A 3D flame reconstruction system and method based on bp neural network algorithm

Abstract

The invention provides a three-dimensional flame reconstruction system and method based on BP neural network algorithm, including a flame acquisition system, a CCD camera, a computer, a flame generator set in the flame acquisition system, a plurality of camera lenses connected to the CCD camera, and a plurality of cameras. The lens is arranged around the flame generator at any angle of view in three-dimensional space coordinates, the CCD camera is connected with multiple camera lenses through optical fibers, the CCD camera is connected with the computer through a data line, and a filter is provided on the surface of the camera lens. The invention omits repeated iteration and time-consuming, and the reconstruction efficiency is greatly improved. In this model, the reconstruction result is used as the hidden layer and the secondary calculation projection is used as the output layer, which omits the additional comparison samples and is more practical, and the mapping relationship of the learning process is more practical. The object image parameters based on the experimental calibration are used instead of the results of the iterative reconstruction algorithm, which relieves the limitation of the accuracy of the original algorithm.

Description

technical field [0001] The invention relates to the field of aerospace technology, in particular to a three-dimensional flame reconstruction system and method based on a BP neural network algorithm. Background technique [0002] With the rapid development of aerospace research in my country, the diagnosis of internal combustion flow mechanism in complex combustion fields such as gas turbines and scramjets is urgent. Combustion diagnosis technology based on chemiluminescence in the combustion chamber is the most practical optical diagnosis method for real engine diagnosis with the help of photon signals produced by high-energy free radicals in combustion intermediate products without external light sources or lasers. Due to the complex flow field of the engine combustion chamber, which has characteristics such as instability, turbulence, and three-dimensionality, the 3D Computed Tomography of Chemiluminescence (3D-CTC) technology based on combustion chemiluminescence combines...

AI Technical Summary

Problems solved by technology

[0005] This new method has greatly improved the speed of 3D reconstruction, but there are still two problems that have not been resolved: one is that at the beginning of learning, it is necessary to use iterative algorithms (such as ART) to obtain a large number of 3D results in advance, and this process is still time-consuming. , that is, it takes a lot of computing time to obtain additional samples

The second is that since the training is based on the results of prior iterative reconstruction, it is obviously difficult to obtain more accurate results than the previously reconstructed data.

Images