大规模秸秆燃烧充分程度智能图像识别仿真

Simulation of Intelligent Image Recognition for Large-Scale Straw Burning Full Degree

对大面积秸秆燃烧过程进行图像监控,有利于大气保护。当前的大区域秸秆燃烧过程中,燃烧是否充分通过烟雾颜色进行判断。一旦秸秆燃烧区域较大,不同区域的烟雾扩散过程及其迅速,采集烟雾灰度特征的衰减过程不可控,传统的智能图像识别方法判定阀值会迅速失效,很难通过烟雾像素特征判断燃烧是否充分。提出基于人工鱼群算法的大规模秸秆燃烧充分程度识别方法。采集大规模秸秆燃烧图像,根据小波变换相关理论,对燃烧图像进行增强处理,提高图像质量。根据人工鱼群算法处理图像细节的优势,引入秸秆燃烧像素灰度分布密度函数,得到离散灰度矩阵,利用上述算法处理像素细节的优势,对大规模秸秆燃烧区域的连接处进行分割,可充分提高识别精度。实验结果表明,利用改进算法进行大规模秸秆燃烧充分程度识别,能够极大的提高识别的准确性。

Based on artificial fish algorithm, a large-scale straw burning full extent recognition method is presented. Firstly, burning images of large-scale straw are collected, and the images are processed to improve their quality based on the wavelet transform thoery. According to the advantage of artificial fish algorithm to process details of image, the straw burning pixel distribution density function is introduced and the discrete grey matrix is obtained. And then, by using the advantage of the algorithm processing pixel details, the joint segmentation of burning area of large -scale straw is obtained. Finally, an improved recognition algorithm is achieved. The experimental results show that the improved algorithm can greatly improve the accuracy of recognition.