Research On Detection Technology For Cracks Of Coal Conveyor Belt

As the carrier of the coal, steel cord conveyor belts are widely used in the coal industry. Ripping on coal conveyor belts brings great potential hazard in coal production. Therefore, studying on detecting cracks of coal conveyor belts plays an important role in practice.On the basis of further study on typical machine vision system and the application environment of the coal conveyor belt, this paper proposes an automatic detection system for surface cracks of coal conveyor belt based on Linear array CCD. According to characteristics of the technical requirements on crack detection, overall systemic structure of the system is introduced. The precision of collection images and adjusting of aperture are studied. Using the double DMA buffers, the real time image acquisition is carried out, and the depth of the DMA is o ptimized.The images of conveyor belts always have some disadvantages in low contrast and plenty of noise. This paper, at first, analyzes the limitations of convention pre-processing method. Basing on this, a one-dimension differential calculation, which is suitable for real time detection, is presented. Since the template design is the most important factor in differential calculation, we have three methods to build them: simple template, median-filtering template and curve-fitting template. These three methods are compared under the same conditions.This paper gives a fast algorithm for the crack's length extracted by which calculates diagonal line in crack's enclosing rectangle, and then an approximate value is obtained. For getting a more accurate value of crack's length, this paper proposes an algorithm based on mathematical morphology. This algorithm refers several steps: first of all, separating cracks in different directions by directional structure element. Secondly, jointing disconnected cracks by dilating. Thirdly, the skeleton of the cracks are got by thinning processing. After all these, the accurate length of the cracks could be obtained. The result of experiment detecting the surface cracks of the belt with above algorithm demonstrates that this method obtains the high precision.