人工智能
阈值
计算机科学
平滑的
模式识别(心理学)
噪音(视频)
阶段(地层学)
离群值
计算机视觉
地质学
图像(数学)
古生物学
作者
Bipin Nair,M U Anusha,J. Anusha
标识
DOI:10.1109/icces54183.2022.9835997
摘要
Knowledge present in rocks is extincting due to degradation. So, the proposed work aims to binarize the historical ancient Kannada rock inscription from ancient temples about old slokas as well as mythological information’s in a readable format. The historical data are very precious so preserving such data is very important task. So proposed work makes the degraded rock inscriptions to readable format using the novel stage wise processing proceeded with smoothing, sharpening, noise removal, outlier detection, filling holes, background subtraction as well as thresholding. After thresholding will get a binarized output image which is more readable. The proposed research work has addressed some of the challenges like oil stains, rock erosion, algae, fungus and uneven illumination. Proposed model gives the highest accuracy of 95% compared to existing model like Otsu, Sauvola, Niblack, Adaptive Thersholding and Adaptive Gaussian Thersholding.
科研通智能强力驱动
Strongly Powered by AbleSci AI