白羽肉鸡肉品新鲜度多维快速测定方法的建立与有效性验证

Establishment and validity verification of a multidimensional rapid determination method for the freshness of white feather broiler meat

中文关键词:  鸡胸肉  新鲜度检测  理化指标  高光谱成像技术  偏最小二乘回归

英文关键词:chicken breast meat  freshness detection  physicochemical indicators  hyperspectral imaging technology  partial least squares regression

中文摘要:

      目的 为了满足白羽肉鸡肉品新鲜度实时、批量检测的企业现实需求, 探索多维快速测定肉品新鲜度并进行有效性验证。方法 本研究以白羽肉鸡胸肉为对象, 以新鲜度检测为核心, 在4 ℃冷藏条件下设置0 h至144 h不同贮藏时间梯度, 依据GB 5009.237—2016《食品安全国家标准 食品pH值的测定》和GB 5009.3—2016《食品安全国家标准 食品中水分的测定》分别对快速测定法测定的pH及水分含量指标进行验证, 分析其随新鲜度下降的变化规律; 同时利用高光谱成像技术采集样品的光谱数据, 采用Savitzky-Golay (S-G)平滑结合一阶导数法进行光谱预处理, 提取理化指标的特征波长, 基于偏最小二乘回归(partial least squares regression, PLSR)算法构建鸡胸肉新鲜度判别模型。结果 快速测定法与GB 5009.237—2016和GB 5009.3—2016测定pH和水分含量的结果均无显著差异(P>0.05), 其中pH、水分含量的P分别为0.8141、0.2586; 同时, 基于PLSR构建的模型对各项理化指标的预测性能良好, 预测集相关系数(R2p)为0.74~0.87, 预测均方根误差(predicted root mean square error, RMSEP)为0.12~1.80。结论 光谱预处理可有效降低仪器噪声和外界干扰并凸显特征吸收峰, 新鲜肉与腐败肉的特征波长光谱响应差异显著; 所构建的PLSR模型能够实现新鲜肉与腐败肉的准确鉴别。本研究建立的多指标协同检测体系, 为鸡胸肉品质安全监控提供了有效可靠的技术支持, 丰富了鸡肉新鲜度无损检测的理论与实践研究。

英文摘要:

      Objective To meet the practical industrial demand for real-time, batch detection of freshness in white feather broiler meat, this study explores a multidimensional rapid method for assessing meat freshness and validates its validity. Methods This study focuses on the breast meat of white-feathered broilers, with freshness assessment as its central theme. Under refrigerated storage at 4 ℃, a gradient of storage durations from 0 to 144 hours was established. The pH and moisture content values obtained via rapid detection methods were validated in accordance with GB 5009.237—2016 National food safety standard-Determination of pH in foods and GB 5009.3—2016 National food safety standard-Determination of moisture in foods, respectively, to analyze their variation patterns as freshness deteriorates. Concurrently, hyperspectral imaging technology was employed to collect spectral data of the samples. Spectral preprocessing was conducted using Savitzky-Golay (S-G) smoothing combined with first-derivative transformation to extract characteristic wavelengths related to physicochemical indices. Subsequently, a freshness discriminant model for chicken breast meat was constructed based on the partial least squares regression (PLSR) algorithm. Results The results of the rapid determination method showed no significant difference from those determined by GB 5009.237—2016 and GB 5009.3—2016 for pH and moisture content (P>0.05). Among them, the P values of pH and moisture content were 0.8141 and 0.2586 respectively. At the same time, the model constructed based on PLSR had good predictive performance for various physicochemical indicators, with the correlation coefficient (R2p) of the prediction set ranging from 0.74 to 0.87, and the predicted root mean square error (RMSEP) ranging from 0.12 to 1.80. Conclusion Spectral pre-processing can effectively reduce instrument noise and external interference and highlight characteristic absorption peaks. There is a significant difference in the characteristic wavelength spectral responses between fresh meat and spoiled meat. The constructed PLSR model can achieve accurate discrimination between fresh meat and spoiled meat. The multi-index collaborative detection system established in this study provides effective and reliable technical support for the monitoring of chicken breast meat quality and safety, and enriches the theoretical and practical research on non-destructive detection of chicken freshness.