摘要
玉米虫害图像采集时存在的二维运动模糊导致采集的虫害图片质量降低,严重影响了机器对玉米病虫种类的自动识别。本文以二维运动模糊图像为研究对象,利用Radon变换和自相关函数计算,确定出模糊方向和模糊尺度,得到相应的点扩展函数PSF(Point spread function),再选用Lucy-Richardson方法复原图像。从实验结果主观和客观的评价上看,经过确定PSF函数复原处理后的图像细节更加丰富,清晰度更高,为后续玉米虫害的特征提取、目标识别奠定扎实的基础。
The two-dimensional motion blurring in the acquisition of corn pests picture results in quality degradation of insect pests picture,seriously affecting automatic identification of maize pest species by machine. The research focuses on the two-dimensional motion blurred image. Uses Radon transform and auto-correlation function to calculate,and determines the blur direction and the blur scale to obtain the corresponding function PSF（ Point spread function）. And finally we can use Lucy-Richardson method for image restoration. From the perspective of the experimental results in subjective and objective evaluation,the image detail after the dispose of the PSF function is more abundant,more clear,It is concluded that the restoration method has a good effect on the restoration of two-dimensional motion blurred images,and lays a solid foundation for the following features of the extraction and recognition of the pests.
作者
张政云
焦俊
孟珠李
李郑涛
高雅
ZHANG Zhengyun JIAO Jun MENG Zhuli LI Zhengtao GAO Ya(College of Information and Computer, Anhui Agricultural University, Hefei 230036, Chin)
出处
《内蒙古农业大学学报（自然科学版）》
CAS
2017年第4期60-68,共9页
Journal of Inner Mongolia Agricultural University(Natural Science Edition)
基金
安徽省教育厅质量工程项目K4A074
2014jyxm091
2014tszy090资助
