Rapid recognition method of red brick content in recycled brick-concrete aggregates and powder based on color segmentation

This paper investigated a color segmentation-based rapid recognition method for determining the red brick content in recycled brick-concrete aggregates and powder. The study explored various influencing factors, such as surface states, placement forms, image acquisition directions, and scale, on the recognition results. Additionally, the research examined the effect of microscope magnification on the accurate recognition of brick content in recycled powder. To reduce the time cost associated with manual processing of a large number of images, reference thresholds for different color channels of coarse aggregates, fine aggregates, and powder were derived through data statistics. It was demonstrated that the recognition results exhibited fluctuations under different factors. However, the prediction error can be effectively reduced by taking multiple predictions and calculating the average value, thus keeping it within a reasonable range. Particularly in engineering applications, a method that combines comprehensive image acquisition from multiple directions and adjustments based on reference thresholds can be employed to reduce time costs and achieve more desirable recognition results.