A New Navigation Technique Based On The Visual Mechanism Of Flying Insects

In order to overcome some disadvantages in the present navigation methods, designing and developing new navigation approaches, especially learning from visual navigation mechanism of insects is becoming the new point and perspective in the future research of the navigation technique. This paper has proposed the concepts of entropic map and entropy flow from physics idea, developed from the deeply research in the flying insects'visual mechanism. And then, a new navigation algorithm based on entropy flow has been presented in this paper.Firstly, integrating to the visual mechanism of flying insects with the concept of entropy in physics, the concept of entropic map is established in this paper. In removing white noise, pepper & salt noise and speckle noise, the denoising performance of entropic map is compared with that of mean filter, median filter, Wiener filter and Gauss filter quantificationally and qualitatively.Secondly, according to the concept of entropic map, the concept of entropy flow is established, which is used to describe image motion. The entropy flow constraint equation is established, similar to the optical flow constraint equation. In order to calculate entropy flow, many different constant assumptions are presented, which are used to construct different energy functionals. Many different computation approaches of entropy flow are compared qualitatively and quantitatively.Thirdly, the geometric model of camera, that is the six parameter model of orthogonal transformation, is chosen as the orthogonal projection model, according to the research background. And then the six-parameter affine transformation model based on entropy flow is proposed in this paper. In order to reduce the computation and improve the accuracy of global motion estimation, the automatic evaluation threshold selection algorithm of entropy flow data is presented in this paper.Fourthly, the translational motion on the direction of X-axis, Y-axis and Z-axis are considered only in this paper. The navigation algorithm based on entropy flow and Kalman filter is proposed in this paper. The acquirable method of experimental data and the performable approach of simulation are discussed. The navigation performance of the presented algorithm, and the relationship between the initial error and the circular error probability(CEP) are discussed in this paper. The possibility of superimposing a matching process before using the presented navigation method is discussed as well. The simulation results suggest that the navigation algorithm can perform real-time rectification of the missile's trajectory well, and can reduce the cost of the missile's hardware. The new navigation method is an optional solution to the future research of navigation system.In the last, similar to the entropic map, the self-information map based on Parzen window is created in this paper. The small target detection algorithm based on the self-information map is also presented in this paper. In that detection algorithm, the segmentation algorithm combining with the adaptive threshold method and region growing technique is proposed to segment small targets. The fast algorithm of local entropy is explored in this paper as well. The local reverse entropic map is built in this paper, and its construction is similar to the entropic map. The small target detection algorithm based on the local reverse entropic map is proposed in this paper.