In my unity project I have an image file which contains information. I would like to be able to read this .bmp file pixel by pixel. So I tried simply loading a sprite, and then reading its texture. However when I do that, the loaded texture appears to be different than the image I'm using, as if it was compressed. I tried modifying the asset settings (making it an advanced texture with point filter mode, and setting format to argb32bit uncompressed) but it doesn't help. Any ideas?
EDIT: Some additional information,
The code with which I load the image into a pixel array
// sprites from which paths will be loaded
public Sprite[] path_maps;
// Color 24 array which contains the texture bitmap
private Color32[] pixels;
// graph with paths
public Graph[] graphs;
void Start () {
// I skipped some code that doesn't matter to the problem
// go through all path maps
for (int i = 0; i < path_maps.Length; i++) {
// load sprite into pixels array
pixels = path_maps[i].texture.GetPixels32 ();
// create new graph
graphs [i] = new Graph ();
// get graph from path map
graphs[i].pixels_to_graph(pixels, path_maps[i].texture.width, path_maps[i].texture.height, camera_width, camera_height);
}
}
The function which reads pixels and creates node for each colored pixel (only relevant part of the function)
public void pixels_to_graph (Color32[] pixels, int width, int height, double camera_width, double camera_height)
{
// assign pixels width and height
pixels_width = width;
pixels_height = height;
// assign camera dimensions
this.camera_width = camera_width;
this.camera_height = camera_height;
// First create as many nodes as there are colored pixels (simply read the pixel map)
// Note that pixel map contains the bitmap from bottom to the top (left bottom to top right)
// create nodes array
pixel_nodes = new Node[width*height];
// current position in pixel array. Made for easier calculations
int carriage = 0;
// go through all rows
for (int i = 0; i < height; i++) {
// go column by column
for (int j = 0; j < width; j++) {
// if it's transparent or white - ignore
if (pixels [carriage].a == 0 || pixels [carriage].Equals (col_white))
continue;
// else add node to the node array
pixel_nodes [carriage] = new Node (j, i, pixels_width, pixels_height, camera_width, camera_height); // note that i is the y position while j is the x position
// add node to nodes list
nodes.Add (pixel_nodes [carriage]);
carriage++;
}
}
I detect the problem in the above pixels_to_graph function. When I debug it with a 1680x1050 image, that has pixels in its bottom left corner (three black squares with 1 pixel distance between them, of size 3x3px, 2x2px, 1x1px) it only detects the 3x3 one, the colour value at position of the other 2 squares is white. When I tried other shapes I also discovered other distortions, some 1px wide lines are not there, as well as single pixels.
My current texture settings (though note I tried different options but some of them don't display anything, this one at least displays something):
