92D-12 |
Prediction of optimum harvest time of peanuts based on a color analysis with machine vision |
S. DAMAR1, M. O. Balaban1, and A. Drew2. (1) Dept. of Food Science & Human Nutrition, Univ. of Florida, 359 FSHN Bldg., Newell Dr., PO Box 110370, Gainesville, FL 32611-0370, (2) Institute of Food & Agricultural Sciences, Univ. of Florida, Levy County Extension, PO Box 219, Bronson, FL 32621 Determination of peanut maturity and optimum harvest time are critical for harvest yield. The conventional method to predict optimum harvest time is to subjectively evaluate the color of the saddle area of scraped pods. A “Peanut Profile Board” is available to facilitate this. An objective alternative is the use of a machine vision system to evaluate colors. Our objective was to quantify the colors of different peanut samples judged by an expert as to their optimum harvest time, using 2 methods, and to correlate the color information with optimum harvest time. Peanut samples, consisting of 60-90 peanuts each, were picked at different times (0,10,14,17 and 21 days) prior to optimum harvest day, as judged by an expert. The outer hull layer was scraped in a tumbler. The pods were placed in a light box and their video image taken. The image was processed by software to measure the L* value of each pod. In the first method, the average L* value of the image was calculated. The data was analyzed by regression. In the second method, the percent area of all pods in an image that had L* values below 10, 20, 30, 40, 50, 60, 70, 80, and 90 were calculated. These areas were plotted against the L* values. The area under the curve was then calculated, and plotted against days to optimum harvest. The plot of L* vs days-to-harvest showed a linear trend (R2=0.98) until day 17. The area vs days-to-harvest also had a linear trend up to day 17 (R2=0.98). Day 21 point was above the line for method 1, below the line for method 2. Results suggests that machine vision color analysis of peanuts can be used to objectively and rapidly predict optimum harvest time.
Session 92D, Quality Assurance: General
|