Versatile restoration and experimental verification for multi-wave band image of object detection

To solve the problem of restoration and clearness for multi-wave band images in object detection, a novel versatile restoration method was proposed, which was applicable for the restoration of multi-wave band images in object detection captured by imaging devices installed on missile, aerocraft or satellite. Some prior knolwedge was not needed in use of the proposed method about modes and models of degraded images. In this method, only some information of images was used, the minimization criterion function was constructed in regard to point spread function constrained by non-negative least square spatial smoothness, the optimization method was used to solve for the PSF, and the degraded images were restorated by non-blind image restoration method. The proposed method was verified by inputting a number of real images. The experimetal results show that the proposed method overcome some disadvantages of the exist image restoration methods such as non-versatility, good result to synthesized images, poor result to real images, time-consuming and so on. The proposed method can recover infrared, visible and millimeter-wave multi-wave band images without knowing the prior knolowdge. When inputing an image, the system output a clear image quickly. The proposed method was effective for the restoration of multi-wave band images in object detection system.