基于图像增强的高湿度环境下野生二粒小麦籽粒硬度检测方法

为有效检测高湿度环境下谷类籽粒硬度,以某谷类食品加工厂生产的野生二粒小麦为试验对象,采用高光谱图像成像系统获取野生二粒小麦高光谱图像,通过基于匹配思想的自适应消噪方法(PLS)去除野生二粒小麦高光谱图像的带状噪声,增强高湿度环境下野生二粒小麦高光谱图像质量,在此基础上,将野生二粒小麦光谱数据的平均值作为光谱数据,构建高湿度环境下野生二粒小麦籽粒硬度预测模型,实现对高湿度环境下谷类籽粒硬度的准确检测。仿真试验结果表明,当加工环境湿度为55%时,本文方法检测野生二粒小麦籽粒硬度值平均结果为1745 g,与标准红外检测方法得到的结果差值仅有3 g;当环境湿度提高到75%时,本文方法检测结果为1712 g,与标准红外检测方法相差42 g,本文方法检测野生二粒小麦籽粒硬度结果精度高,优于声振频带幅值特性法,是一种高精度的高湿度环境谷类籽粒硬度检测方法。

Testing Method of Wild Emmer Wheat Hardness in High Humidity Environment Based on Image Enhancement

In order to effectively detect the grain hardness in high humidity environment, the hyperspectral image of wild two-grain wheat produced by a cereal processing factory was obtained by hyperspectral image system. The adaptive denoising method (PLS) based on matching idea was used to remove the banded noise of hyperspectral image of wild two-grain wheat and enhance the quality of hyperspectral image of wild two-grain wheat in high humidity environment. On this basis, the average value of spectral data of wild two-grain wheat was taken as spectral data, and the prediction model of grain hardness of wild two-grain wheat in high humidity environment was constructed to realize the accurate detection of grain hardness in high humidity environment. The simulation results showed that when the humidity of processing environment was 55%, the average value of grain hardness of wild two-grain wheat detected by the proposed method was 1745 g, and the difference between the proposed method and the standard infrared detection method was only 3 g. When the ambient humidity was increased to 75%, the result of the proposed method was 1712 g, which is 42 g different from the standard infrared detection method. The proposed method had a high precision in detecting grain hardness of wild two-grain wheat, which is better than the acoustic vibration frequency band amplitude characteristic method.