Research On Microfluidic Visual Camouflage System And Camouflage Characteristics Based On Image Completion

Camouflage has become an essential means of self-preservation for military weapons and military equipment,and the continuous improvement of reconnaissance technology has imposed higher requirements for camouflage means.At present,dynamic camouflage technologies based on the thermochromic,photochromic,and electrochromic materials have limitations in dynamically adapting to environmental changes due to the limited number of color changes,high power consumption,and unstable performance.In contrast,microfluidic visual camouflage systems are suitable for both static and dynamic camouflage due to their small size,low energy consumption,high efficiency,and ease of integration.To achieve the microfluidic dynamic camouflage that is highly matched with the background and has a vivid effect,this paper combines the deep-learning image completion algorithm with the microfluidic technology and proposes a microfluidic visual camouflage system,which is able to camouflage dynamically by virtue of the deep-learning image completion.With the chameleon skin regarded as the camouflaged object,the "destructive coloring" and "fusion coloring" design methods of the chameleon skin camouflage film are studied,and the structure and working principle of the microfluidic visual camouflage system are illustrated.As to the hyperelastic camouflage film characterized by a nonlinear stress-strain relationship,a mathematical model of liquid flow resistance in microchannels and a theoretical calculation method of pressure loss considering the film deformation are established.Then,the unidirectional and bidirectional fluid-structure interaction models are employed to simulate and analyze the microfluidic liquid flow characteristics of the vis ual camouflage system,as well as the effects of the flow rate on pressure distribution,velocity distribution,equivalent stress,equivalent strain,and film deformation.Meanwhile,dedicated experimental tests are taken to confirm the simulation results of the pressure loss characteristics and film deformation characteristics of liquid flow in the microchannels.Also,the dynamic response time calculation model of colored liquid filling in the microchannels is deduced with the influence of nonlinear high elasticity of the camouflage film considered,and it is validated by the method of measuring the deformation of the film using a laser displacement sensor.The visual camouflage properties of the camouflage films are investigated so as to optimize the microchannel parameters for achieving the best camouflage performance.To this end,a comprehensive evaluation index of the visual camouflage characteristics of the camouflage films is established by the entropy weight method considering the color,brightness,edge shape,and texture features,and the influence of the microchannel parameters on the visual camouflage effect of the film is analyzed.Then,the response time of visual camouflage discolouration of the film samples is explored based on the method of image brightness contrast.Meanwhile,the influences of the microchannels on the pressure loss and deformation characteristics of the film are revealed by means of the bidirectional fluid-structure interaction simulation and experimental test.At last,the optimum microchannel parameters are derived from a comprehensive evaluation of the visual camouflage effect of the film,the visual camouflage discolouration response time,and the pressure loss characteristics under different microchannel parameters.The deep-learning image completion algorithm applied to microfluidic visual camouflage is probed by constructing a variety of initial camouflage images,which have a high degree of confusion and different color block distributions.An investigation is conducted to clarity the influence of the different combinations of the K-means color quantization algorithm,watershed algorithm,and main color extraction algorithm on generation of the discolouration signals.Furthermore,the color and color block distribution information of the camouflage images is used to create the signals for the microfluidic visual camouflage system.Afterwards,a logic control method of the electromagnetic switch is designed to drive the colored liquid to match the corresponding microchannel of the camouflage film.Through a microfluidic visual camouflage test system,the optimal camouflage image is determined for the environmental background image by the deep-learning image complete method,and different discolouration signals are generated for different camouflage images.Meanwhile,experimental tests are performed to verify the visual camouflage characteristics of the chameleon skin based on the comprehensive evaluation index.Additionally,the pressure loss characteristics and discolouration response time of the camouflage film of the microfluidic visual camouflage system are tested,and an analysis is carried out to explain the reasons why the pressure loss is different for the camouflage film with the same number of color patches.