Machine Vision Method for Quantitative Statistics Analysis of Industrial Product Images
Keywords:Image processing, Machine vision, Industrial application, Sensing technique
To address the problems of unstable accuracy, low efficiency, and subjective influence of manual counting, a machine vision-based method to count the quantity of tobacco shreds is proposed for the first time. In this paper, the complex tobacco shred image is obtained by backlight imaging. The adaptive threshold segmentation method is used to segment tobacco shreds. The pixel area of the tobacco shred area is calculated by connected domain labelling. Second, independent tobacco shreds and adhesive tobacco shreds were identified based on the pixel area, and the quantity of segmented tobacco shreds was counted for the first time. Subsequently, in complex scenarios (such as tobacco shreds adhesive and overlapping), an image is usually obtained by manually drawing the contours of the adhesive and overlapping tobacco shreds on the basis of primary statistics. Finally, different individuals are distinguished, segmentation is completed, and tobacco shred quantity statistics are realised. The experimental results show that the average accuracy is 100.0 % for quantitative statistics of independent tobacco shred images. For tobacco shred images with adhesive and overlapping interference, the minimum accuracy is 90 %, and the accuracy increases with the increase in tobacco shred quantity. Furthermore, the efficiency of the tobacco shred quantity statistics conducted by the method in this paper was only affected by complex scenarios. Compared to artificial processing, the efficiency was increased by more than 100 %. The work in this paper can provide the technical basis for measuring the dimensions of tobacco shreds.
How to Cite
The copyright for the paper in this journal is retained by the author(s) with the first publication right granted to the journal. The authors agree to the Creative Commons Attribution 4.0 (CC BY 4.0) agreement under which the paper in the Journal is licensed.
By virtue of their appearance in this open access journal, papers are free to use with proper attribution in educational and other non-commercial settings with an acknowledgement of the initial publication in the journal.