| The image processing has extensive industrial applications instead of the manual inspection. In this paper , some image processing methods, especially the mathematical morphology, are studied and applied in the industrial online inspection system.Considering the requests and characteristics of the boiler liquid level online inspection system, the pretreatment of the liquid level image, extraction of the scale, sub-pixel edge detecting and correcting the result are studied in detail. After contrasted and analyzed the result of these methods, some of them are chosen to apply in the liquid level online inspection system, and get the value of liquid level finally. Many measuring experiments were carried on and the results indicated that the project can basically satisfy the request of the online inspection system on accuracy. The system's feasibility and validity were verified.The images collected from strip steel product line have the disadvantages such as low contrast and plenty of noise, as well as the quite high requirement of the speed in the real-time inspection system. For these problems, a new surface defects inspection of strip steel based on the scale Top-Hat transform is presented in this paper. The experimental results show that the algorithm proposed in this paper can considerably reduce noise and accurately extract the defects, comparing with the conventional algorithms. Designing of structure elements is also analyzed and studied, and two adaptive methods are presented. These principles of designing were used in the processing and got good result. The feasibility of adaptive methods of structure elements designing was verified.Finally, the principle of hardware selection and designing are discussed. Two software and hardware designing schemes of online inspection system based on mathematical morphology are proposed. |
PDF Full Text Request