| The importance of military development to a country is obvious.Modern military weapons are the symbol of national deterrence,and the existence of firearms,a basic weapon,is also essential and plays a decisive role in war.When shooting,there are many problems,such as noise,smoke,muzzle flame,harmful gas emission and so on.Smoke is one of the harmful phenomena.The smoke produced by gunshot can easily expose the position of the shooter,which on the one hand will threaten the life safety of the shooter,on the other hand will affect the accuracy of the shooter’s second firing.In addition,incomplete combustion of propellant will cause more smoke at the exit of the gun,and the energy can not be effectively used,resulting in a lot of waste.Therefore,the study of muzzle smoke is of great value to the development of weapons in China.This paper designs and implements an image-based muzzle smoke detection system.Based on the analysis of various requirements of the system,the main modules are designed and optimized in detail,including data acquisition module,data calibration module,target plate background module,smoke transmittance calculation module and smoke area module.Aiming at the problem of low accuracy of the measurement results of light detection method,the system uses the continuity of smoke in time and space to segment the target image into multiple small areas,so that the calculated value of its transmittance is more consistent with the actual situation;Aiming at the problem that it is difficult to characterize the muzzle smoke area,the smoke area parameters with a certain range of smoke transmittance can be obtained by combining the smoke transmittance characteristics with the spatial characteristics;In view of the irregularity of the muzzle smoke image and the fast speed of smoke diffusion,the traditional edge detection algorithm can not effectively extract the smoke edge contour.An edge detection algorithm based on dyadic wavelet and mathematical morphology is designed to detect smoke,which improves the accuracy of smoke edge detection;Aiming at the problem of inaccurate target area recognition,the method of making difference between smoke frame,background frame and non target image is studied to make it easier to recognize smoke target and improve the accuracy of measurement results.In this paper,the muzzle smoke simulation model is also established.Through the smoke particle transmittance calculation model simulation,the transmittance versus time curve is obtained.Through the smoke area algorithm,the area versus time curve is obtained.The accuracy of this method is verified by comparing with the transmittance and area curve obtained from the actual field test experiment;Analyze the experimental data,study the specific factors affecting the change of transmittance and area curve,analyze the distribution characteristics of smoke,and complete the theoretical verification.Field experiments show that the muzzle smoke detection method designed in this paper can ensure the accuracy and stability of the measurement.The standard deviation of the measured transmittance is less than 8%,and the standard deviation of the measured smoke area is less than 5%,which fully conforms to the technical indicators.The detection method can provide a basis for the research of barrel weapons. |
