vioft2nntf2t|tblJournal|Abstract_paper|0xf4ffb00b2d0000002fb60c0001000100 Seed is a little embryonic plant that can be used to introduce plant infections to new areas while also allowing them to survive from one cropping season to the next. Seed health is a well-known component in modern agricultural science for achieving the required plant population and yield. Seed-borne fungus are a major biotic constraint in seed production around the world. The detection of seed-borne pathogens by seed health testing is a crucial step in the treatment of crop diseases. Speed and accuracy are critical requirements for longterm economic growth, competitiveness, and sustainability in agricultural output. Because human judgements in identifying objects and situations are variable, subjective, and delayed, seed prediction activities are costly and unreliable. Machine vision technology provides a nondestructive, cost-effective, quick, and accurate option for automated procedures. Seed variety, seed type (country seed or hybrid seed), seed health, and purity prediction were the four basic processes. We began the first procedure by aligning the seed bodies in the same direction using a seed orientation approach. Then, to detect atypical physical seed samples, a quality screening procedure was used. Their physical characteristics, such as shape, colour, and texture, were retrieved to serve as data representations for the prediction. This research introduces a new fuzzy cognitive map (FCM) model based on deep learning neural networks that predicts seed purity tests using data from biological investigations. The relevant data features from the seed test are extracted by FCM, which then effectively initialises the deep neural networks. The Levenberg–Marquardt (LM) technique for deep neural networks was discovered to improve seed purity test prediction. Four statistical machine learning algorithms (BP-ANN, Multivariate regression, and FCMLM deep learning). Furthermore, we demonstrated an improvement in the system''''s overall performance in terms of data quality, including seed orientation and quality screening. In independent numerical testing, the correlation coefficient between predicted values and true values acquired from experiments reached 0.9.
M. Suganthi and J.G.R. Sathiaseelan
Bishop Heber College, India
Fuzzy Cognitive Map, Deep Learning, FCMLM Deep Learning, BPANN, Multivariate Regression, Seed Purity
| January | February | March | April | May | June | July | August | September | October | November | December |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 2 | 0 | 0 |
| Published By :
ICTACT
Published In :
ICTACT Journal on Soft Computing
( Volume: 12 , Issue: 4 , Pages: 2715-2722 )
Date of Publication :
July 2022
Page Views :
561
Full Text Views :
4
|