| Because the ablation mechanism of gun propellant to weapon barrel is complex,the ablation research of gun propellant depends on experiments to a great extent.In the traditional semi closed ablation tube method,the performance of gun propellant is only measured by the mass change of ablation tube before and after the ablation experiment.Therefore,in order to comprehensively evaluate the performance of gun propellant,the change of the orifice diameter of ablation tube before and after the experiment and the ablation tube during the experiment are introduced The nozzle flame size is a new evaluation index.It has important practical value to measure the performance of propellants by new evaluation index in the development of propellants.According to the measured characteristics and the corresponding technical indicators,the ablation tube parameters and flame measurement technology are studied.The ablation tube orifice aperture measurement system based on industrial camera and the flame measurement system based on high-speed camera are designed.The hardware selection and software scheme of the measurement system are completed,and the key technologies in the system are studied,including sub-pixel edge detection Edge detection technology,semantic segmentation technology and camera calibration technology.The paper compares and analyzes several common pixel edge detection algorithms,selects adaptive Canny algorithm to locate the image of ablation tube aperture,and combines the improved Zernike moment sub-pixel edge detection algorithm to complete the edge detection of the nozzle aperture of the ablation pipe.Through the contrast experiment,the improved algorithm has been improved in accuracy.The image segmentation method based on semantic segmentation model is studied.Combined with the characteristics of ablation nozzle flame,UNET network is selected to realize flame segmentation.In order to improve the flame segmentation accuracy of the measurement system,this paper improves the UNET network,introduces the depth separable convolution and residual network,and through the experiment,verifies that the improved UNET network has improved the segmentation accuracy,F1 score has increased by 0.15 to0.985.Finally,the measurement system was built,and the nozzle diameter of the ablation tube and the nozzle flame during the experiment were measured.The results show that the accuracy of sub-pixel edge detection algorithm based on Zernike moment can meet the technical requirements;the flame segmentation algorithm based on UNET network can measure the maximum single frame flame area of20.1288 m in the experimental process,and the flame superposition area of20.1727 m in the whole experimental process.The experimental results verify the effectiveness of the proposed method,which can provide experimental basis for the development of propellants. |
