@article{article_780103, title={Adaptive 2-D LMS Filter Embedded Edge Detection Application}, journal={Avrupa Bilim ve Teknoloji Dergisi}, pages={343–351}, year={2020}, DOI={10.31590/ejosat.780103}, author={Paralı, Ufuk}, keywords={2-D Adaptif Fitre, TDLMS, Kenar Algılama Algoritması, Canny, Sobel, Prewitt, Roberts, LoG}, abstract={In this numerical study the effect of embedding two-dimensional least mean square (TDLMS) adaptive filter into various edge detection systems is discussed. TDLMS and edge detection modules are arranged in the system scheme in a manner such that they work sequentially. TDLMS algorithm is commonly used in many various image processing applications. Due to its ability of updating filter coefficients without needing any a priori assumptions, TDLMS provides superior advantegeous in 2-D signal processing applications. We investigated the performance increment of TDLMS especially on the commonly used edge detection algortihms in the literature such as Canny, Sobel, Prewitt, Roberts and LoG (Laplacian of Gaussian). It is observed that embedding TDLMS is particularly useful in edge detection for low SNR images comparing to high SNR images. The simulation results clearly show TDLMS filter provides significant improvement for the edge detection implementation on a relatively lower SNR case comparing to a higher SNR case. Especially, TDLMS embedded Sobel, Prewitt and Roberts implementations have relatively better results than TDLMS embedded Canny and LoG implementations for a low SNR image. On the other hand, for relatively higher SNR case, embedding TDLMS filter into the edge detection system does not provide as much significant improvement as in relatively lower SNR case. But still, for a high SNR case, TDLMS embedded Canny implementation have relatively better results than TDLMS embedded Sobel, Prewitt, Roberts and LoG implementations.}, publisher={Osman SAĞDIÇ}