A contraption endorsed Indian pulse seed recognition system with quality determination prioritizing compound imaging techniques

The main objective of presenting this context is to identify and verify the quality of the seed for future fertilization in the field of agriculture. This research paper proposes a novel image processing technique that includes two phases depicts an optimized selection of feature extraction and classification algorithm that enhances the quality, exactness of the seed variety and realization of the most excellent classification production percentage of 98.9%. In the identification and classification phase, image representation of the pulse seed varieties is pre-processed to enhance the seed image with S-component Conversion and the escalated seed image is processed to detect the outer boundary of the seed and also the inner region of the pulse seed image is extracted for the more accurate and closed boundaries to detect and fill the inner holes of the pulse seed image. This proposed algorithm employed with extraction of 256 features by evaluating the shape, color, texture and seed specialized features. And to attain significant features selection by weighting and ranking techniques for the particular varietal classification. In the second phase, the quality determination of the pulse seed is established to find good viability condition of seed using thermal imaging techniques. The quality aspects are estimating water content in the seed, germination level, and vigorness of the seed and trueness of the seed. In our proposed work average utilization time for processing of identification and classification for each seed is 0.21s. Findings: This work implemented two main phases: Pulse seed Identification and classification and Pulse Quality Determination. In Phase-I, the seed image is acquisitioned and pre-processed by image enhancement and noise removal. Then the enhanced image is processed to detect the inner and outer region of the seed image for the identification process. Then features are extracted from the segmented image for the classification of the pulse seed variety. To minimize the complexity and time consumption of the feature extraction is significant feature are selected using feature selection for classification. In Phase-II, the quality of the seed is determined by evaluating the quality test aspects of the healthy seed such as Germination, Vigorness and Seed color Purity.