After the shadow detection has been reached, without training can directly remove the shadows. This helps in misclassification of shadows in cerpospora diseased maize leaf as region of interests. The Accuracy, misclassification and Precision values before shadow removal and after shadow removal are 0.2, 0.8, 0.2 and 0.85, 0.15, 1 respectively.