我国拥有丰富的地方鸡遗传资源，不同品种的地方鸡在营养、口感和风味上表现出显著区别。该研究利用近红外高光谱成像技术研究地方鸡品种的鉴别方法。以龙游麻鸡、江山乌骨鸡和丝羽乌骨鸡为研究对象，采集3个品种鸡肉在900~1 700 nm波段的高光谱图像并提取光谱数据。对光谱数据进行预处理后选取特征波长，建立PLS和SVM鉴别模型。结果显示，Detrending-SNV为最优预处理方法，ACCc、ACCp和ACCt 分别为94.17%、95%和94.59%。基于MCUVE-SPA算法选取的16个特征波长建立的PLS鉴别模型为最优模型，ACCc、ACCp和ACCt分别为95.83%、100%和97.92%，与全光谱模型相比分别提高了1.76%、5.26%和3.52%。研究结果表明，利用近红外高光谱成像技术鉴别地方鸡品种具有一定的可行性。

China has abundant indigenous chicken breeds, and different breeds of indigenous chicken vary significantly in terms of nutrition, taste, and flavor. In this study, near-infrared hyperspectral imaging technology was used to study the discrimination method for indigenous chicken breeds. Longyou Partridge Chicken (LPC), Jiangshan Wugu Chicken (JWC) and Silkies Chicken (SC) were used as the research objects. Hyperspectral images of three chicken breeds in the band range of 900~1 700 nm were collected and spectral data were extracted. After preprocessing the spectral data, characteristic wavelengths were selected and PLS and SVM discrimination models were established. The results showed that the Detrending-SNV pre-treatment method was the optimal preprocessing method, with ACCc, ACCp and ACCt being 94.17%, 95% and 94.59%, respectively (which increased by 1.76%, 5.26% and 3.52%, respectively, compared with the full spectrum model). The research results showed that it is feasible to use near-infrared hyperspectral imaging technology to discriminate indigenous chicken breeds.

国家自然科学基金项目（31702153）；浙江农林大学创新训练项目（S202210341201）