Edge detection is a key process in computer vision that identifies significant variations in image intensity, which correspond to physical phenomena. This process is crucial for applications such as 3D reconstruction, motion analysis, image enhancement, and registration. Edge detection typically involves smoothing and differentiation of the image, but these operations can introduce noise and information loss. Therefore, various edge detectors have been developed, each with different mathematical and algorithmic properties. This paper provides an overview of current understanding of edge detection, including edge definition, detector properties, detection methodology, and existing detectors. Edge detection aims to locate variations in image intensity and identify the physical phenomena causing them. These variations include step edges, line edges, and junctions. Step edges are the most common type and are localized at inflection points of the image. Line edges result from mutual illumination or thin objects against a background. Junctions occur where at least two edges meet and are defined by the meeting of edges in different directions. Edge detectors process images to identify these edges, involving differentiation, smoothing, and labeling. Differentiation is the computation of image derivatives to locate edges, while smoothing reduces noise and regularizes differentiation. Labeling involves suppressing false edges to enhance the signal-to-noise ratio. The performance of these operations is interdependent, with smoothing and differentiation being commutative in some cases. Smoothing filters, such as Gaussian and cubic spline, are used to reduce noise while preserving image structure. Differentiation operators, including gradient and Laplacian, are used to detect edges. The gradient is a first-order operator, while the Laplacian is a second-order operator. The choice of differentiation operator affects the type of edges detected. Edge labeling involves suppressing false edges to enhance the signal-to-noise ratio. This process is influenced by the properties of the differentiation operator and the characteristics of the image. Multi-detector and multi-scale approaches are used to improve edge detection by considering different scales and detectors. These approaches help in identifying edges more accurately by combining information from different scales and detectors. The mutual influence between edge detectors and edges is significant, as the properties of the detector affect the detected edges. The performance of edge detectors is evaluated based on criteria such as signal-to-noise ratio and delocalization error. Differentiation operators, such as gradient and Laplacian, have different effects on edge detection, with the Laplacian being sensitive to edge curvature and illumination conditions. The performance of edge detectors is also influenced by the characteristics of the image, such as subpixel and edge orientation. Overall, edge detection is a complex process that requires careful consideration of various factors to ensure accurate and reliable results.Edge detection is a key process in computer vision that identifies significant variations in image intensity, which correspond to physical phenomena. This process is crucial for applications such as 3D reconstruction, motion analysis, image enhancement, and registration. Edge detection typically involves smoothing and differentiation of the image, but these operations can introduce noise and information loss. Therefore, various edge detectors have been developed, each with different mathematical and algorithmic properties. This paper provides an overview of current understanding of edge detection, including edge definition, detector properties, detection methodology, and existing detectors. Edge detection aims to locate variations in image intensity and identify the physical phenomena causing them. These variations include step edges, line edges, and junctions. Step edges are the most common type and are localized at inflection points of the image. Line edges result from mutual illumination or thin objects against a background. Junctions occur where at least two edges meet and are defined by the meeting of edges in different directions. Edge detectors process images to identify these edges, involving differentiation, smoothing, and labeling. Differentiation is the computation of image derivatives to locate edges, while smoothing reduces noise and regularizes differentiation. Labeling involves suppressing false edges to enhance the signal-to-noise ratio. The performance of these operations is interdependent, with smoothing and differentiation being commutative in some cases. Smoothing filters, such as Gaussian and cubic spline, are used to reduce noise while preserving image structure. Differentiation operators, including gradient and Laplacian, are used to detect edges. The gradient is a first-order operator, while the Laplacian is a second-order operator. The choice of differentiation operator affects the type of edges detected. Edge labeling involves suppressing false edges to enhance the signal-to-noise ratio. This process is influenced by the properties of the differentiation operator and the characteristics of the image. Multi-detector and multi-scale approaches are used to improve edge detection by considering different scales and detectors. These approaches help in identifying edges more accurately by combining information from different scales and detectors. The mutual influence between edge detectors and edges is significant, as the properties of the detector affect the detected edges. The performance of edge detectors is evaluated based on criteria such as signal-to-noise ratio and delocalization error. Differentiation operators, such as gradient and Laplacian, have different effects on edge detection, with the Laplacian being sensitive to edge curvature and illumination conditions. The performance of edge detectors is also influenced by the characteristics of the image, such as subpixel and edge orientation. Overall, edge detection is a complex process that requires careful consideration of various factors to ensure accurate and reliable results.