Flotation froth feature extraction method and flotation froth abnormal working condition automatic detection method

The invention discloses a flotation froth feature extraction method and a flotation froth abnormal working condition automatic detection method. The method comprises the steps that 1, a froth real-time video image is acquired; step 2, constructing an improved YloV5 model to carry out target identification; 3, inputting the video image data in the step 1 into the YoV5 model improved in the step 2; 4, output data of the step 3 are obtained, wherein the output data comprise the flotation foam size grading proportion, the flow direction and flow speed and the life cycle; standard flotation froth and optimized flotation control are achieved. The method is used for solving the problems that the foam size, shape, color and other characteristics change greatly, the difference is obvious, and the area difference between the large and small foams is large, and detecting whether the working condition is abnormal when the characteristic extraction method is used. 本发明公开了一种浮选泡沫特征提取方法及浮选泡沫异常工况自动检测方法，步骤1：获取泡沫实时视频图像；步骤2：构建改进的