Li Tao, Tong Jinjie, Liu Muhua, Yao Mingyin, Xiao Zhifeng, Li Chengjie
College of Engineering, Jiangxi Agricultural University, Nanchang 330045, China.
Jiangxi Institute of Science and Technology Information, Nanchang 330000, China.
Foods. 2022 Dec 11;11(24):4009. doi: 10.3390/foods11244009.
Online detection of impurities content in the corn deep-bed drying process is the key technology to ensure stable operation and to provide data support for self-adapting control of drying equipment. In this study, an automatic approach to corn image acquisition, impurity classification and recognition, and impurities content detection based on machine vision technology are proposed. The multi-scale retinex with colour restore (MSRCR) algorithm is utilized to enhance the original image for eliminating the influence of noise. HSV (Hue, saturation, value) colour space parameter threshold is set for image segmentation, and the classification and recognition results are obtained combined with the morphological operation. The comprehensive evaluation index is adopted to quantitatively evaluate the test results. Online detection results show that the comprehensive evaluation index of broken corncobs, broken bracts, and crushed stones are 83.05%, 83.87%, and 87.43%, respectively. The proposed algorithm can quickly and effectively identify the impurities in corn images, providing technical support and a theoretical basis for monitoring impurities content in the corn deep-bed drying process.
在线检测玉米深床干燥过程中的杂质含量是确保干燥设备稳定运行并为其自适应控制提供数据支持的关键技术。本研究提出了一种基于机器视觉技术的玉米图像自动采集、杂质分类识别及杂质含量检测方法。利用带颜色恢复的多尺度视网膜算法（MSRCR）对原始图像进行增强，以消除噪声影响。设置HSV（色调、饱和度、明度）颜色空间参数阈值进行图像分割，并结合形态学运算得到分类识别结果。采用综合评价指标对检测结果进行定量评估。在线检测结果表明，破碎玉米棒、破碎苞叶和碎石的综合评价指标分别为83.05%、83.87%和87.43%。所提算法能够快速有效地识别玉米图像中的杂质，为玉米深床干燥过程中杂质含量监测提供技术支持和理论依据。
