Prediction of Seedcotton Yield with Digital Imagery

One of the most common and likely least reliable selection processes in most cotton breeding programs involves visually selecting single progeny rows for advancement within the program. A study was conducted in 2014 that attempted to correlate the number of white pixels in a digital image of a row of cotton to known values of seedcotton that was machine harvested and weighed. Images and weights were taken from a drought study at College Station and Thrall, TX, as well as the National Regional Standard cultivars in the commercial variety trials at College Station, Thrall, and Commerce, TX.  If these comparisons between white pixels and seedcotton weight prove to be highly related, potential applications could include assistance in single row selections, crop maturity ratings , and even crop insurance adjustments.