Fractional Differential Image Enhancement Technology And Its Application In Copper Flotation Monitoring System

In copper flotation industrial environment, the cameras are working in a noisy environment caused by mechanical vibrations, illumination changing, dust, smog, etc. Thus, the obtained froth images are seriously affected, which are blurred with obvious shadow and reflection and almost no background information. Moreover, the flotation images are blurred with jitter phenomenon, complex texture, stable grayscale and indistinct edge. Therefore, it is difficult to process the froth images with traditional methods. For one thing, traditional integer-order differential image enhancement operators are not sensitive to the image edge with stable grayscale, which fails to achieve the expected effect. For another, burgeoning image enhancement methods such as neutral network and wavelet transformation, which would take up too much time and energy, are rarely applied in industrial field. However, fractional differential is suitable for processing images with unobvious edge, large noise and unclear textures. Therefore, a fractional differential technology is introduced in the paper to enhance the copper flotation froth images. Considering the characteristics of froth images, an image enhancement operator of fractional differential is constructed, and an automatically-adapted image enhancement method is proposed, showing good simulation results. The main contents of this paper are as follows:(1) The image characteristics of copper flotation and Fractional differential image enhancement technology is firstly described, fractional based on the image structure and shape, an image enhancement mask with rotational invariance is constructed. The mask is characterized with fractional differential, as well as simple calculation, low time complexity, operative convenience and applicability in industrial field.(2) Since manually retrieving the order of fractional differential is time-consuming and inaccurate, a self-adaptive image enhancement method based on fractional differential is proposed. It takes advantage of dichotomy and adaptively selects the fractional differential order, based on the image texture. Compared with manual retrieval of fractional differential order, the project method obtains more precise fractional order, produces more favorable after-process image and save numerous time.(3) By contrasting the performance of proposed method with those of popular image enhancement methods in simulation experiment, it shows that the proposed method in the paper is characterized with higher PSNR, SSIM and IE values, which can also obtain more accurate bubble numbers and more reasonable suggestion when using it in copper flotation image monitoring system, laying a foundation for optimal control of copper flotation process based on image characteristics. The paper consists of35figures,3tables and75references.