From 5720604042454c3c001afc32519728f7bd61a82a Mon Sep 17 00:00:00 2001 From: Nonannet Date: Wed, 5 Feb 2020 17:26:35 +0100 Subject: [PATCH] Create README.md --- README.md | 60 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 60 insertions(+) create mode 100644 README.md diff --git a/README.md b/README.md new file mode 100644 index 0000000..7dff6de --- /dev/null +++ b/README.md @@ -0,0 +1,60 @@ +# imagerec +```text +A simple and lightweight image recognition tool for small Linux based +embedded systems like the MT7688AN SoC. + +Images can be captured from v4l2 camera devices and recognizes lines +and circles using hough transformation. + +The tool can be controlled by command line parameters or text over TCP/IP. + +The TCP/IP interface is dessigned to be used by scripts (e.g. micro python) +localy on the embeded device or remotly from a PC for easy development. +Netcat can be very usefull for testing. + +Is uses only fixed point operations and is optimized for low memory devices. + +Usage: imagerec [-options] +options: + -h show help + -c commands ASCII command string + -d port TCP/IP-port to listen on + -p device capture device name + -r whidth*hight image resolution + -n pixels max number of pixels for l command + -f file file path for w command (# for index) + -d return result data only + +examples: imagerec -d /dev/video0 -r 640*480 -p 5044 + imagerec -d /dev/video0 -r 640*480 -c cgnexCngml -s + imagerec -d /dev/video0 -r 640*480 -c cgnexCnw -f result.tif + imagerec -d /dev/video0 -r 640*480 -c cgnwexsonwrCnw -f result#.tif + +single byte ASCII commands: + c capture image + n normalize image + g apply gausian blur + e edge detection with sobel filter, must be + folowd by 'x', 'o' or a hough transformation + x remove non edge pixel, must be folowd by + 'o' or a hough transformation + o convert directional slope to absolute slope + C circle hough transformation + L line hough transformation + H line hough transformation (horizontal only) + V line hough transformation (vertical only) + M miniscus hough transformation + b binarize + m remove non-local-maxima pixels + l list brightes pixels (from max. 32 non black pixels) + p list brightes pixel clusters (3x3) + q close connection + s store a copy of the current buffer + r recall a copy of the stored buffer + w write buffer to disk (TIF format) + z set index for output file name to zero + d show result data only + i show info and result data (default) + +example: echo \"cngexCngmlq\" | nc localhost 5044 +```