|
| Design of Image Acquisition System for Bubbles in Pulp of Flotation Machine |
|
Received:June 23, 2022
Revised:July 22, 2022
Accepted:July 26, 2022
Published Online:April 21, 2023
|
| View Full Text View/Add Comment Download reader |
| DOI： |
| KeyWord:flotation; pulp phase; bubble image; image segmentation |
| Author | Institution |
| XU Peipei |
BGRIMM Machinery and Automation Technology Co.,Ltd. |
| RAN Hongxiang |
BGRIMM Machinery and Automation Technology Co.,Ltd. |
| YANG Wenwang |
BGRIMM Machinery and Automation Technology Co.,Ltd. |
| YANG Yihong |
BGRIMM Machinery and Automation Technology Co.,Ltd. |
| LI Qiang |
BGRIMM Machinery and Automation Technology Co.,Ltd. |
|
| Hits: 1811 |
| Download times: 1126 |
| Abstract: |
| In the field of flotation optimization control, the research on froth image acquisition and processing has been widely used. Most of the existing research focuses on the image data collected from the surface of the froth phase of the flotation machine. For the pulp phase, due to the harsh environment, extremely poor light transmittance, and corrosiveness, it is difficult to obtain internal bubble images, thus affecting the development of technology. For this problem, this paper designs a new bubble image acquisition system, which can directly collect image data in situ in the pulp phase, and proposes to use a deep learning semantic segmentation model to achieve accurate segmentation of the pulp phase bubble image. The experimental results show that this system can collect clear pulp phase bubble images in the laboratory environment. The pre-trained model is used for transfer learning based on small samples, and the bubble segmentation accuracy of the validation set is 0.943, which has important research and application value. |
| Close |
|
|
|