image-processing - エッジ検出は画像の特徴に依存しますか?

Question

こんにちは、エッジ検出が画像条件 (機能) に依存するかどうかを調べようとしています。

エッジ検出演算子には巨大な数学的根拠があることを私は知っています。また、エッジ検出は画像のノイズに敏感であることも知っています。

明るさ、コントラストはどうですか？ポイントは、画像の品質を推定する方法を探していることです。画像の品質はエッジ検出にとって重要ですか?

Answer 1

Edges are detected where there is a change in pixel value in either x or y direction in an image. The maths behind this is simple differentiation. Any variation or noise that can change the pixel value can reduce the chances of detecting an edge but then morphological operations can help.

For example, blur is one operation that can reduce the image quality by changing the pixel values. Figure 1 represents an image and its edges.. As I have already mentioned edges are detected where the pixel value changes in one direction, you can see the white lines as edges corresponding to these changes in pixel value. Figure 2 is blurred image of input image, the edges of this blur image is much less than the actual number of edges..

That is just one example, noise while capturing the image, illumination of object or dark object can give different edges. Depending upon how the noise is affecting the image the edges can increase or decrease.

There are some basic methods of detecting edges, I have used canny edge detection. You can refer to review of classic edge detectors to understand them further.