摘要
用多数表决电路解决图像处理中孤立噪声处理是比较直接而且可靠的方法 ,但是 ,随着变量数的增多 ,电路的规模急剧增大 (呈指数规律 ) ,而少变量的表决电路 (变量数小于 4 )在孤立噪声处理中又没有实用价值 ,为解决该问题 ,提出了一种新型神经元MOS多数表决电路 ,它更适用于变量数大于 4的情况。与传统的CMOS方案相比 ,该电路具有速度快、结构简单、功耗低等特点。T SPICE仿真结果证明该设计的可行性。
It is a simple and credible method using multivariable vote circuit to solve the problem of isolating noise on image processing. With the increasing of variables, the scale of the circuit enlargeds rapidly (on exponential formula). Vote circuit, with variables less than 4. has no valuc for using in isolating noise processing. For this reason, a new voting circuit on neuron MOS is researched, suit for the case of variables greater than 4. Comparing with the normal CMOS method, this method has the characteristic of fast speed、simple structure and low power consumption etc. The result of T-SPICE simulation shows the practicability of this design.
出处
《电子器件》
CAS
2004年第2期217-219,227,共4页
Chinese Journal of Electron Devices
基金
国家自然科学基金资助项目 (6 0 2 76 0 17) .
