Convert PPM Image to Greyscale C++

Question

Trying to convert a PPM Image to greyscale by indexing the pointer containing the pixel data:

void PPMObject::greyScale()
{
    const float r = 0.299F;
    const float g = 0.587F;
    const float b = 0.114F;

    int size = this->width * this->height * 3;
    for (int i = 0; i < size; i++)
    {
        this->m_Ptr[i] = (this->m_Ptr[i] * r) + (this->m_Ptr[i] * g) + (this->m_Ptr[i] * b);
        this->m_Ptr[i+1] = (this->m_Ptr[i+1] * r) + (this->m_Ptr[i+1] * g) + (this->m_Ptr[i+1] * b);
        this->m_Ptr[i+2] = (this->m_Ptr[i+2] * r) + (this->m_Ptr[i+2] * g) + (this->m_Ptr[i+2] * b);
    }
}

Where I've read the PPM Image file using the >> overload:

istream& operator >>(istream &inputStream, PPMObject &other)
{
    inputStream >> other.magicNum >> other.width >> other.height >> other.maxColorValue;
    inputStream.get();
    size_t size = other.width * other.height * 3;
    other.m_Ptr = new char[size];
    inputStream.read(other.m_Ptr, size);
    return inputStream;
}

I write the data as follows:

ostream& operator <<(ostream &outputStream, const PPMObject &other)
{
    outputStream << other.magicNum  << " "
    << other.width          << " "
    << other.height         << " "
    << other.maxColorValue  << " "
    ;   
    outputStream.write(other.m_Ptr, other.width * other.height * 3);
    return outputStream;
}

There are no problems with reading or writing the PPM file.

The problem is simply converting the PPM Image to greyscale -- indexing is not the approach. The file is unchanged.

The question is likely: How do I get the values from the pointer to manipulate them?

E.g., Where are the pixels located in the char pointer?

Averaging the RGB component values is certainly another approach but then how do I assign the average back into the pointer?

score 2 · Accepted Answer · answered Mar 08 '15 at 06:53

Inside your greyScale() function you need to increment i by 3 each time you step through the loop, because each pixel takes up 3 bytes and you process one pixel at a time:

for (int i = 0; i < size; i+=3)

Also, the formula you are currently using will leave the values unchanged (ignoring rounding and floating point errors).

This:

this->m_Ptr[i] = (this->m_Ptr[i] * r) + (this->m_Ptr[i] * g) + (this->m_Ptr[i] * b);

simplifies to:

 this->m_Ptr[i] = (this->m_Ptr[i] * 1.0F);

The correct formula is something like this (ignoring casts and assuming the data is in RGB order):

for (int i = 0; i < size; i+=3)
{
    float greyscaleValue = (this->m_Ptr[i] * r) + (this->m_Ptr[i+1] * g) + (this->m_Ptr[i+2] * b);
    this->m_Ptr[i] = greyscaleValue;
    this->m_Ptr[i+1] = greyscaleValue;
    this->m_Ptr[i+2] = greyscaleValue;
}

Thanks! With more experience I should catch these mistakes myself. — user4640007, Mar 08 '15 at 07:13

score 0 · Answer 2 · answered Feb 10 '21 at 03:23

The following program I wrote earlier today successfully converts a color image (.ppm) file to a grayscale image (.ppm) file. I added a lot of comments so you could easily follow along.

#include <fstream>
#include <iostream>
#include <string>

using namespace std;

int main() {
    //File input and output streams
    ifstream fin("colorImage.ppm");
    ofstream fout("grayscaleImage.ppm");
    //Check if input file was successfully opened
    if (!fin) {
        cout << "Error - Nonexistent Image (.ppm) File" << endl;
        system("pause");
        return -1;  // Error exit
    }
    //Declare necessary variables
    string magic_number;
    int pixel_per_row, num_rows, color_depth, red, green, blue;
    //Read in values for the following variables from input file
    fin >> magic_number >> pixel_per_row >> num_rows >> color_depth;
    //Write the following variables to the output file
    fout << magic_number << endl << pixel_per_row << " " << num_rows << endl << color_depth << endl;
    //Read in input file until file end putting values into appropriate variables
    while (fin >> red >> green >> blue) {
        red = green = blue = int(0.3 * red + 0.59 * green + 0.11 * blue); //Covert each pixel to grayscale
        fout << red << endl << green << endl << blue << endl; //Write converted values to output file
    }
    //Close files
    fin.close();
    fout.close();
    return 0;
}

Convert PPM Image to Greyscale C++

2 Answers2