Research On Automatic Volume Measurement Of Irregular Coal Pile Based On Binocular Stereo Vision

The measurement of coal pile volume is an important process of coal mines,power plants,heating and other companies.It is the prerequisite for calculating coal supply and production planning.The active application of new science and technology to improve the ease and accuracy of coal pile measurement is the current trend in the coal pile measurement business.The accurate calculation of irregular coal pile volumes has long been a challenge.Traditional measurement methods are labour-intensive and time-consuming,and the low density of measurement data makes it difficult to achieve fine management,resulting in low volume measurement accuracy and the inability to obtain accurate 3D real-world data of the heap.To address these shortcomings,this paper proposes a research method for irregular coal pile volume measurement based on binocular stereo vision technology.The research approaches are:(1)Image pre-processing.To address the issue of noise in the image pairs obtained by the ZED binocular camera,median filtering is first used to denoise the coal pile image to reduce noise;Then,the Histogram equalization algorithm is used to enhance the local features of the coal pile image;Finally,an improved semi global stereo matching algorithm is used to obtain disparity maps with a low mismatch rate.(2)Research on irregular coal pile stereo matching algorithm.In response to the discontinuity of the parallax map obtained by the traditional semi-global algorithm,this paper proposes a new semi-global stereo matching algorithm,which uses multi-feature fusion technology to combine several similarity evaluation parameters such as Census transform and image gradient to evaluate the image matching degree in a more comprehensive way.And it also uses the summation of 16 directional path generation values to obtain the final generation values,and performs steps such as sub-pixel enhancement and left-right consistency check on the initial parallax map to obtain the final optimized parallax map.(3)A method of volume measurement based on parallax maps is proposed.For the calculation method of using the average height of coal piles,the volume of the volume will lead to a large error.This article searches the average depth information of each point of each two-dimensional graphics by searching for the three-dimensional value of its relative position.On the basis of the value,the volume of the coal pile based on pixel-based volume measurement is used.When measuring volume,the height of the pixel represents the actual height of the coal pile.This technology uses the vision image as the main input.First,the segmentation of the visual images is completed through the watershed segmentation algorithm,and then the three-dimensional reconstruction of the target(coal pile)area is adopted by a three-dimensional imaging method.Finally The volume calculation method has implemented the volume estimation of the coal pile.(4)Build an experimental platform.It aims to measure the volume of irregular coal piles.Through experiments,the volume calculation method of irregular coal piles meets the measurement accuracy requirements of less than 10%,providing a solid theoretical and technical foundation for further practical applications in this field.