Research On Recognition And Track Algorithm On The Laser Scattering Rate Of Sedimentation Particles

The size and shape of the particles play a crucial role in food testing,healthcare,materials science,and transportation construction.In the materials science industry,the size of the material particle has an important influence on the material.Therefore,the material particle size analysis has great research and application value.The sedimentation particles in this paper are relatively small and similar to the background color,which is not conducive to recognition.The particle will be covered,merged,disappeared in the process of falling,which is not conducive to tracking,so this paper analyzed the characteristics of the sedimentation and combining the fuzzy set theory can effectively solve the fuzzy problem of sedimentation particles,improve the recognition efficiency,and then puts forward the maximum distance method to complete the sedimentation particles tracking.The main contents of this paper are as follows:(1)The characteristics of the sedimentation particles are extracted.By analyzing the video of the sedimentation particles and the sequence images of the sedimentation particles,the gray characteristics,morphological features and motion characteristics of the sedimentation particles are analyzed.By analyzing these features,different sedimentation particles can be distinguished for subsequent identification and tracking.(2)The genetic algorithm is improved by improving the encoding rules and termination conditions,and apply it to the segmentation of the sedimentation image based on fuzzy partition entropy.Experiments show that the improved algorithm has better segmentation performance.(3)Fuzzy partitioned entropy threshold segmentation algorithm has the problems that the running time increases with the increase of the number of thresholds and the noise is large.Through the analysis of the calculation process of the fuzzy partition entropy,it is found that there are a large number of repeated calculations,this article will recur the idea using fuzzy partitioning entropy to reduce part of the repeated calculations,then using the optimization algorithm to search for the global optimal threshold.(4)The similarity between the sedimentation particles and the background studied in this paper is very high and there are many uncertain ambiguities.Traditional identification methods are difficult to identify the sedimentation particles.This paper applies the fuzzy comprehensive evaluation method to the identification of sedimentation particles,and then selects the appropriate characteristic amount according to the characteristics of the sedimentation particles,and then establishes the membership function of the characteristic amount to identify the sedimentation particles.(5)In(1),the gray-scale characteristics,morphological characteristics,and motion characteristics of the sedimentation particles were analyzed,and other features such as the polar axis and the angle between the polar axes were not taken into account because they were difficult to measure,and thus the characteristics of sedimentation particles extraction is not very rich,so this paper uses YOLO convolutional neural network and improves the YOLO loss function,inception structure model and spatiotemporal pyramid pooling,and theoretically elaborated the feasibility of the improved YOLO convolutional neural network.(6)The maximum distance method is proposed to track the sedimentation particles,and the phenomenon of aggregation,dispersion and disappearance of settling particles during settlement is also analyzed.The correction strategy is also put forward for the failed particle trajectories and settling particles.