Rice Leaf Image Contrast Enhancement through Joint Occurrence of Spatial Gray Levels

Abstract

Rice leaf images are the main attributes in the diagnosis of several rice related diseases. As their acquirement in real time imposes several artifacts, they needs to be pre-processed before subjecting them for further processing. Towards such objective, this paper proposes a new method called as Spatial Relation Assisted Contrast Enhancement (SRCE). SRCE is a simple and effective method that considers the Joint Spatial Spread (JSS) to perform contrast enhancement. For every gray level, its JSS is measured through 2D Spatial Joint Histogram (2DSJH) and Mutual Information. Based on the mutual information, SRCE constructs a hyperlink matrix and assigns a rank which denotes the close occurrence of gray levels. Further, the rank is used for mapping input gray levels to output gray levels.  Simulation Experiments on different types of rice leaf mages through qualitative and quantitative evaluation shows the effectiveness of SRCE in improving the quality. For performance assessment, different metrics including Contrast Improvement Index (CII) and Structural Similarity Index Measure (SSIM) are used and compared with different state-of-the art methods. On an average, the CII of proposed method is observed as 6.6733 while for conventional methods, it is observed as 2.2978, 3.1767, 3.7322, 3.9166 and 5.1385 for HE, CLAHE, CLAHE + HF, BPDFHE and SECE respectively. Further the average PSNR is observed as 16.9312 dB while for HE, CLAHE, CLAHE + HF, BPDFHE and SECE, it is observed as 9.4264 dB, 12.6688 dB, 13.4062 dB, 14.4186 dB and 15.5586 dB respectively.