2 Observations and data reduction

The ${\rm H}_{\alpha}$ observations used in this work were taken at Sacramento Peak Observatory on December 14, 1996. Our image (Fig. 1) has dimensions $128\times128$ pixels with 42 different intensities. The minimum value is 32 and the maximum is 222. The resolution is 0.4 arcsec/pixel and thus our image contains an area of $51.2 \times 51.2$ $\mbox {arcsec}^2$.As a first step we sharpen this image to enhance the local contrast by convolving it with the following template:

$$\left(\begin{array} {ccc} -1 & -2 & -1 \\ -2 & 13 & -2 \\ -1 & -2 & -1\end{array}\right)\cdot$$

The new image (Fig. 2) has a minimum of -20 and a maximum of 330. Let I(x,y) be our image matrix of pixel intensities of (x,y) (Fig. 1). The image after sharpening (Fig. 2) has intensities I₁ (x,y). Let now ${\cal M} = {\rm max}\left(I_1 (x,y) \right)$ and ${\it m} = {\rm min}\left(I_1 (x,y) \right)$. We reduce our image (Fig. 2) into q classes (Figs. 3, 4) using the following equation:

 

$$I_2 (x,y)= {\rm int}\left( \frac{I_1 (x,y) - \it {m}}{\cal {M} - \it {m}}~(q-1) \right).$$ (1)

  

Figure 1: Selected sunspot image $(128 \times 128)$ containing an area of $51.2 \times 51.2$ $\mbox {arcsec}^2$

  

Figure 2: Sunspot image $(128 \times 128)$ after sharpening of Fig. 1 with template of Sect. 2

  

Figure 3: Sunspot image $(128 \times 128)$ after application of Eq. (1) in Fig. 2 with q=4 labels

  

Figure 4: Sunspot image $(128 \times 128)$ after application of Eq. (1) in Fig. 2 with q=8 labels