Research On The Visible Light Environment Integration Algorithm And Intelligent Control System Of Ground Target

People have learned the static bionic camouflage design method from the creatures with color changing ability in nature,but when the target comes to a new scene,the integration effect of the original environment will be reduced,and even the risk of exposure will be increased.Focusing on the visible light reconnaissance means,thesis designs and implements an intelligent control system which can adaptively generate different textures with the change of background environment,dynamically display with electrochromic device array,and has the robustness of fog environment.First,considering that different textures can be adaptively generated in multiple scenes and multiple seasons,an algorithm for integrating the visible light environment of ground targets is designed.The YOLACT model is used to segment the target from the background environment,and then the generated mask image and the original image are input into the HIFILL image inpainting model,and the K-nearest neighbor algorithm is used to standardize the initially generated texture.The texture generated by the proposed method is compared with the texture generated by the traditional method,and the saliency detection is performed to verify the effectiveness of the proposed algorithm.Secondly,the drone is used as the image acquisition device and the electrochromic device array is used as the display device to realize the ground target visible light environment integration system.The system realizes the long-distance real-time transmission of the collected images from the drone to the laptop through the image transmission system and the image transmission software.After receiving the captured image,use the environment fusion algorithm to generate adaptive texture,and divide the rectangular area where the generated texture is located into multiple small rectangular areas,respectively identify the color in each small rectangular area.According to the identified color,Arduino sends different control commands to the drive circuit,and applies the specified voltage across the electrochromic device corresponding to the small rectangular area,so as to display the color consistent with the environment color.Finally,taking into account the problem of the reduction of the sharpness of the collected images in the foggy environment,a dark channel dehazing method based on bright area detection is proposed.The fog image is converted to the LAB color space,and the image are divided into light and dark areas based on the luminance component L,and the threshold value is obtained by the threshold regression network.The transmittance of the bright area is corrected and the initially restored image is enhanced in different areas.The linear enhancement method is used in the bright area,and the relative fog concentration weighted aggregation method is used in the dark area.