| 基于图像的化工厂重度污染等级判定研究与仿真 |
| |
| 引用本文： | 王海顺,吴华.基于图像的化工厂重度污染等级判定研究与仿真[J].科技通报,2012,28(7):98-101. |
| |
| 作者姓名： | 王海顺 吴华 |
| |
| 作者单位： | 安阳师范学院,河南安阳,455002 |
| |
| 摘 要： | 在不同化工厂检测环境中,空气中的环境变化情况较为复杂,带有颜色特征的污染气体浓度会被迅速稀释,造成转化的像素特征强度衰减。传统算法多是基于采集到的某种像素特征强度进行污染等级的判断,一旦气体被稀释,颜色特征发生退化,检测准确率会降低。提出了一种基于像素支持向量机增量学习算法。通过灰度差分的支持向量机增量学习。建立对不同像素等级信号进行对应增强学习,克服像素衰退的弊端。实验证明,这种算法能够避免由于化工厂内气体大量扩散,造成的像素衰减的缺陷,提高了化工污染程度检测的准确率。
|
| 关 键 词： | 化工厂污染 支持向量机 像素衰退 |
Based on Image Chemical Plant Severe Pollution Level Determine Research and Simulation |
| |
| WANF Haishun
,
WU Hua.Based on Image Chemical Plant Severe Pollution Level Determine Research and Simulation[J].Bulletin of Science and Technology,2012,28(7):98-101. |
| |
| Authors: | WANF Haishun WU Hua |
| |
| Institution: | 2(Anyang Normal University,Anyang City,Henan 455000,China) |
| |
| Abstract: | In different chemical plant test environment,environmental changes in the air condition is relatively complex,with color characteristics of the concentration of the gas pollution will be diluted rapidly,causing the transformation of the pixels feature strength attenuation.More than traditional algorithm is based on the collected a pixels intensity on the level of the pollution feature judgment,once the air is diluted,color features occur degradation,detection accuracy can be decreased.Puts forward a pixel based on support vector machine(SVM) incremental learning algorithm.Through the gray difference of support vector machine(SVM) incremental learning.To set up different pixel level signal corresponding improving learning,overcome the disadvantages of pixels recession.Experiments show that the algorithm can avoid the chemical plant gas inside the large diffusion,the attenuation of the defects caused by pixels,improve the chemical pollution degree detection accuracy. |
| |
| Keywords: | chemical plant pollution support vector machine(SVM) pixel recession |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
