A Study On Applying Computer Vision In The Evaluation Of Stem Volume

At present, the domestic detection technology of stem volume generally laggedbehind, the evaluation of stem volume based on computer vision can effectivelyimprove the detection efficiency while reducing costs. Although the theoretical studyof the technology has made some achievements, but so far the domestic product of ithas not yet been realized, so this study has the significance both in economic andtheory.Machines replace human labor is the trend of the new century in which humanprogress, This paper conform to the direction of technology development,comprehensively analyze the status of evaluation of stem volume at home and abroad.Make a deep study on the problem of estimates of stem volume of equal length heap,As an important supplement to intelligent detection of stem volume, this studydescribes an overview of computer vision computer vision, analyze the request of thedata collection phase.Then, make an innovative research on the methods of data processing:1, Optimize multi-template filtering method based on MSE, select the bettertemplate;2, improved the classical median filtering method, the improved methodimproves the robustness to impulse noise;3, the proposed combination of these two methods of image filtering and imagefor the pre-timber end, experimental results show that this method is better than thecomparison method;4, comprehensively analyze a variety of adaptive threshold segmentationalgorithm, proposes segmentation method combined with two-dimensional entropysegmentation method and the filtering method this paper introduced, the experimentresults show that this method compared with comparison algorithm has betteraccuracy and better adaptive level;5, analyze several edge detection operator, and compare the application of thevarious methods in this task;6, proposes an improved adaptive edge detection method based on block andmean gradient, that the quality is better than compared detection algorithms;7, proposes a new scanning algorithm for better performance; improved scanningefficiency of label processing; achieves the removal of interference region; comparedwith the comparison methods, the method proposed has better robustness andefficiency.8, proposes the method with combination of proposed adaptive thresholdingalgorithm and introduced edge detection methods to improve the accuracy and the adaptive ability;Then introduced the estimates are based on computer vision using themathematical model, calculation methods. Analyze the advantage witch computervision method has to the manual testing. Finally, analyze the deviation during thestudy and shows the main results of this paper. Lastly, discusses the shortcomings ofthe work, points out the direction for the follow-up work.