Automatic Assessment of Seed Germination Percentage
Keywords:color analysis, object detection, machine vision, seed germination, Agriculture wastes
This research was designed to investigate an automatic seed germination rate for the top of paper germination method. Chili and guinea were adopted to be used in the experiment with a 4-time repetition and 2 sets of the germination group (4-separated plates with 50 seeds per plate, 2 sets per seed type, totally 400 seeds of chili and 400 seeds of quinea). Two detection methods were proposed binary thresholding and maximum likelihood; based on color analysis. An uncontrolled environment image taking was the way to collect image data. The results were compared to a hand-labeling groundtruth. Both methods achieved accuracy rate higher than 93% which was promising to implement this system. The binary thresholding was a lightweight method suitable for a very limited resource software environment system. The maximum likelihood was more complex. The method had more potential than the binary thresholding, it was flexible to the light condition, returned few false alarms per image (less than 3 false alarms per image). Maximum likelihood could be adopted to implement in a proper environment which still could be in a mobile device.
Authors who publish with Engineering Journal agree to transfer all copyright rights in and to the above work to the Engineering Journal (EJ)'s Editorial Board so that EJ's Editorial Board shall have the right to publish the work for nonprofit use in any media or form. In return, authors retain: (1) all proprietary rights other than copyright; (2) re-use of all or part of the above paper in their other work; (3) right to reproduce or authorize others to reproduce the above paper for authors' personal use or for company use if the source and EJ's copyright notice is indicated, and if the reproduction is not made for the purpose of sale.