This is a simple framework that allows for developers to quickly test out their python opencv programs and view the results on a website. The python programs are separated into 3 image.py, camera.py, video.py. Each script captures a image from a separate source i.e. image.py captures a static image saved in the res folder, video.py captures a series of images from a video file saved in the res folder and camera.py captures a series of images from a camera connected to the device.
To view the Output form the different scripts, once you run the server according to the instructions below, go to the urls e.g. localhost:5800/image.html or localhost:5800/camera.html or localhost:5800/video.html
To add your program, simply insert your image manipulation algorithm in the spaces designated as shown below
'''
if (message == '1'):#Multiple Image displaying algorithms can be added below #Replace your own code here. Below is a simple example
img = cv2.imread('res/facetest.jpg')
#The code below sends the image to the Image viewer. DO NOT DELETE
#the image is passed to the encoder as img, you can pass any other image matrix
encode_param=[1,90]
res, image = cv2.imencode('.jpg',img,encode_param)
sample = base64.b64encode(image)
self.sendMessage(sample, isBinary)
#if (message == '2'): #add extra algorithm like this
'''
The img variable stores the image that can be manipulated on command from the website. The python script is basically a websocket server so you can run this remotely on your raspberry pi or your computer and access from a different server, provided you connect to the ip address of the raspberry pi.
##Installation and Running (DEBIAN)
-
The prerequisite programs required to run the program are python, pip, opencv, pyopencv, numpy, autobahn(websocket library), twisted. To simplify the installation process I have made a script to install the various components. After installing nodejs and npm just run "npm install" followed by "npm run setup-deb" on your command line which should do the rest. This is made for the linux operating system, specifically debian based systems i.e. with apt-get.
-
In your command line typing "npm run server" will start up the web server, although you can start it from a remote pc.
-
To start the python scripts just run "npm run imageServer" for the image server, "npm run cameraServer" for the camera server and "npm run videoServer" for the video server onto your command line. This will run image.py, video.py, camera.py log their outputs to image.log, video.log and camera.log respectively.
##Installation (Windows)
-
To run the programs in windows, you can still run the "npm install" then "npm setup-win" This is assuming that you have already installed opencv, python, pip, nodejs and npm.
-
In your command line typing "npm server" will start up the web server, although you can start it from a remote pc.
-
To start all the python scripts just run "npm run-scripts" onto your command line. This will run image.py, video.py, camera.py log their outputs to image.log, video.log and camera.log respectively.
##Installation and Running (ARMSTRONG--not tested)
-
The prerequisite programs required to run the program are python, pip, opencv, pyopencv, numpy, autobahn(websocket library), twisted. To simplify the installation process I have made a script to install the various components. After installing nodejs and npm just run "npm install" followed by "npm run setup-edison" on your command line which should do the rest. This is made for the linux operating system, specifically debian based systems i.e. with apt-get.
-
In your command line typing "npm run server" will start up the web server, although you can start it from a remote pc.
-
To start the python scripts just run "npm run imageServer" for the image server, "npm run cameraServer" for the camera server and "npm run videoServer" for the video server onto your command line. This will run image.py, video.py, camera.py log their outputs to image.log, video.log and camera.log respectively.
##Installation in other systems
- To add installation scripts for more systems please feel free to contribute.