一种无需匹配的多目重构方法

针对立体匹配算法具有二义性,对遮挡、光照条件缺乏鲁棒性的问题,提出了一种两步多目重构新方法.首先,基于视觉外型的第二基本属性把立体匹配问题转化为颜色方差最小化问题,使用方差阈值来平衡采样点数与采样噪声之间的矛盾.然后,结合运动曲面的位置势能与弹性势能建立目标泛函,再采用二阶中心差分与Hamilton-Jacobi方程的迎风设计技术极小化该目标泛函,促使水平集函数逐步逼近目标表面,同时消除采样噪声,最终实现目标表面完全重构.新方法不仅避免了立体匹配算法固有的二义性问题,而且增强了对遮挡、复杂光照条件的鲁棒性.实验结果表明,新方法能够从实际复杂图像集推断出目标形状,与Hoppe的经典算法比较,表现出更好的准确性,且节省了26%~33%的重构时间.

采用窄带图切割的多目重构方法

针对目前多目重构研究中物体表面获取和凹陷区域恢复的难题,提出采用窄带图切割的可见外壳和多目立体匹配相融合的方法.在可见外壳的拓扑约束下,通过最小化能量泛函使得重构表面与多目立体匹配得到的稠密点云形状一致;并利用窄带图切割计算能量泛函的全局最优解.实验结果表明,该方法能重构封闭表面和凹陷区域;窄带图切割在获得最优解的同时提高了计算效率,与全图切割相比速度提高了3倍以上.

Multi-objective planning model for simultaneous reconfiguration of power distribution network and allocation of renewable energy resources and capacitors with considering uncertainties

This research develops a comprehensive method to solve a combinatorial problem consisting of distribution system reconfiguration, capacitor allocation, and renewable energy resources sizing and siting simultaneously and to improve power system's accountability and system performance parameters. Due to finding solution which is closer to realistic characteristics, load forecasting, market price errors and the uncertainties related to the variable output power of wind based DG units are put in consideration. This work employs NSGA-II accompanied by the fuzzy set theory to solve the aforementioned multi-objective problem. The proposed scheme finally leads to a solution with a minimum voltage deviation, a maximum voltage stability, lower amount of pollutant and lower cost. The cost includes the installation costs of new equipment, reconfiguration costs, power loss cost, reliability cost, cost of energy purchased from power market, upgrade costs of lines and operation and maintenance costs of DGs. Therefore, the proposed methodology improves power quality, reliability and security in lower costs besides its preserve, with the operational indices of power distribution networks in acceptable level. To validate the proposed methodology's usefulness, it was applied on the IEEE 33-bus distribution system then the outcomes were compared with initial configuration.

复杂X光照条件下非凸表面完全重构

针对复杂光照条件下非凸表面完全重构问题,提出了一种基于两步策略的多目重构新方法。把立体匹配过程转化为颜色方差最小化问题,利用颜色方差阈值来平衡采样点数与采样噪声之间的矛盾,同时回避了立体匹配算法的二义性。采用水平集方法建立曲面运动方程,通过符号距离函数的演化过程逼近目标表面;通过弹性势能的最小化过程滤除采样噪声,增强对遮挡、光照条件的稳键性。实验结果表明,该方法能够在复杂光照条件下推断出非凸目标形状,且与Jin的算法比较,表现了更好的准确性,节省31%～39%的重构时间。