74 Gray scale and Sobel filter :: Quicker, easier and cheaper to make your own chip!

How it works

This project performs grayscale conversion and Sobel filtering with the aim of detecting edges in an image.

Below is a block diagram of the implementation:

arc

How to test

It is necessary for the pixels to be sent via an SPI protocol; for this purpose, the input ui_in[2:0] is designated as follows:

ui_in[0] → SPI Clock
ui_in[1] → Chip Select
ui_in[2] → Input Pixel

As shown in the previous image, there are some processing options:

Bypass → Returns the input pixel unprocessed.
Grayscale → Returns the pixel converted to grayscale, so it is recommended that the input pixel be RGB.
Sobel → Returns the edge detection corresponding to the input pixel, so it is recommended that the input pixel be grayscale.
Grayscale + Sobel → Returns the edge detection of the input pixel by performing both grayscale processing and the Sobel filter, so it is recommended that the input pixel be RGB.

To select one of the processing options, the input ui_in[4:3] is designated as follows:

ui_in[4:3] = $00$ → Grayscale + Sobel
ui_in[4:3] = $01$ → Sobel
ui_in[4:3] = $10$ → Grayscale
ui_in[4:3] = $11$ → Bypass

To perform the Sobel filter processing, it must be enabled according to the selected processing. This can be enabled or disabled as needed through the input ui_in[5], where $1$ enables and $0$ disables.

The result of the processing corresponds to the output uo_out[0].

There is also a functionality for the input to the different processing options to come from an internal LFSR block; for this purpose, the pins uio_in[3:2] are dedicated for input.

External hardware

Any device that allows sending data via an SPI protocol, like a Raspberry Pi.

#	Input	Output	Bidirectional
0	spi_sck_i	spi_sdo_o	LFSR_enable_i
1	spi_cs_i	lfsr_done	seed_stop_i
2	spi_sdi_i	ena	lfsr_en_i
3	select_process_i[0]	output_px[0]
4	select_process_i[1]	output_px[1]
5	start_sobel_i	output_px[2]
6		output_px[3]
7		output_px[4]

74 Gray scale and Sobel filter