Comprehensive defect detection of bamboo strips with new feature extraction machine vision methods

Bamboo strips, as assembling parts of sleeping mats, cushions and other decorative components, play an important role in humans’ everyday life and social economy now. Therefore, quality control for bamboo strips production is very critical. Traditional manual sorting technology owns many disadvantages such as high production cost and low sorting accuracy. This work deals with an automatic sorting system for comprehensive defect detection of bamboo strips based on machine vision. Differing from the present feature extraction methods of the bamboo strip in image processing, contour features considering area and geometrical symmetry and texture feature considering average gradient are newly introduced. An experimental automatic sorting system is designed to verify the feasibility and superiority of the proposed comprehensive defect detection method. Experimental results show that total defect detection accuracy, contour defect detection accuracy, surface texture defect detection accuracy and sorting accuracy reach 99.1%, 98.33%, 95.2% and 95.125%, respectively. The designed sorting system finishes one time sorting in 197 ms with a comparable low-speed computation processor in laboratory and it can be utilized instead of three skilled workers in practice.