use bytes_per_pixel instead of hardcoded value

Author: floe

examples/protonect/Protonect.cpp

@@ -90,7 +90,7 @@ int main(int argc, char *argv[])
     cv::imshow("ir", cv::Mat(ir->height, ir->width, CV_32FC1, ir->data) / 20000.0f);
     cv::imshow("depth", cv::Mat(depth->height, depth->width, CV_32FC1, depth->data) / 4500.0f);
 
-    uint8_t registered[depth->height*depth->width*3];
+    uint8_t registered[depth->height*depth->width*rgb->bytes_per_pixel];
     registration.apply(rgb,depth,registered);
     cv::imshow("registered", cv::Mat(depth->height, depth->width, CV_8UC3, registered));
 

examples/protonect/src/registration.cpp

@@ -100,7 +100,7 @@ void Registration::apply(Frame* rgb, Frame* depth, unsigned char* registered)
     for (int y = 0; y < depth->height; y++) {
 
       d_off = y*depth->width + x;
-      r_off = d_off*3;
+      r_off = d_off*rgb->bytes_per_pixel;
 
       float z_raw = depth_raw[d_off];
       if (z_raw == 0.0) {
@@ -113,8 +113,8 @@ void Registration::apply(Frame* rgb, Frame* depth, unsigned char* registered)
       apply(x,y,z_raw,cx,cy);
 
       // is rounding the right way to go here - no subpixel precision?
-      c_off = (round(cx) + round(cy) * rgb->width) * 3;
-      if ((c_off < 0) || (c_off > rgb->width*rgb->height*3)) continue;
+      c_off = (round(cx) + round(cy) * rgb->width) * rgb->bytes_per_pixel;
+      if ((c_off < 0) || (c_off > rgb->width*rgb->height*rgb->bytes_per_pixel)) continue;
 
       registered[r_off+0] = rgb->data[c_off+0];
       registered[r_off+1] = rgb->data[c_off+1];