This paper studies the restoration of fragmented 2D digital pictures. Automatic reassembly of image fragments can facilitate many tasks in archaeology, forensics, and digital media. We propose an integrated algorithm for this important yet challenging problem. Our algorithm first determines the adjacency relationship among fragments, then effectively matches subpieces and refines the reassembly iteratively.

An experimental study was performed to investigate the resistance characteristics of the trilobal throttling element in narrow annular channel, while corresponding flow fields analyzed using Fluent software with RNG k-ε turbulence model based on the structural grid for the trilobal throttling element with different placket angle. A relationship between the packet angle θ and the local resistance coefficient ξ is obtained. Results show that the local resistance coefficient obtained by numerical simulation of the trilobal throttling element in narrow annular channel have a good agreement with the experimental data, and their relative error is less than 8%. This also fully proves that the RNG k-ε turbulence model is suitable for the simulation of the trilobal throttling element in the narrow annular channel. The local resistance coefficient of the trilobal throttling element in narrow annular channel increases with rising packet angle θ, and the relationship is similar to the power function change.