典型猪舍光环境下机器视觉图像量化评价及筛选

发布时间：2019-07-01 08:36

【摘要】：随着数字化农业进程的发展,利用机器视觉技术进行生猪养殖方面研究已成为许多研究人员关注的热点之一。机器视觉中外界光环境对采集图像的质量具有较大影响,但是较多猪体质量监测方面研究的焦点集中在图像算法方面,忽视外界光环境的重要性,缺乏量化评价机器视觉光环境优劣的方法。该文基于猪体质量检测平台,利用LabVIEW软件编程分析实际光环境条件下拍摄图像质量。经现场试验分析,理想图像、曝光过度图像、自然光图像、阴阳图像4类图像在曝光参数和照度均匀度参数有明显区别:理想图像和曝光过度图像的照度均匀度参数均满足照明工程标准要求,理想图像曝光正常比率较高;自然光图像和阴影图像的照度均匀度参数不满足照明工程标准要求,且曝光正常率偏低。该文采用曝光正常,最小照度与最大照度比值U10.7和最小照度与平均照度比值U20.8判断实现理想图像筛选,便于研究者前期图像预处理工作。
[Abstract]:With the development of digital agriculture, the use of machine vision technology in pig breeding has become one of the hotspots of many researchers. The boundary light environment of machine vision has a great influence on the quality of image acquisition, but more pig body quality monitoring focuses on image algorithm, neglects the importance of external light environment, and lacks a quantitative method to evaluate the quality of machine vision light environment. In this paper, based on the pig body mass detection platform, LabVIEW software is used to program and analyze the image quality under the actual light environment. Through field test analysis, the ideal image, overexposed image, natural light image and yin and yang image are obviously different in exposure parameters and illuminance uniformity parameters: the illuminance uniformity parameters of ideal image and overexposed image meet the requirements of lighting engineering standard, and the ideal image exposure normal ratio is high; The illuminance uniformity parameters of natural light image and shadow image do not meet the requirements of lighting engineering standard, and the normal exposure rate is on the low side. In this paper, normal exposure, the ratio of minimum illuminance to maximum illuminance U10.7 and the ratio of minimum illuminance to average illuminance U20.8 are used to determine the ideal image screening, which is convenient for researchers to preprocess the image in the early stage.
【作者单位】：中国农业大学水利与土木工程学院农业部设施农业工程重点实验室;
【基金】：“十二五”国家科技支撑项目(2014BAD08B05)
【分类号】：TP391.41

【相似文献】