What are some challenges and limitations of Canny edge detection in real-world scenarios?

由人工智能和领英社区提供技术支持

Canny edge detection is a popular technique for finding the boundaries of objects in an image. It uses a series of filters and thresholds to produce a binary image where the edges are marked as white and the rest as black. However, Canny edge detection is not perfect and has some challenges and limitations in real-world scenarios. In this article, you will learn about some of these issues and how to overcome them.

此文章中的业界达人

由社区从 2 条内容中精选。了解更多

Dylan Brown

Graduate Electronic Engineer. BEng Honours in Intelligent Systems. Masters postgraduate student at the University of…

1 Noise sensitivity

One of the main challenges of Canny edge detection is that it is sensitive to noise in the image. Noise can be caused by various factors, such as low lighting, compression, or sensor defects. Noise can create false edges or blur the true edges, affecting the accuracy and quality of the edge detection. To reduce the noise, Canny edge detection applies a Gaussian blur filter to the image before applying the gradient and the non-maximum suppression. However, this filter also reduces the sharpness and detail of the edges, creating a trade-off between noise reduction and edge preservation.

添加您的观点

Jafar A.
举报内容
I would say there are two main limitations: (1) this is a parametric method that means you need to devise a kind of logic to define parameters like threshold, which is not easy for real-time applications (2) this is a time consuming algorithm due to its sequential structure

已翻译

赞

2 Parameter selection

Another challenge of Canny edge detection is that it requires the selection of two parameters: the low and high thresholds for the hysteresis thresholding. These parameters determine which pixels are considered as edges, weak edges, or non-edges. The optimal values of these parameters depend on the characteristics of the image, such as the contrast, the edge density, and the noise level. Choosing too low or too high values can result in either missing edges or too many false edges. Therefore, finding the best parameters for a given image can be difficult and time-consuming. One possible solution is to use adaptive thresholding, which adjusts the parameters based on the local statistics of the image.

添加您的观点

Dylan Brown

Graduate Electronic Engineer. BEng Honours in Intelligent Systems. Masters postgraduate student at the University of Pretoria.
举报内容
One of the primary challenges is its requirement for thresholding parameters. Although adaptive thresholding can be used, there does not currently exist a method that can select the perfect parameters for every possible scene/image. This becomes a problem when implementing the edge detector in dynamic environments, where scene and setting changes occur regularly. When implemented on a set of images with similar characteristics, an optimal set of parameters can be found, but this is not viable for real-world usage.

已翻译

赞

3 Edge thinning

A limitation of Canny edge detection is that it does not guarantee that the edges are thin and continuous. The non-maximum suppression step aims to thin the edges by keeping only the pixels with the maximum gradient magnitude along the edge direction. However, this step can also introduce gaps or breaks in the edges, especially if the image is noisy or the edge direction is not well defined. Moreover, the hysteresis thresholding step can also create thick or double edges if the weak edges are connected to the strong edges. To overcome this limitation, some post-processing techniques can be applied, such as morphological operations, edge linking, or contour tracing.

添加您的观点

4 Edge semantics

Another limitation of Canny edge detection is that it does not provide any semantic information about the edges. It only detects the pixels where there is a significant change in intensity, but it does not distinguish between different types of edges, such as object boundaries, shadows, textures, or occlusions. This can make it difficult to use Canny edge detection for higher-level tasks, such as object recognition, segmentation, or scene understanding. To address this limitation, some extensions of Canny edge detection have been proposed, such as incorporating color, texture, or depth information, or using machine learning models to learn edge features.

添加您的观点

5 Edge orientation

A final limitation of Canny edge detection is that it does not capture the orientation of the edges. It only computes the direction of the gradient, which is perpendicular to the edge direction. However, the orientation of the edges can be useful for describing the shape, pose, or perspective of the objects in the image. For example, horizontal edges can indicate a flat surface, while vertical edges can indicate a wall or a pillar. To obtain the orientation of the edges, some additional steps can be performed, such as computing the Hough transform, fitting lines or curves, or using orientation histograms.

添加您的观点

6 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

添加您的观点

Image Processing

+ 关注

给文章评分

我们借助人工智能创建了此文章。您认为这篇文章怎么样？

很棒不太好

举报此文章

What are some challenges and limitations of Canny edge detection in real-world scenarios?

1

2

3

4

5

6

1 Noise sensitivity

2 Parameter selection

3 Edge thinning

4 Edge semantics

5 Edge orientation

6 Here’s what else to consider

Image Processing

给文章评分

感谢您的反馈

更多Image Processing相关文章

更多相关阅读内容