<p>Efficient and accurate segmentation of leaf images and image enhancement is crucial in various domains including nutrient deficiency identification, disease detection, plant species classification, environmental monitoring and agricultural research. Traditional segmentation techniques often struggle with challenges such as complex backgrounds, variable leaf shapes and inconsistent lighting conditions. In this work, we propose a new approach that integrates the DeepLabV3, a sophisticated deep learning model for semantic segmentation with AO-GrabCut, a powerful graph-cut-based method for accurate and precise segmentation of turmeric leaf images known as DeepLabV3 with AO-GrabCut (DLV3AO-GrabCut). Also, various image enhancement methods such as histogram equalization, logarithmic transformation and digital unsharp masking are explored and applied on turmeric dataset. DLV3AO-GrabCut demonstrate its superior performance by considering intersection over union (IoU), dice similarity coefficient (DSC), pixel accuracy (PA), precision(P), and recall(R) as 0.86, 0.94, 0.96, 0.97 and 0.98 respectively compared to traditional methods and individual segmentation models. The performance of the image enhancement methods is evaluated using peak signal-to-noise ratio (PSNR) and mean squared error (MSE), and structured similarity index matrix (SSIM). The results demonstrated that digital unsharp masking yields superior enhancement quality making it suitable for disease identification and nutrient deficiency identification. The improvement in accuracy is shown for segmented and enhanced images on a dataset of leaf images in the analysis.</p>