Sistem Pendeteksi Viabilitas Benih Kacang Tanah Berdasarkan Luas Area HSV Color

Abstract

Peanut seed tetrazolium test evaluation is usually by eye and a microscope. This method has a weaknesses in the accuracy of reading the color intensity, and  is more subjective. The seeds was observed one by one so that the observation is not effective. To make observations more accurate, efficient, and effective, digital image processing can be applied to the seed viability evaluation. The method can be used was the detection of the Hue, Saturation, and Value color area in reading the red color pattern resulting from tetrazolium test.The result is the system can detect a maximum of 25  seeds with an operational time of 22-25 seconds in one detection. Seed classification is the seeds are predicted to normal, abnormal, and dead. The process of classifying seeds is identified based on the red color pattern resulting from the detection of the area of 4 HSV color ranges, namely red (175,100,20:180,255,255), pink (160, 100,20 : 174,150,255), white 1 (175,0,0 : 180,100,255), and white 2 (0,0,0 : 100,255,255). The results show that the accuracy of the system in reading the total number of seeds is 100% with the detection error of  HSV color area is 1.54%.