Documentation imprivement

This commit pushes additional documentation for the repo.

Doxyfile

@@ -771,7 +771,7 @@ WARN_LOGFILE           =
 # spaces. See also FILE_PATTERNS and EXTENSION_MAPPING
 # Note: If this tag is empty the current directory is searched.
 
-INPUT                  = include/ include/ball_tracking
+INPUT                  = include/ include/ball_tracking README.md
 
 # This tag can be used to specify the character encoding of the source files
 # that doxygen parses. Internally doxygen uses the UTF-8 encoding. Doxygen uses
@@ -920,7 +920,7 @@ FILTER_SOURCE_PATTERNS =
 # (index.html). This can be useful if you have a project on for instance GitHub
 # and want to reuse the introduction page also for the doxygen output.
 
-USE_MDFILE_AS_MAINPAGE =
+USE_MDFILE_AS_MAINPAGE = README.md
 
 #---------------------------------------------------------------------------
 # Configuration options related to source browsing

README.md

 Ball Tracking Code
 ==================
 
+This library for ball tracking contains the implementation of the methods presented on this 
+[paper](https://www.mdpi.com/2218-6581/8/4/90). Some of the design goals of the library are:
+
+* Support different camera hardware
+* Support different algorithms along the pipeline for object detection
+* Easy to configure from a configuration file (JSON format)
+* Publisher-Subscriber architecture servers for the vision system
+* The algorithms can run on a CPU or GPU
+
+### Support for different camera hardware
+
+* The camera support is implemented in a separate library called 
+[libcamera](https://gitlab.tuebingen.mpg.de/sgomez/camera).
+* At the time of this writing we only implemented the **PvAPI** interface for Prosillica cameras. 
+* To support new cameras you only need to implement a class that Inherits from **CamSetDriver**
+* Objects from the CamSetDriver class receive the camera set configuration and a callback to be called
+when new images are received.
+
+## Algorithm Pipeline
+
+Implemented in "tracker.hpp":
+
+* **BallLogLikelihood**: Receives an image and returns a likelihood image (of each pixel being the ball)
+* **Binarizer**: Turns the likelihood image into a binary image (Ball, Not ball)
+* **FindBallBlob**: Returns all candidate locations of the ball
+
+**Note**: The algorithms used are implemented both in the CPU and the GPU
+
+### Configuration
+
+How do we select:
+* Which algorithms to use on each step of the pipeline
+* The model parameters (when using ML)
+* The hyperparameters of **Binarizer** and **FindBallBlob**
+* If the algorithms should run in the CPU or GPU
+
+We do all that using a configuration file in JSON format. See an example in 
+[examples/tracking/cpu_track_conf.json](./examples/tracking/cpu_track_conf.json)
+
+## Publisher - Subscriber
+
+We have two separate important processes that communicate using the Publisher-Subscriber architecture
+* [server](./examples/tracking/server.cpp): Broadcasts the 2D pixel location of the ball in all the images to
+all its subscribers
+* [server3d](./examples/tracking/server3d.cpp): Subscribes to the 2D server and broadcasts the 3D position
+in world coordinates of the ball to all its subscribers
+
+Both services use configuration files (JSON). You can see examples in the example folder.
+
+
 ## Installation
 
 The first step to install this library is to install the pre-requisits. The easiest way to install these

src/stereo.cpp

@@ -169,8 +169,11 @@ namespace ball_tracking {
               if (curr.is_obs[i]) BOOST_LOG_TRIVIAL(trace) << "Obs " << i << ": (" << curr.obs2d[i][0] << ", " << curr.obs2d[i][1] << ")";
               else BOOST_LOG_TRIVIAL(trace) << "Obs " << i << ": null";
             }
-            robotics::pt3d pos3d; 
+            robotics::pt3d pos3d;
+            auto t1 = std::chrono::high_resolution_clock::now();
             bool success = stereo(pos3d, curr);
+            auto t2 = std::chrono::high_resolution_clock::now();
+            double proc_time = std::chrono::duration_cast<std::chrono::milliseconds>(t2-t1).count();
             zmqpp::message out;
             json out_obs;
             if (success) {
@@ -179,6 +182,7 @@ namespace ball_tracking {
             json out_msg {
               {"num", num},
               {"time", time},
+              {"proc_time", proc_time},
               {"obs", out_obs}
             };
             BOOST_LOG_TRIVIAL(trace) << "Sending message: " << out_msg.dump();