Support Website Contact Support Blog

Does Image Calibration means Aerial Triangulation?


I am MSc student, I have employed Pix4D during my thesis for image processing, I would like to ask about image calibration:

Does the calibration means Aerial Triangulation?

Does image calibration includes exterior orientation?

Does the Bundle Block Adjustment performs after camera optimization?


Traditionally the term image or camera calibration is related to calibrating the internals of a camera in a lab. Aerial triangulation is then very often the optimization of the externals of a project only.

Since we are dealing with cameras that are usually not pre calibrated in a lab and that are not so stable in their internals (temperature, shock, non-fixed focal length), we use the therm image calibration as a combination of internal and external parameter estimation. This includes Aerial Triangulation and Bundle Block adjustment.


Thanks a lot

Dear still I need the following informations:

  1. Does the exterior orientation parameters are computed after AAT or after BBA?
  2. Does the camera optimization performs after BBA?
  3. After performing Reoptimize, does the Automatic Tie Points has rebuilt again and verified for matching?
  4. What is different between generated sparse point cloud and point cloud densification in Step 2? 
  5. Does the Image matching process employs only  Automatic Tie Points? 



Hello Rojgar,

  1. The parameters of the exterior orientation are computed after both AAT and BBA. After the AAT the 3D position and the orientation of the cameras are computed, as well as the 3D coordinates of the Automatic Tie Points in an arbitrary coordinate system. The outputs of the AAT are used as inputs for the BBA, which optimizes them using the GCPs and the camera parameters. The final exterior orientation parameters are the ones computed after the BBA.

  2. The internal camera parameters (focal length, principal point coordinates, distortions) are optimized during BBA. 

  3. The Reoptimize option, reoptimizes the camera positions and the internal camera parameters.The ATPs positions are also reoptimized during this process.

  4. The sparse point cloud is the result of the BBA, while the densified point cloud is the output of the densification algorithm that takes as an input the sparse point cloud (Automatic Tie Points).

  5. The image matching process uses the keypoints, that are the characteristic points on the images, to find pairs and compute the matches. 


Dear Antigoni, 

I have some doubts regarding the role of the internal camera parameters during the process. I’m not sure whether they are optimized during the BBA or during the AAT. If I understood correctly, in your comment you wrote that the outputs of the AAT are optimized in the BBA with the help of the GCPs and the camera parameters, which ones are you refering to, the external or the internal camera parameters?


Thank you so much


Hi Luis, 

From my understanding the optimization procedure of both the internal and external camera parameters is iterative. This means that it starts from an initial estimate and gets improved during the AAT and then during the BBA. Each time using more information (such as GCPs,…) to make the estimates more accurate.