An Algorithm to Obtain the Stained Area Ratio Based on Digital Image Processing Techniques for Organic Straw Classification

The present work proposes the development of an algorithm to obtain the relationship of stained areas based on digital image processing techniques for the classification of organic straws. The area ratio in the straw is a parameter that indicates how stained the organic straw is. This parameter is calculated from the information of the image taken, considering how many pixels are part of the straw and how many pixels are part of the unstained part of it. The sorting of straws is usually done by visual perception of the operator. However, this method is very ineffective and time-consuming, and it should be noted that operators have three-hour breaks due to eyestrain. This affects the speed of the sherbet sorting process and therefore the entire manufacturing process. The proposed algorithm aims to reduce the subjectivity of the analysis and improve its speed. The algorithm is composed of processing steps that involve image conditioning using Sobel gradient filter, sharpening filter, morphology, angle correction, histogram skew analysis, decontamination by the Multi-Diagonal Lambda Noise Filter, detection of the straw and not stained areas of the straw, and finally, obtaining the area ratio. Finally, the proposed algorithm obtained a precision of 70%.