The detection and identification of impact craters on a planetary surface are crucially important for planetary studies and autonomous navigation. Crater detection refers to finding craters in a given image, whereas identification means to actually mapping them to particular reference craters. However, no method is available for simultaneously detecting and identifying craters with sufficient accuracy and robustness. Thus, this study proposes a novel end-to-end fully convolutional neural network (CNN), namely, CraterIDNet, which takes remotely sensed planetary images of any size as input and outputs detected crater positions, apparent diameters, and identification results. CraterIDNet comprises two pipelines, namely, crater detection pipeline (CDP) and crater identification pipeline (CIP). First, we propose a pre-trained model with high generalization performance for transfer learning. Then, anchor scale optimization and anchor density adjustment are proposed for CDP. In addition, multi-scale impact craters are detected simultaneously by using different feature maps with multi-scale receptive fields. These strategies considerably improve the detection performance of small craters. Furthermore, a grid pattern layer is proposed to generate grid patterns with rotation and scale invariance for CIP. The grid pattern integrates the distribution and scale information of nearby craters, which will remarkably improve identification robustness when combined with the CNN framework. We comprehensively evaluate CraterIDNet and present state-of-the-art crater detection and identification performance with a small network architecture (4 MB).

This study provides an optimal high-performance image processing algorithm for a miniaturized independent optical navigation sensor, which combines the functions of a star tracker and a navigation camera. This novel image processing algorithm is capable of extracting two different types of optical navigation measurements from a raw image. The aim is to simultaneously extract stars and target celestial body features with high accuracy and reliability to estimate observer-to-body relative position in subsequent navigation process. This paper presents star and celestial body imaging models and a novel slope edge model. We propose a high-performance algorithm to achieve the synchronous extraction of star and celestial body image features based on the aforementioned models. Double-window variance difference method is proposed to segment and classify stars and edge image regions of a celestial body with strong robustness. The sub-pixel level position of star centroid and celestial body edges are then simultaneously extracted by using the same operator on the basis of the consistency of the derivative distribution of star and celestial body edge profiles. The edge extraction deviation when using the slope edge model is also analyzed and compensated, and the accuracy of the celestial body edge extraction is improved to a higher level. The proposed algorithm has excellent feature extraction performance in terms of qualitative and quantitative measurements. This study has established a technical foundation for the development of the miniaturized independent optical navigation sensor, which is low cost, light weight and has flexible applicability due to its high accuracy and robustness.

The simultaneous extraction of optical navigation measurements from a target celestial body and star images is essential for autonomous optical navigation. Generally, a single optical navigation sensor cannot simultaneously image the target celestial body and stars well-exposed because their irradiance difference is generally large. Multi-sensor integration or complex image processing algorithms are commonly utilized to solve the said problem. This study analyzes and demonstrates the feasibility of simultaneously imaging the target celestial body and stars well-exposed within a single exposure through a single field of view (FOV) optical navigation sensor using the well capacity adjusting (WCA) scheme. First, the irradiance characteristics of the celestial body are analyzed. Then, the celestial body edge model and star spot imaging model are established when the WCA scheme is applied. Furthermore, the effect of exposure parameters on the accuracy of star centroiding and edge extraction is analyzed using the proposed model. Optimal exposure parameters are also derived by conducting Monte Carlo simulation to obtain the best performance of the navigation sensor. Finally, laboratorial and night sky experiments are performed to validate the correctness of the proposed model and optimal exposure parameters.