Cotton is an important natural fiber resource contributing to economies around the world, and its production is an important economic factor for the United States as the leading exporter, accounting for over one-third of global trade in raw cotton (USDA, 2013). Cotton quality, affecting both profitability and marketability, is of concern to the industry. Cotton trash is one of the important factors in cotton quality. Although most large-sized cotton trash (e.g. hulls) can be directly removed after ginning, a lot of cotton trash is cut into smaller pieces and remains mixed with lint fibers. These small pieces of trash include various types of things which cause different damage to cotton fiber or textile machines during textile processing. Therefore, cotton trash detection and identification is urgent to the cotton and textile industry. Currently available equipment for measurement of cotton quality includes the High Volume Instrument (HVI) and the Advanced Fiber Information System (AFIS), but none of them can quantify and differentiate cotton trash at present. In order to solve this issue, our lab has been developing a new system based on hyperspectral imaging to detect and identify cotton trash.