Straightness measurement algorithm based on machine vision for coal longwall face

In the mining process of fully mechanized coal face with few hundred meters,long,narrow and undulating curved,face straightness is a non-visual measurement of large moving target.Studying the trapezoidal windows,grayscale,recognition and feature search vision algorithms as real-time measuring form of scraper conveyor,the straightness of local coordinate system is established with measurement of structure light pixel value to a straightness deviation from set.By chain-transfer algorithm,the face layout of multiple local straightness with camera coordinate system is transformed to global coordinate system with matrices and translation vector,as the conveyor section is the face of straightness measurement markers.Experiments show that visual inspection achieves accuracy of ±50mm for the whole face straightness,it meets requirements of highly effective shearer running along straight face.