Creating a grayscale png image from a vector of int.

[J.Hilk]

@Uday-More said in Creating a grayscale png image from a vector of int.:

This should be ... 4096.0 ....

why ? you wrote:

values range from 0 - 4000.

Able to generate images now.

Great ! Don't forget to use the topic tools to set the topic to solved then

[Uday More]

@J-Hilk . Thnx for the input.
I have a doubt arranging the vector of the "mappedData"
We are forming the image with width 128 and height as 32 and the 3rd argument is "bytesPerLine" which was given as 128.
What is the importance of "bytesperline".

Generated Images are below.

It is repeating the sections of image.

[J.Hilk]

@Uday-More

What is the importance of "bytesperline".

well its part of the constructor :P

for a Format_Grayscale8 Image, the bytes per line is equal to your pixels for your width,
for a Format_Grayscale16 image, the bytes per line would be double the pixels for your width, because your Pixel information is composed of 2 bytes instead of 1

[Uday More]

@J-Hilk . I tried to convert the values from 0-65536 (16 bit)

1. There is no Format_Grayscale16 . It is written (// reserved for future use). Therefore, I use Format_RGB16. Is it correct.
2. I used a QVector<quint16>. It gives be a build error "QImage::QImage(QImage &&) noexcept': cannot convert argument 1 from 'unsigned short *' to 'uchar *".

What type of vector should be used in such cases to have values ranging from 0-65536 ?

[J.Hilk]

@Uday-More said in Creating a grayscale png image from a vector of int.:

Format_RGB16

You'll need to know exactly, what you want to do and what you're doing.

Format_RGB16 means, It expects a 16 bit value for Red and 16 bit value for Green and 16 bit value for Blue -> 46 bits (6Bytes) per pixel -> 768 BytesperLine (with a width of 128 from before)

Therefore your seeder QVector<quint16> needs to have a size of 12.288. Does it?

I used a QVector<quint16>. It gives be a build error "QImage::QImage(QImage &&) noexcept': cannot convert argument 1 from 'unsigned short *' to 'uchar *".

Please show the correlating code segment

What type of vector should be used in such cases to have values ranging from 0-65536?

I assume you mean 65535, because 65536 requires more than 16 bit.
QVector<quint16> sounds appropriate

[Uday More]

1. I, earlier need 0-255 (8 bit) which worked properly
2. Yes Format_RGB16 is what I need now. Similar to image given below. We can see there is blue element.

3. I meant 65535

4)As explained earlier, i derived this statement of
QVector<quint16> mappedData;
QImage greyScaleImage(mappedData.data(), 32, 128, 768, QImage::Format_RGB16);

This gives me an error "QImage::QImage(QImage &&) noexcept': cannot convert argument 1 from 'unsigned short *' to 'uchar *"

From qglobal.h header i found this.
typedef unsigned char quint8; /* 8 bit unsigned /
typedef unsigned short quint16; / 16 bit unsigned */

[jsulm]

@Uday-More said in Creating a grayscale png image from a vector of int.:

greyScaleImage

This is not clear to me: if this is a greyscale image how can it have blue color?
What format is it actually: 16bit per pixel greyscale? Or RGB with 16bit per color channel?

[Uday More]

@Uday-More said in Creating a grayscale png image from a vector of int.:

768

Code goes below

// Fetching the intensity values from the cloud
int pointsSize = m_pCurrentSetCloud.get()->points.size();
QVector<float> intensityVector;
for(int i = 0; i < pointsSize; ++i)
{
    intensityVector.push_back((m_pCurrentSetCloud.get()->points[i].intensity));
}

float maxValue = *std::max_element(intensityVector.begin(), intensityVector.end());

// Scaling the values to 0-65535
auto mapRange = [maxValue](float input)->quint16 {
    return static_cast<quint16>((input / maxValue) * 65536);
};

QVector<quint16> mappedData(pointsSize);
for(int i(0); i < pointsSize; i++){
    mappedData[i] = mapRange(intensityVector.at(i));
}

// Creating the 16bit RGB image and rotating it by 270 degrees fro display
QImage rgbImage(mappedData.data(), 32, 128, 768, QImage::Format_RGB16);
QImage rotatedImage = rgbImage.transformed(QMatrix().rotate(270.0));
rotatedImage.save("intensity.png");

[Uday More]

May be I am complicating a little bit. I don't have much knowledge on the QImage and bytes needed for information.

1. I generated a grayscale image previously from which i understand that each pixel is holding a 8 bit value (0-255). Similar to this can I have 8 bit RGB ? How will this be arranged ? How many bytes foreach component ? If we arrange this in 32*128, will it be 32 bytes per line ?

[jsulm]

@Uday-More said in Creating a grayscale png image from a vector of int.:

If we arrange this in 32*128, will it be 32 bytes per line ?

No. It will be 32 * number_of_columns bytes.
I assume you mean QImage::Format_RGB32 which uses 32 bytes per pixel.

From the documentation: "QImage::Format_RGB32 4 The image is stored using a 32-bit RGB format (0xffRRGGBB)."
So, first byte 0xFF, then red, green, blue.