I am trying to build my custom Photosphere viewer to run using SDL2 and a custom IMU I purchased. So far, I have managed to read IMU values, open the .jpg and display it using SDL2.
My issue is how to make sense IMU data to read parts of the jpg appropriately. Basically, I do not want to display the whole jpg but just parts of it based on IMU data (I receive Euler angles or Quaternions). Right now, I am just using a single mono photosphere (I am not concerned with stereo yet), which is stored as a equirectangular projection, and I need to use the IMU to get it to a polar projection (I believe?)
I am not sure how to index the jpg based on IMU data to create a working photosphere viewer and I cannot seem to find a good explanation of how to address the jpg. Can anyone point me into the right direction? Thanks!
I was able to find a really great OpenGL based simple Python photosphere viewer here. I just then needed to create a rotation matrix from the sensor IMU. There are good tutorials to convert from Quaternion to Matrix like this one.
Related
I would like to convert a 3d image (hyperspectral cube) into a 2d one with 3 channels.
I have all images in the source and target – paired.
What should be changed in the code in order to support this?
(see good code with explanation here (Jason Brownlee):
https://machinelearningmastery.com/how-to-develop-a-pix2pix-gan-for-image-to-image-translation)
Thanks,
Eli
I am working on 3D reconstruction with tango. Our system is quite similar to KinectFusion, which uses voxel representation, but use Tango as tracker. Left image (in video linked below) is rendered by Raycast at current pose (given by tango) in real time. Raw pose converted by GetOC2OWMat() as in code examples, in addition sign of tx and rx are flipped to cope with our system. Everything works fine except ration in Z axis, which changes angle in rendered image. I guess coordinate system conversion is not done properly, but depth integration is working if no Z rotation is involved. I have also checked det(R) is always 1.
Video
It sounds like you are not factoring in intrinsics - have you accounted for camera and device IMU frames ? You need these to fully re-establish original viewpoint, i.e. both camera and device imu frame matrices need to be multiplied in to your stack
Sorry that I just find the place where things goes wrong. When the image is displayed with opengl, the rendered gl size does not have same aspect ratio as Raycasting image.
Do you program with Java/C/Unity? I'm curious because my device has problems with the camera data and you seem to capture it without problems. I am quite sure it's a bug but I would like to make sure it really is one.
I am working with disparity maps (1024 x 768) obtained via stereo and I am able to get point clouds with XYZRGB pcl::Points. However not all pixels from the disparity map are valid depth hence there will never be 1024x768 = 786432 XYZRGB points. Fortunately I am able to save the point clouds unorganized (i.e. height=1). Unfortunately, some normal estimation methods etc, are tailored for organized pointclouds. How can I create organised pointclouds from this ?
I believe that this is not possible.
First of all unorganized point cloud (PC) is just list of points in random order written in file
On the other hand organized PC carries information of in which order orginal points were obtained by depth camera and some other information. This information is stored in lets call it grid.
Once you destroy this grid omiting some points theres no algorithm that can put it back together as it originally was
You can use other methods which provides PCL that doesnt take OPC as an argument. Result will be same as if you would use organized point cloud only little bit slower (depends on size of your input cloud)
I assume that you do have the calibration parameters that are necessary to transform the image points and their depth into 3D points, right?
In this case, you simply create a 2D point cloud and do the following for each pixel of the disparity map:
If the point is valid:
set the corresponding point in the point cloud to the 3D point
else:
set the corresponding point in the cloud to NaN (i.e. a 3D point with NaN as coordinates)
I am completely new to 3D and started with Jeff Lamarche's tutorials as an introduction to openGL ES for iPhone, then so far, I am able to draw a spinning sphere, which will the base of my application.
What I want to do is render a planet Earth, thanks to 2D GIS vector data (polygones, lines or points with latitude/longitude or x/y coord).
I want to be able to turn different layers on/off and maybe able to identify an object that wold be touched.
My questions are :
would it be easier to rasterize my vector data to use them as image texture or apply the vector data onto the sphere (keeping in mind that I want to turn on/off the layers, the touch-enabled objects being optional)?
would it be easier to use a software like blender to draw the planet and add the layers rather than starting with the sphere I already have (procedural sphere)?
do the export tool from blender to opengl work well?
This kind of question is difficult to answer in general. Technically your intention sounds a lot like if you would like to write a program like Google Earth or KDE Marble. Since you're referring to GIS data you will require very high resolution. Textures only make sense for limited resolution data.
GIS applications usually work using hybrid approaches where some vector data are rendered directly (roads, waters, borders), while others are rendered to texture and the texture, or more accurately texture tiles, being used as caches, for example for building outlines in dense cities or the like. However data as it comes from say OSM can be directly rendered as vector data, since they are not very dense.
Do you know what would be the best approach to generate 3D output for one of these new "3D ready" televisions from software. Our application have some nice 3D visualizations, we want these to look nice.
Also, how feasible is it to generate it from a Flash (Flex) app.
I believe that the gaming and 3DTV industries have paved the way for you. As long as your app already outputs 3D visualizations, it may just be a matter of installing a driver. You can get started with this NVIDIA 3D Stereo User’s Guide, but I believe there's tons of other stuff out there if you look.
See also the answers to this question.
3D televisions can display 3D output only for images shot in 3D. This means "intended for simulated 3D," not just a two-dimensional projection of a 3D image.
Stereoscopy is produced by generating two completely separate images per frame (one for each eye) in which the foreground objects are offset to simulate a 3D image. You cannot take a 2D image and make it into a 3D image, the source frames must be produced as 3D frames from the beginning.
More information:
http://en.wikipedia.org/wiki/3D_television
http://en.wikipedia.org/wiki/Stereoscopy