Novel method for online dense reconstruction by rotation-translation decoupling optimization
To address the optimization instability caused by the rotation-translation-coupling-optimization in traditional algo-rithms,this paper proposed a novel dense reconstruction algorithm based on rotation-translation decoupling optimization.The algorithm divided the camera pose space into independent rotation and translation subspaces and independently searches for the optimal solution within each subspace.During each iteration,the algorithm set search neighborhoods for the estimated values of rotation and translation,samples candidate solutions within these neighborhoods,and selected the optimal solutions through evaluation to update the estimates.Iterate this process until the termination condition was met.Experimental results verify that the algorithm effectively reduces interference between rotation and translation during the optimization process,thereby enhan-cing the efficiency and accuracy of the optimization.It demonstrates that the algorithm has certain advantages in the camera pose estimation stage of dense reconstruction.
万方数据
维普
