In recent years,due to frequently accidents in mine,thus,there is an urgent need for an effective and robust recognition algorithm that can be used in applications with low definition and high noises.Under severe conditions inmine,video signals are tend to be polluted by strong noises during sampling and transmission.This study presents a mine safety monitoring algorithm based on computer vision technique.It involves two steps,namely preprocess andrecognition of visual information.Our contribution lies in two aspects.First,a video enhancement method is addressedusing non-local means accelerated by inter-frame similarity prediction.Second,an object recognition approach is proposed based on improved Hough forest.It speeds up recognition process by reducing the size of search window.And it increases the recognition accuracy through introducing time-dimensional analysis.Experimental results illustrate that the proposed algorithm obtains 66％ reduction in denoising time and an improved recognition rate.It is easonable and reliable to apply our algorithm to mine afety monitoring.