Computer-Vision System for Control of Drying Processes

Abstract

Computer-vision system (CVS) for control of a drying process with a portable CCD camera with IEEE-1396 interface and configurable software LabView 7.0 and IMAQ^TM 6.1 was developed. An object area was continuously monitored through the CVS by extracting the green plane from the RGB color space followed by thresholding and pixel counting. An object color was continuously monitored through the CVS as color intensity in the hue-saturation-intensity (HSI) color space. The observability of a drying process was provided due to online image analysis and correlation of image attributes (area, color, texture) with physical parameters of drying (moisture, quality). A relationship between area shrinkage and moisture content was used for online estimation of actual moisture content. A relationship between color intensity and quality was used for online estimation of quality degradation.

Experimental study of the CVS for ginseng drying showed advantages of computer-vision for online monitoring of important state variables, such as moisture content and material quality. Color measurements demonstrated high sensitivity of quality to drying conditions: drying at 50°C resulted in significant color changes and unacceptable quality degradation. The quality of roots in three-stage (38-50-38°C) drying process was compatible with recommended isothermal (38°C) drying due to significant (30–40%) reduction of drying time. This control strategy was used in a pilot batch dryer for temperature control with respect to quality. Testing of a pilot dryer with embedded CVS proved stability and robustness of control strategy, combined with high accuracy in the estimation of moisture content (8–14% of error with 95% confidence). The composite moisture measurements at the endpoint demonstrated uniform drying of root mixture to target moisture content 0.1 g/g (db) with minor variations between individual roots in the range of 0.07–0.12 g/g.

Keywords:

• Color
• Ginseng
• Machine-vision
• Moisture
• Quality
• Shrinkage

ACKNOWLEDGMENTS

I would like to express my gratitude to Dr. Valerie Davidson and Dr. Ralph Brown, who helped me very much with valuable comments. I would like to express special acknowledgement to Scott Noble for his help in the development of IMAQ Vision Builder control applications.

Notes

^a Mean values and standard deviations (95% confidence) were calculated from four identical drying experiments (38°C temperature, 1 m/s air velocity, and 12% relative humidity).