Wheat is one of the most important crop species with the highest production and consumption

worldwide. However, today's demand for wheat is increasing due to epidemics and wars. Since different

types of wheat have different characteristics, their usage areas also differ. Determining the most suitable

wheat type for the usage areas will also be beneficial in terms of pricing the wheat. Classification is needed

in order to determine a wheat type as durum or bread. Today, these classification processes are carried out

by experiments in laboratory environments, but these methods are time-consuming and require long

processes. Convolutional Neural Networks (CNN) deep learning-based approaches are used to make faster

and more precise classification. When other studies were examined, the fact that the CNN model gave high

accuracy results in classification studies contributed to the realization of this study.

The aim of this study; is the classification of reflectance values obtained by near infrared (NIR)

spectroscopy of different types of wheat using CNN model. Within the scope of this study, classification of

24 wheat types collected from certain regions of Turkey was carried out. Data sets were created with

reflectance values obtained from wheat using NIR Spectrometer device. 1D Convolutional Model was used

in the classification of the created datasets. In the classification of the NIR reflectance data set, a test

accuracy value of 96.10% was obtained. As a result of these studies, the use of CNN models has proven to

be successful in classification studies.