315 Starch Degradation Analysis of Apple Fruits measured with a Hyperspectral (NIR) Imaging System A. Peirs, N. Scheerlinck, B.M. Nicola Laboratory of Postharvest Technology, Catholic University of Leuven, Belgium J. De Baerdemaeker Laboratory for Agro- Machinery and Processing, Catholic University of Leuven, Belgium Keywords: hyperspectral imaging, near infrared, starch, apple Abstract A hyperspectral near infrared imaging device was developed to measure the starch content of apple fruit during their maturation. A good relation was found between the score plots of the first PC after calculation of a PCA analysis of the hyperspectral images of three apples of different maturity. This was observed for 6 different cultivars. Further, the instrument indicated that internal disorders such as water core could also be detected. INTRODUCTION Because of the consumer demand, apple fruit is provided all year round. Modern postharvest techniques, such as controlled or modified atmosphere combined with low temperatures can extend the storage life up to nine months after harvest with minimal quality losses. To ensure high quality after long-term storage, it is essential that the fruits are harvested within a well-defined optimal harvest period. As the apple is a climacteric fruit this optimal picking period corresponds to the period at which the respiratory activity of the fruits is minimal -the pre-climacteric minimum- or slightly later. A key characteristic on which the optimal harvest date prediction is based is the almost complete conversion of insoluble starch into simple sugars. This phenomenon was well examined by Fan et al. (1995) who observed a decrease in total starch content, which was synthesised during growth and reached a maximum around the end of August (Berüter, 1989). The iodine test is a cheap and fast commercial test that provides an indication of the starch distribution although the starch concentration is not exactly measured. However, the disadvantage of this technique is its subjective character. The scope of this work was to develop an objective measurement technique to replace the traditional starch measurements. As it was already found before that reflectance spectroscopy can measure internal quality characteristics of fruit (Peirs et al., 2001) a hyperspectral imaging NIR-instrument was developed to measure the starch content. MATERIALS AND METHODS Apple Fruit Starch measurements were performed on a number of Jonagold, Boskoop, Braeburn, Elstar, Royal Gala and Golden Delicious picked at the orchard Velm (Belgium) in 2001. Construction of the Hyperspectral-Imaging Device Rather than collecting a single spectrum at one spot on a sample, an imaging spectrometer records a spectral volume, or a hyperspectral image, which contains a complete spectrum for every pixel in the image of the sample. The selected approach was to measure a sequence of lines of spatial information with a full spectral range per spatial pixel (Figure 1b) rather than to capture a sequence of full spatial scenes at each spectral band to form a three-dimensional image cube (Fig. 1a). The advantage of the former approach is that the interesting spectral regions do not need examination prior to the image analysis. Martinsen and Schaare (1998) examined the sugar distribution in kiwifruit in this way. The hyperspectral scanning system was developed as follows. Proc. Postharvest Unltd Eds. B.E. Verlinden et al. Acta Hort. 599, ISHS 2003