Quality only happens when you care enough to do your best.
High-quality silk comes only with a superior grade cocoon. A fast reliable process that can ascertain the quality of cocoons is of high importance in the silk supply chain.
ReshaMandi is using Artificial Intelligence or Image processing to predict the cocoon quality with the help of technology. Cocoons quality is determined by shell percent, filament length, reelability and the percentage of defective cocoons. Using just the image of a cocoon through initial image preprocessing and extracting features from the image, all these parameters are checked accurately and swiftly.
Initial image preprocessing
ReshaMandi employs a deft initial image processing in which the RGB colours of cocoons are converted to grey images. Bilateral filters are applied to soothe the image and reduce the noise in it. Further, through the thresholding technique, in an open CV, pixel values in relation to the threshold values are provided for the image. The adaptive threshold method calculates the thresholding value based on its surrounding pixels. Through one more technique, adjust_gamma, the image is made clearer and is also used to restore the faded pictures to their previous depth of colour.
How are features extracted from an image?
ReshaMandi uses an open CV function, finds contours and the boundary of the image is determined. Through the source image, contour retrieval mode, contour approximation method and Open CV- fitElipse , the major and minor axis diameters of the cocoon ellipse and angle with the horizontal line are deciphered. The image can also be further analyzed for any dents if any.
ReshaMandi is not just addressing the problems related to quality but is able to offer the best solution to the scalability problem. Cocoon grading is not just merely confined to a single cocoon, millions of cocoons need to be accessed. Through AI-enabled image processing, several cocoons can be analyzed easily and quickly.
By leveraging AI, ReshaMandi in its noble endeavor is eliminating complete manual intervention and, saving valuable time in checking the quality of cocoons. With this tech-enabled solution, ReshaMandi is building a fair, uniform procedure and better negotiations in cocoon trading for two important stakeholders of the silk industry: seri farmers and silk reelers. This offers real-time benefits to both. Seri farmers are able to get their cocoons graded without any hassles and can obtain fair prices by selling them, while the silk reelers will procure high-quality cocoons for the best yarn.
