Flotation bubble identification method based on cascade classifier

Abstract

The invention discloses a flotation bubble recognition method based on a cascade classifier. The method comprises the steps that firstly, collecting original bubble images in a flotation tank under various working conditions; marking a bubble region from the collected original foam image as a region of interest, and generating a bubble detection positive sample set; intercepting a non-bubble areafrom the collected original foam image to generate a background negative sample set; sending the generated positive and negative sample sets into a cascade classifier at the same time to obtain a bubble detection model; loading the trained bubble detection model into an upper computer on a flotation site, detecting bubbles in the collected original bubble image in real time, and displaying a detection result; and identifying the position of the complete bubbles in the original foam image according to the detection result, and counting the size distribution of the bubbles. The invention has lowrequirements for image quality, is not affected by a plurality of bright spots, color blocks, collapse and the like on the surface of the bubble, and can quickly and accurately identify the bubble.

Description

technical field [0001] The invention relates to the technical field of flotation bubbles, in particular to a method for identifying flotation bubbles based on a cascade classifier. Background technique [0002] Flotation is a beneficiation technology that uses the differences in the physical and chemical properties of the mineral surface to separate the target minerals from the pulp. At the flotation site, the size and state of the bubbles are the most intuitive representations that reflect whether the current flotation conditions are operating normally. The traditional method of adjusting the operation through manual observation of foam is greatly affected by subjective experience, has low precision, and has hidden dangers to personal safety. Therefore, the study of automatic foam identification technology is of great significance for reducing labor intensity, guiding on-site production, and improving the utilization rate of mineral resources. important practical significan...

AI Technical Summary

Problems solved by technology

The traditional method of adjusting the operation through manual observation of foam is greatly affected by subjective experience, has low precision, and has hidden dangers to personal safety. Therefore, the study of automatic foam identification technology is of great significance for reducing labor intensity, guiding on-site production, and improving the utilization rate of mineral resources. important practical significance

[0003] The research hotspots of the existing technology mainly focus on identifying bubbles through image segmentation, and make improvements based on traditional methods such as watershed, edge detection, and region growing. Problems such as different and uneven illumination make these methods easy to over-segment and difficult to adapt to changing working conditions.

Images