This study investigated the use of infrared thermography (IRT) to identify the dark, firm, and dry (DFD) phenomenon in Brazilian beef, which is a significant concern for the industry because of its inferior quality and reduced shelf life. This study examined 113 Nellore bulls and analyzed their minimum and maximum ocular temperatures using IRT. The results highlight the efficacy of thermal images (IRTmax) as a significant predictor, with R values ranging from 0.84 to 0.88 for calibration models. The inclusion of parameters such as glucose and lactate further enhanced prediction accuracy. The models also revealed that the combination of features, such as lightness (L*), redness (a*), and yellowness (b*), contributed to the precise prediction of pHu, with an R of 0.88. In model validation, RMSEP ranged from 0.104 to 0.158, indicating good generalization capability. The RPD, ranging from 1.7 to 2.6, suggests satisfactory quantitative prediction. The statistical significance of all models, evidenced by values <0.001, strengthens the reliability of the results. In conclusion, the models support the use of IRT as a tool for identifying pHu alterations in carcasses. When combined with blood parameters, they may exhibit even greater efficiency in predicting pHu in Nelore cattle carcasses, highlighting the potential applicability of these methods in the beef industry.

中文翻译：

---

非侵入性肉类质量评估：探索眼部红外热成像技术预测内洛尔肉牛最终 pH 值的潜力

这项研究调查了使用红外热成像 (IRT) 来识别巴西牛肉中的深色、坚硬和干燥 (DFD) 现象，这种现象因其质量低劣和保质期缩短而受到业界的严重关注。这项研究检查了 113 头内洛尔公牛，并使用 IRT 分析了它们的最低和最高眼温。结果凸显了热图像 (IRTmax) 作为重要预测因子的功效，校准模型的 R 值范围为 0.84 至 0.88。葡萄糖和乳酸等参数的包含进一步提高了预测准确性。模型还显示，亮度 (L*)、红度 (a*) 和黄度 (b*) 等特征的组合有助于精确预测 pHu，R 为 0.88。在模型验证中，RMSEP范围为0.104至0.158，表明具有良好的泛化能力。RPD 范围为 1.7 至 2.6，表明定量预测令人满意。所有模型的统计显着性（由值 <0.001 证明）增强了结果的可靠性。总之，这些模型支持使用 IRT 作为识别屠体 pHu 变化的工具。当与血液参数相结合时，它们在预测 Nelore 牛尸体中的 pHu 方面可能会表现出更高的效率，这凸显了这些方法在牛肉行业中的潜在适用性。