Sometimes the learning curve stays stuck for a long time.... Force the GA to reset and start over

FAIL! : Vision_Test::TestCase_LabelBlobs() Compared values are not the same
Actual (totalNoOfBlobs) : 3940
Expected (uint16_t(NO_OF_BLOBS)): 379

Takes a lot of processing time

VSA_UT: /usr/include/boost/archive/basic_binary_iprimitive.hpp:98: void boost::archive::basic_binary_iprimitive<Archive, Elem, Tr>::load(bool&) [with Archive = boost::archive::binary_iarchive; Elem = char; Tr = std::char_traits]: Assertion `0 == i || 1 == i' failed.

Make an analysis class for the PSD function

When the copy constructor is called for the Stats class. all the data point are also copied. This is inefficient and unnecessary.

In accordance with the new working height

Presumably something goes wrong with the depth search

  // Down the rabbit hole
  for (uint32_t i = 0; i < connectedLabels.size(); i++) {
      for (uint32_t j = 0; j < connectedLabels[i].size(); j++) {
          uint16_t CurrentVal = connectedLabels[i][j];
          uint16_t PrevVal = i;
          while (CurrentVal != PrevVal) {
              PrevVal = CurrentVal;
              CurrentVal = connectedLabels[PrevVal][0];
            }
          connectedLabels[i][j] = CurrentVal;
        }
    }

in roundness en angularity

When the PrepImg function is called a sigsev fault occurs

Sample doens't load and gets stuck on PSD gives a sigsev fault

http://collegerama.tudelft.nl/Mediasite/Play/556ad0089bfd415a949ef9d75993986f1d?PlayFrom=5150000

When the merged HDR mat (expected value range 0. till 1.) is converted to a uchar (0 - 255) a SIGSEV fault occurs.

Only redraw the changed graphs
Work with std::vector without converting

Algorithm that determines if blob consist of multiple particles so that it can exclude these blobs from the PSD.

http://www.frogatto.com/?p=26

Using a vector of shared_ptr removes the possibility of leaking memory because you forgot to walk the vector and call delete on each element. Let's walk through a slightly modified version of the example line-by-line.

typedef boost::shared_ptr gate_ptr;
Create an alias for the shared pointer type. This avoids the ugliness in the C++ language that results from typing std::vectorboost::shared_ptr and forgetting the space between the closing greater-than signs.

std::vector<gate_ptr> vec;

Creates an empty vector of boost::shared_ptr objects.

gate_ptr ptr(new ANDgate);

Allocate a new ANDgate instance and store it into a shared_ptr. The reason for doing this separately is to prevent a problem that can occur if an operation throws. This isn't possible in this example. The Boost shared_ptr "Best Practices" explain why it is a best practice to allocate into a free-standing object instead of a temporary.

vec.push_back(ptr);

This creates a new shared pointer in the vector and copies ptr into it. The reference counting in the guts of shared_ptr ensures that the allocated object inside of ptr is safely transferred into the vector.

What is not explained is that the destructor for shared_ptr ensures that the allocated memory is deleted. This is where the memory leak is avoided. The destructor for std::vector ensures that the destructor for T is called for every element stored in the vector. However, the destructor for a pointer (e.g., gate*) does not delete the memory that you had allocated. That is what you are trying to avoid by using shared_ptr or ptr_vector.

Since the data eventually has to be presented as a CFD. add the matrix to the class

The Stats class works with a uniform binrange but the sieves in PSD differ in range.

• 2
• 1.4
• 1
• 0.71
• 0.5
• 0.355
• 0.25
• 0.18
• 0.125
• 0.09
• 0.075
• 0.063
• 0.045
• 0.038
• 0

Make a calibration discs, and write a calibrate function that determines the radius (Hough Circle Transform) and compares it to the known SI dimension. So to determine the SIfactor

Progressbar update should work correctly

Neural network where the input is the chromatic a and b of the current pixel and neighbors (maybe described as statistics) the network is learned by letting the user mark the particles... And letting that be the outcome.

Use seperate save and load for the matrix and make array. Not the for loop

The current Sample class saves all the images, including the original image and all the processing steps.
This doesn't really when the sample consist of one picture taken but it starts to count when the sample will consist of multiple shot with front light and projected light.

The user isn't interested in the complete picture but only the particles. The reworked GUI will take this into account.

Also a lot of learning faults have to be rectified in the new GUI

So that a sample can be emailed to another computer

Integrate the beaglebone and ledlighting from below

The GetEdgeErosion function works but when it standalone with a bigger matrix there're issues

http://doc.qt.io/qt-5/qtextdocument.html
http://www.qcustomplot.com/index.php/tutorials/specialcases/textdocument

output to printer, email or disk

Find optimum settings to segment the particles from the background

This function gets called a lot and is relatively costly

Move Welch test from test to Soilmath::Stats

http://www.frogatto.com/?p=26

vector::iterator end_vaild = remove_if(v.begin(), v.end(), particle_expired);
v.erase(end_valid, v.end());