A first attempt at edge detection is shown, including a scaled, low- resolution copy of the unprocessed image, for reference.
| A first attempt at edge detection is shown, including a scaled, low- resolution copy of the unprocessed image, for reference. |
|
| -1 | -1 | -1 |
| -1 | 9 | -1 |
| -1 | -1 | -1 |
Low- Pass Filter:
| 1/16 | 1/16 | 1/16 |
| 1/16 | 1/2 | 1/16 |
| 1/16 | 1/16 | 1/16 |
Edge- Detector (as in the image above:)
| -1 | -1 | -1 |
| -1 | 8 | -1 |
| -1 | -1 | -1 |
Strictly speaking, convolution is defined as the time integral from negative infinity to time t1 of the product of two signals, v1(t).v2(t1-t), where one of the signals has been folded and delayed by an amount of time t1; t1 is a constant as far as the integration is concerned.You are really interested in the value of the convolution for different values of t1. However, for digital signal processing, you are more likely to calculate a finite and discrete version.
The following image is intended to demonstrate that quite a lot of texture information can be retained in a binary (two- tone black- and- white) image, if the threshold is chosen well enough:
Electronics links:
Computing Projects:
The distinction is rather loose: Most of the electronics sections involve use of a computer to some extent, and vice versa.