Intelligent detection method for small target in celestial body optical image

Abstract

The invention discloses an intelligent detection method for a celestial body optical image small target, and belongs to the technical field of astronomical image processing and target inspection. The method comprises the following steps: 1, reading and processing an FITS format file of a celestial body optical image, and converting the FITS format file into a format file; 2, training and test data set making; 3, adopting a Yov5 network model, and performing model training on the Yov5 network model by using the training set; 4, detecting the trained Yolov5 model by using the test set; and 5, inputting the acquired celestial body optical image into the model, and outputting a target detection frame and confidence. According to the method, the neural network model is established based on the Yolov5 target detection algorithm, and the neural network model is trained by using deep learning, so that a traditional method is replaced to carry out celestial body optical image target detection, simple and repetitive manual work is avoided, and the detection precision and speed are improved while automation and intelligentization are realized.

Description

technical field [0001] The invention belongs to the technical field of astronomical image processing and target inspection, and in particular relates to an intelligent detection method for small targets in astronomical optical images. Background technique [0002] In space science and exploration missions, there are urgent technical requirements for spacecraft artificial intelligence control, autonomous navigation and control, real-time on-orbit processing speed, storage data throughput, massive data processing, complex scientific calculations, and astronomical image processing. . [0003] Astronomical data are often in FITS format. FITS (Flexible Image Transport System) is a unified standard format for data transmission and exchange between observatories around the world determined by the International Astronomical Union (IAU) in 1982. It describes the definition of data and the general method of data encoding. It is a machine-independent, independent method using magneti...

AI Technical Summary

Problems solved by technology

For the sequence of star images obtained by the star observation camera, the brightness of the target under different images varies greatly, the background noise is large, and the target only occupies a few or a dozen pixels on the grayscale image, and no information such as texture, structure, and color can be used , and the target star point in the image will occasionally be accompanied by "smearing", the use of classic image processing and detection methods often does not work well

However, due to the complexity and large number of astronomical images, it is often necessary to complete a large number of simple repetitive tasks if they are analyzed and processed manually.

Images