Machine vision automating cotton gins

June 9, 2005
Patented machine vision system offers real-time monitoring of "trash" content that will enable cotton ginners to move closer to an automated system.

AN AGRICULTURAL RESEARCH scientist has developed a patented machine vision system for automating cotton gins that is now being manufactured commercially. The system uses color video cameras for a hands-off monitoring of the "on-the-fly" flow of cotton and attendant debris through the many machines that make up a modern cotton gin.

Mathew G. Pelletier, an agricultural engineer with the ARS Cotton Production and Processing Research Unit in Lubbock, Texas, invented the system. Its real-time monitoring of "trash" content will enable cotton ginners to move closer to an automated system that seems inevitable in the near future. It can advise ginners or automatically turn cleaners on and off, depending on trash content.

By knowing the trash content at each stage of ginning, gin operators will know exactly which stages they can or cannot skip to reach particular quality goals without under- or over-cleaning cotton. Under-cleaning lowers cotton's value and makes the cotton more difficult to use at textile mills, while over-cleaning wastes cotton. Added weight from the saved cotton provides an additional $3 to $5 per 500-pound bale, which could lead to more than $100,000 in increased revenue for a typical cotton gin.

Knowing the trash content also helps gin operators judge how much heat is needed to dry the cotton. The system includes ARS-developed software that enables computers to recognize trash on video images fed live to the computer.

Not only can the computer distinguish cotton from trash, it can also distinguish among various trash components--the sticks, leaves and burrs that cling to cotton after harvest.

The system films 50 square feet of cotton a minute, giving gin operators a very accurate measurement of trash levels.