The heterogeneity of apple fruit was highlighted by near-infrared hyperspectral imaging (NIR-HSI) using a data analysis in two successive steps. First, NIR-HSI images were acquired on the cut surface of six transverse slices per apple, which were then systematically sampled with 5 or 6 cylinders per slice. PCA carried out on the NIR-HSI images allowed to select 141 representative cylinders from the total dataset (1056 samples), in which the contents of dry matter (DMC), total sugars (TSC), fructose, glucose, sucrose, malic acid and polyphenols were quantified by spectrophotometry and chromatography. In a second step, leave-one-out PLS models were developed and successfully used to describe the distribution of DMC (R_cv² = 0.83, RPD = 2.39) and TSC (R_cv² = 0.81, RPD = 2.20) in each apple slice. A strong heterogeneity of DMC and TSC was detected inside each fruit. Such a simple and rapid method reduced the needs of numerous chemical characterizations to demonstrate the distribution of quality traits within and between fruit and contributed to better manage the fruit quality measurements.

通过两个连续步骤中的数据分析，近红外高光谱成像（NIR-HSI）突出了苹果果实的异质性。首先，在每个苹果六个横切面的切面上获取NIR-HSI图像，然后用每个切面5或6个圆柱体对它们进行系统采样。在NIR-HSI图像上进行的PCA允许从总数据集中选择141个代表性圆柱体（1056个样本），其中干物质（DMC），总糖（TSC），果糖，葡萄糖，蔗糖，苹果酸和多酚通过分光光度法和色谱法定量。第二步，开发了留一法式PLS模型，并成功地用于描述DMC（R _cv ²  = 0.83，RPD = 2.39）和TSC（R _cv ² = 0.81，RPD = 2.20）。在每个水果中都检测到DMC和TSC的强烈异质性。这种简单，快速的方法减少了许多化学表征的需求，以证明水果内部和水果之间的质量性状分布，有助于更好地管理水果质量测量。