I am writing a ray tracer program for my computer graphics class. So far I only have spheres implemented and a shadow ray. The current problem is that when i move my sphere off center it stretches. here is the code that i use to calculate if a ray is intersecting a sphere:
bool Sphere::onSphere(Ray r)
{
float b = (r.dir*2).innerProduct(r.pos + centre*-1);
float c = (r.pos + centre*-1).innerProduct(r.pos + centre*-1) - radius*radius;
return b*b - 4*c >= 0;
}
here is the code that i use to spawn each ray:
for(int i = -cam.width/2; i < cam.width/2; i++)
{
for(int j = -cam.height/2; j < cam.height/2; j++)
{
float normi = (float)i;
float normj = (float)j;
Vector pixlePos = cam.right*normi + cam.up*normj + cam.forward*cam.dist + cam.pos*1;
Vector direction = pixlePos + cam.pos*-1;
direction.normalize();
Vector colour = recursiveRayTrace(Ray(pixlePos, direction), 30, 1, 0);
float red = colour.getX()/255;
float green = colour.getY()/255;
float blue = colour.getZ()/255;
fwrite (&red, sizeof(float), 1, myFile);
fwrite (&green, sizeof(float), 1, myFile);
fwrite (&blue, sizeof(float), 1, myFile);
}
}
recursiveRayTrace:
Vector Scene::recursiveRayTrace(Ray r, float maxDist, int maxBounces, int bounces)
{
if(maxBounces < bounces)
return Vector(0,0,0);
int count = 0;
for(int i = 0; i < spheres.size(); i++)
{
if(spheres.at(i).onSphere(r))
{
Vector colour(ambiant.colour);
for(int j = 0; j < lights.size(); j++)
{
Vector intersection(r.pos + r.dir*spheres.at(i).getT(r));
Ray nRay(intersection, lights.at(i).centre + intersection*-1);
colour = colour + lights.at(i).colour;
}
return colour;
}
}
return Vector(0,0,0);
}
What i get is an sphere that is stretched in the direction of the vector from the center to the center of the circle. I'm not looking for anyone to do my homework. I am just having a really hard time debugging this on. Any hints are appreciated :) Thanks!
Edit: cam.dist is the distance from the camera to the view plane
The stretching is actually a natural consequence of perspective viewing and it is exaggerated if you have a very wide field of view. In other words moving the camera back from your image plane should make it seem more natural.
Related
Currently I try to write code for calculating the parts of the screen you can see and those who can't because of objects that block light in 2d, like in Among Us:
The code should run on a processor with very low specs (at least in 2020), the C64. On such a simple CPU it's not possible to do such complex math fast enough for a game, so I came up with an idea: First of all, I make everything tile based, that makes processing easier and also means that I can just change entire characters or their color cells. Then I just write code for the PC in Processing (that's a coding language similar to Java but easier to use) to calculate how rays of light would move (the following graphic should make that more understandable), first just with a rectangle (and a single quadrant):
Then I wrote some completely messy assembler code for using the recorded coordinates to just keep filling the tiles with an inverted character based on the number of the ray currently being drawn on the ray until they hit an object (/ the tile it wants to fill is not inverted and not a space) and then just go to the next ray. I reduced the radius to 7 so it just takes up 256 bytes, useful for ASM. And that totally worked, I was able to fix every single bug and the result was quite impressive, since I needed to add pause statements or everything ran so fast that you couldn't see anything.
After that worked, I tried it with a circle, setting the points using this code:
int pointNum = ceil(radius * PI * 2); // calculates the circumference
for(int i = 0;i < pointNum;i++){
float angle = map(i, 0, pointNum, 0, PI*2);
setPixel(sin(angle) * radius, cos(angle) * radius);
}
I previously used the Bresenham circle algorithm but that didn't quite work so I tried a more simple way. So ...
All the marked black tiles never get hit by any light, which is a pretty big issue, because it wouldn't make much sense in a game that you just can't see those tiles. The code I used, written in Processing, is:
float[] xPoints = new float[0];
float[] yPoints = new float[0];
float[] xPointsT;
float[] yPointsT;
float[] xPointsHad = new float[0];
float[] yPointsHad = new float[0];
int pos = 0;
float interpolPos = 0;
int radius = 12;
float tileSize = 800.0 / (2*radius+1);
String output = " !byte ";
int pointNum = ceil(radius * PI * 2);
void setup() {
size(800, 800);
frameRate(60);
xPointsT = new float[0];
yPointsT = new float[0];
/*for(int i = 0;i <= radius;i++){
setPixel(radius, i);
setPixel(i, radius);
}*/ //Uncomment this and comment the next 4 lines to get the rectangle version
for(int i = 0;i < pointNum;i++){
float angle = map(i, 0, pointNum, 0, PI*2);
setPixel(sin(angle) * radius, cos(angle) * radius);
}
xPoints = concat(xPoints, xPointsT);
yPoints = concat(yPoints, yPointsT);
}
void draw(){
if(interpolPos > radius){
pos++;
interpolPos = 0;
println(output);
output = " !byte ";
}
float x=0, y=0;
float interpolMul = interpolPos / radius;
x = xPoints[pos] * interpolMul;
y = yPoints[pos] * interpolMul;
interpolPos+=1;//sorta the resolution
background(0);
stroke(255);
for(int i = 0;i < 2*radius+1;i++){
for(int j = 0;j < 2*radius+1;j++){
if((round(x) + radius) == i && (round(y) + radius) == j){
fill(0, 255, 0);
if(output != " !byte ")
output += ", ";
output += i-radius;
output += ", ";
output += j-radius;
xPointsHad = append(xPointsHad, i);
yPointsHad = append(yPointsHad, j);
}
else{
int fillVal = 0;
for(int k = 0; k < xPoints.length;k++){
if(round(xPoints[k])+radius == i && round(yPoints[k])+radius == j){
fillVal += 64;
}
}
fill(0, 0, fillVal);
if(fillVal == 0){
for(int k = 0; k < xPointsHad.length;k++){
if(round(xPointsHad[k]) == i && round(yPointsHad[k]) == j){
fill(128, 0, 0);
}
}
}
}
rect(i * tileSize, j * tileSize, tileSize, tileSize);
}
}
strokeWeight(3);
stroke(0, 255, 255, 64);
for(int i = 0;i < xPoints.length;i++){
line((float(radius)+0.5) * tileSize, (float(radius)+0.5) * tileSize, (float(radius)+0.5+xPoints[i]) * tileSize, (float(radius)+0.5+yPoints[i]) * tileSize);
}
strokeWeight(1);
fill(255, 255, 0);
ellipse((x + radius + 0.5) * tileSize, (y + radius + 0.5) * tileSize, 10, 10);
}
void setPixel(float _x, float _y){
for(int i = 0; i < xPoints.length;i++){
if(_x == xPoints[i] && _y == yPoints[i]){
return;
}
}
for(int i = 0; i < xPointsT.length;i++){
if(_x == xPointsT[i] && _y == yPointsT[i]){
return;
}
}
xPointsT = append(xPointsT, _x);
yPointsT = append(yPointsT, _y);
}
(Instructions to get the rectangle are in the code)
Those mentioned tiles seem to be never hit because the rays on them just jump over them, but what can I do to prevent that? You can decrease interpolPos+=x; to hit more tiles because that way your steps are smaller, but that wastes quite some space, so I don't think that's a good solution. Ideally you could also just decrease the number of coordinates you draw to get a smaller vision. Has anyone a good idea how to do that?
You have chosen wrong method to find all touched cells - instead of point-based way you need cell(squares)-based approach - ray intersects rectangle rather than point.
There is article of Amanatides and Woo "A Fast Voxel Traversal Algorithm for Ray Tracing" for 2D.
Practical implementation.
Example:
Quick-made tracing example. Rays emitted from left top corner go to blue points. If ray meets black cell obstacle, it stops. Pink cells are lighted by rays, grey ones are not.
Okay, I found something that worked for me in my situation: I just used the part that totally works (the rectangle) and then just make that a circle by ignoring every tile hit that's further away from the light source then the radius + 0.5, because without + .5 the circle looks weird. You can try it yourself, here's the code:
float[] xPoints = new float[0];
float[] yPoints = new float[0];
float[] xPointsT;
float[] yPointsT;
float[] xPointsHad = new float[0];
float[] yPointsHad = new float[0];
int pos = 0;
float interpolPos = 0;
int radius = 7;
float tileSize = 800.0 / (2*radius+1);
int pointNum = ceil(radius * PI * 2);
String standardOutput = " !align 15,0\n !byte ";
void setup() {
size(800, 800);
frameRate(60);
xPointsT = new float[0];
yPointsT = new float[0];
for(int i = 0;i <= radius;i++){
setPixel(radius, i);
setPixel(i, radius);
} //Uncomment this and comment the next 4 lines to get the rectangle version
/*for(int i = 0;i < pointNum;i++){
float angle = map(i, 0, pointNum, 0, PI*2);
setPixel(sin(angle) * radius, cos(angle) * radius);
}*/
xPoints = concat(xPoints, xPointsT);
yPoints = concat(yPoints, yPointsT);
xPointsT = new float[0];
yPointsT = new float[0];
}
void draw(){
if(interpolPos > radius){
pos++;
interpolPos = 0;
String output = standardOutput;
for(int i = 0;i < radius + 1;i++){
int indexPos = floor(map(i, 0, radius + 1, 0, xPointsT.length));
output += round(xPointsT[indexPos]);
output += ",";
output += round(yPointsT[indexPos]);
if(i < radius){
output += ", ";
}
}
println(output);
xPointsT = new float[0];
yPointsT = new float[0];
}
float x=0, y=0;
float interpolMul = interpolPos / radius;
x = xPoints[pos] * interpolMul;
y = yPoints[pos] * interpolMul;
interpolPos+=1;//sorta the resolution
background(0);
stroke(255);
for(int i = 0;i < 2*radius+1;i++){
for(int j = 0;j < 2*radius+1;j++){
if((round(x) + radius) == i && (round(y) + radius) == j && sqrt(sq(round(x)) + sq(round(y))) < radius + 0.5){
fill(0, 255, 0);
xPointsT = append(xPointsT, i-radius);
yPointsT = append(yPointsT, j-radius);
xPointsHad = append(xPointsHad, i);
yPointsHad = append(yPointsHad, j);
}
else{
int fillVal = 0;
for(int k = 0; k < xPoints.length;k++){
if(round(xPoints[k])+radius == i && round(yPoints[k])+radius == j){
fillVal += 64;
}
}
fill(0, 0, fillVal);
if(fillVal == 0){
for(int k = 0; k < xPointsHad.length;k++){
if(round(xPointsHad[k]) == i && round(yPointsHad[k]) == j){
fill(128, 0, 0);
}
}
}
}
rect(i * tileSize, j * tileSize, tileSize, tileSize);
}
}
strokeWeight(3);
stroke(0, 255, 255, 64);
for(int i = 0;i < xPoints.length;i++){
line((float(radius)+0.5) * tileSize, (float(radius)+0.5) * tileSize, (float(radius)+0.5+xPoints[i]) * tileSize, (float(radius)+0.5+yPoints[i]) * tileSize);
}
strokeWeight(1);
fill(255, 255, 0);
ellipse((x + radius + 0.5) * tileSize, (y + radius + 0.5) * tileSize, 10, 10);
}
void setPixel(float _x, float _y){
for(int i = 0; i < xPoints.length;i++){
if(_x == xPoints[i] && _y == yPoints[i]){
return;
}
}
for(int i = 0; i < xPointsT.length;i++){
if(_x == xPointsT[i] && _y == yPointsT[i]){
return;
}
}
xPointsT = append(xPointsT, _x);
yPointsT = append(yPointsT, _y);
}
Besides the main difference to ignore tiles that are not in the circle, I also changed that I store the coordinates not in a String but in two arrays, because then I use code to stretch them when there are fewer then radius + 1 points, so I don't have to store multiple circles with different sizes in the C64's RAM, so it meets my main requirements: It should fill every tile and it should be downscalable by ignoring some points at the end of rays. And is if efficient? Uh ... there could be a better solution that fills the circle with fewer rays, but I don't care too much. Still, if you have an idea, it would be nice if you could tell me, but otherwise this question is solved.
Edit: I forgot to add a picture. Don't be confused, I modified the code after posting it so you can also see the blue tiles on the circle.
I'm trying to make a real-time advertisement billboard detection in road using android smartphone. The goal is to crop the area of the advertisement billboard object (regions of interest) and save it to database.
For example:
enter image description here
enter image description here
For preprocessing, I used grayscaling and Canny Edge Detection (Otsu thresholding is used to set the upper and lower threshold). Then, I used contour-based method to detect whether the object is rectangular by checking the point. I use Java OpenCV in android studio for implementation. When I run the program, it only detect rectangular object in plain background and if the rectangular having a high contrast from the background. Currently, it can only detect rectangle with 90 degree and it failed to detect object with rounded rectangle shape. Furthermore, my program failed completely to detect rectangular object in a more complex background, like road scene where the object I'm trying to detect is having similar color to the background/low contrast or there are many occlusions like tree, traffic light, and cables which caused the detection to fail.
This is the code I use for edge detection
Mat destination = new Mat(oriMat.rows(), oriMat.cols(), oriMat.type());
Imgproc.cvtColor(oriMat, destination, Imgproc.COLOR_RGBA2GRAY);
Imgproc.GaussianBlur(destination, destination, new Size(3,3), 0, 0, Imgproc.BORDER_DEFAULT);
double otsuThresholdValue = Imgproc.threshold(destination, destination, 0, 255, Imgproc.THRESH_BINARY | Imgproc.THRESH_OTSU);
double lowerThreshold = otsuThresholdValue*0.5;
double upperThreshold = otsuThresholdValue;
Mat canny = new Mat();
Imgproc.Canny(destination, canny, lowerThreshold, upperThreshold);
Mat abs = new Mat();
Core.convertScaleAbs(canny, abs);
Mat result = new Mat();
Core.addWeighted(abs, 0.5, abs, 0.5, 0, result);
Here is the code I use for contour-based detection
ArrayList<MatOfPoint> contours = new ArrayList<>();
// find contours and store them all as a list
Imgproc.findContours(matData.monoChrome.clone(), contours, new Mat(), Imgproc.RETR_EXTERNAL, Imgproc.CHAIN_APPROX_SIMPLE);
final int width = matData.monoChrome.rows();
final int height = matData.monoChrome.cols();
int matArea = width * height;
for (int i = 0; i < contours.size(); i++) {
double contoursArea = Imgproc.contourArea(contours.get(i));
MatOfPoint2f approx = new MatOfPoint2f();
MatOfPoint2f contour = new MatOfPoint2f(contours.get(i).toArray());
double epsilon = Imgproc.arcLength(contour, true) * 0.1;
// Imgproc.minAreaRect(contour);
// approximate contour with accuracy proportional to the contour perimeter
Imgproc.approxPolyDP(contour, approx, epsilon, true);
if (Math.abs(contoursArea) < matArea * 0.01 ||
!Imgproc.isContourConvex(new MatOfPoint(approx.toArray()))) {
continue;
}
Imgproc.drawContours(matData.resizeMat, contours, i, new Scalar(0, 255, 0));
List<Point> points = approx.toList();
int pointCount = points.size();
LinkedList<Double> cos = new LinkedList<>();
for (int j = 2; j < pointCount + 1; j++) {
cos.addLast(angle(points.get(j % pointCount), points.get(j - 2), points.get(j - 1)));
}
Collections.sort(cos, (lhs, rhs) -> lhs.intValue() - rhs.intValue());
double mincos = cos.getFirst();
double maxcos = cos.getLast();
if (points.size() == 4 && mincos >= -0.3 && maxcos <= 0.5) {
for (int j = 0; j < points.size(); j++) {
Core.circle(matData.resizeMat, points.get(j), 6, new Scalar(255, 0, 0), 6);
}
matData.points = points;
break;
}
}
Is there any method I can use to recognize advertisement billboard in road?
I would appreciate any answers and ideas. Thank you!
I'm making a program to blur a 16 bit grayscale image in CUDA.
In my program, if I use a Gaussian blur function with sigma = 20 or 30, it takes a lot of time, while it is fast with sigma = 2.0 or 3.0.
I've read in some web site that Guaussian blur with FFT is good for large kernel size or large sigma value:
Is It really true ?
Which algorithm should I use: simple Gaussian blur or Gaussian blur with FFT ?
My code for Guassian Blur is below. In my code , is there something wrong or not ?
enter code here
__global__
void gaussian_blur(
unsigned short* const blurredChannel, // return value: blurred channel (either red, green, or blue)
const unsigned short* const inputChannel, // red, green, or blue channel from the original image
int rows,
int cols,
const float* const filterWeight, // gaussian filter weights. The weights look like a bell shape.
int filterWidth // number of pixels in x and y directions for calculating average blurring
)
{
int r = blockIdx.y * blockDim.y + threadIdx.y; // current row
int c = blockIdx.x * blockDim.x + threadIdx.x; // current column
if ((r >= rows) || (c >= cols))
{
return;
}
int half = filterWidth / 2;
float blur = 0.f; // will contained blurred value
int width = cols - 1;
int height = rows - 1;
for (int i = -half; i <= half; ++i) // rows
{
for (int j = -half; j <= half; ++j) // columns
{
// Clamp filter to the image border
int h = min(max(r + i, 0), height);
int w = min(max(c + j, 0), width);
// Blur is a product of current pixel value and weight of that pixel.
// Remember that sum of all weights equals to 1, so we are averaging sum of all pixels by their weight.
int idx = w + cols * h; // current pixel index
float pixel = static_cast<float>(inputChannel[idx]);
idx = (i + half) * filterWidth + j + half;
float weight = filterWeight[idx];
blur += pixel * weight;
}
}
blurredChannel[c + r * cols] = static_cast<unsigned short>(blur);
}
void createFilter(float *gKernel,double sigma,int radius)
{
double r, s = 2.0 * sigma * sigma;
// sum is for normalization
double sum = 0.0;
// generate 9*9 kernel
int m=0;
for (int x = -radius; x <= radius; x++)
{
for(int y = -radius; y <= radius; y++)
{
r = std::sqrtf(x*x + y*y);
gKernel[m] = (exp(-(r*r)/s))/(3.14 * s);
sum += gKernel[m];
m++;
}
}
m=0;
// normalize the Kernel
for(int i = 0; i < (radius*2 +1); ++i)
for(int j = 0; j < (radius*2 +1); ++j)
gKernel[m++] /= sum;
}
int main()
{
cudaError_t cudaStatus;
const int size =81;
float gKernel[size];
float *dev_p=0;
cudaStatus = cudaMalloc((void**)&dev_p, size * sizeof(float));
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
}
createFilter(gKernel,20.0,4);
cudaStatus = cudaMemcpy(dev_p, gKernel, size* sizeof(float), cudaMemcpyHostToDevice);
if (cudaStatus != cudaSuccess) {
fprintf(stderr, "cudaMemcpy failed!");
}
/* i read image Buffere in unsigned short that code is not added here ,becouse it is large , and copy image data of buffere from host to device*/
/* So, suppose i have unsigned short *d_img which contain image data */
cudaMalloc( (void**)&d_img, length* sizeof(unsigned short));
cudaMalloc( (void**)&d_blur_img, length* sizeof(unsigned short));
static const int BLOCK_WIDTH = 32;
int image_width=1580.0,image_height=1050.0;
int x = static_cast<int>(ceilf(static_cast<float>(image_width) / BLOCK_WIDTH));
int y = static_cast<int>(ceilf(static_cast<float>((image_height) ) / BLOCK_WIDTH));
const dim3 grid (x, y, 1); // number of blocks
const dim3 block(BLOCK_WIDTH, BLOCK_WIDTH, 1);
gaussian_blur<<<grid,block>>>(d_blur_img,d_img,1050.0,1580.0,dev_p,9.0);
cudaDeviceSynchronize();
/* after bluring image i will copied buffer from Device to Host and free gpu memory */
cudaFree(d_img);
cudaFree(d_blur_img);
cudaFree(dev_p);
return 0;
}
Short answer: both algorithms are good with respect to image blurring, so feel free to pick the best (fastest) one for your use case.
Kernel size and sigma value are directly correlated: the greater the sigma, the larger the kernel (and thus the more operations-per-pixel to get the final result).
If you implemented a naive convolution, then you should try a separable convolution implementation instead; it will reduce the computation time by an order of magnitude already.
Now some more insight: they implement almost the same Gaussian blurring operation. Why almost ? It's because taking the FFT of an image does implicitly periodize it. Hence, at the border of the image, the convolution kernel sees an image that has been wrapped around its edge. This is called circular convolution (because of the wrapping). On the other hand, Gaussian blur implements a simple linear convolution.
Lets say I have 100 one-colored A4 sheets of paper, that are cut into different shapes and figures (2D), scanned, saved as an image file, and then needs to be sorted in ascending order of area.
Is there an effective way to find the area of the figures and arrange them?
If all pictures have the same size and all shapes the same color (that´s the situation if I don´t missunderstand your question), you can calculate the average color value.
The nearer the calculated color comes to the figures´s color, the bigger is the shape on the Image.
Some code:
private Color GetAverageImageColor(Image img)
{
double[] rgb = new double[3];
Color col;
Bitmap bmp = new Bitmap(img);
for(int y = 0; y < bmp.Size.Height; y++)
{
for(int x = 0; x < bmp.Size.Width; x++)
{
col = bmp.GetPixel(x, y);
rgb[0] += col.R;
rgb[1] += col.G;
rgb[2] += col.B;
}
}
for (int i = 0; i < 3; i++)
{
rgb[i] /= (bmp.Size.Height * bmp.Size.Width);
rgb[i] = Math.Round(rgb[i]);
}
return Color.FromArgb((int) rgb[0], (int) rgb[1], (int) rgb[2]);
}
first post here, and probably an easy one.
I've got the code from Processing's reference site:
float a = 0.0;
float inc = TWO_PI/25.0;
for(int i=0; i<100; i=i+4) {
line(i, 50, i, 50+sin(a)*40.0);
a = a + inc;
}
http://processing.org/reference/sin_.html
However, what I need is a line that follows the curve of a Sin wave, not lines representing points along the curve and ending at the 0 axis. So basically I need to draw an "S" shape with a sin wave equation.
Can someone run me through how to do this?
Thank you in advance,
-Askee
To draw a curve you need to store the previous point's position.
float a = 0.0;
float inc = TWO_PI/25.0;
float prev_x = 0, prev_y = 50, x, y;
for(int i=0; i<100; i=i+4) {
x = i;
y = 50 + sin(a) * 40.0;
line(prev_x, prev_y, x, y);
prev_x = x;
prev_y = y;
a = a + inc;
}