10 results about "Solar image" patented technology

The application discloses a kind of global light path adjusting devices and adjusting methods of correction lens group for solar imaging spectrometer, including correction lens group and adjustment shim, the front end of correction lens group is formed as first reference surface, the rear end of correction lens group is formed as second reference surface;The first reference surface of correction lens group is used as the reference of off-axis three-mirror optical system of global light path system of solar imaging spectrometer;The second reference surface of correction lens group is used as the reference of spectral scanning imaging system of global light path system of solar imaging spectrometer.The adjusting device and method of the application utilize the design of the form of self connection interface and the introduction of adjusting reference, reduce the difficulty of angle adjusting of lens group by means of external reference mirror in the prior art, can quickly and accurately connect the optical system of front end and rear end, ensure that the imaging quality of imaging spectrometer optical system is optimal.

The application relates to a device and a method for detecting whether the focusing of a tower type solar power generation system is accurate, which comprises a first detection body and a second detection body, the first detection body is provided with a first plane, a second plane and a third plane which can both reflect light and allow light to pass through, the first plane, the second plane and the third plane are arranged at 90 degrees and intersect at a point O; the second detection body is provided with a fourth plane which can both reflect light and allow light to pass through, and the fourth plane is arranged on the front side of the first detection body far away from the point O. The application fully utilizes the reflection and light transmission principle of light, the irradiation direction of the sunlight is fed back through the image of the sun, and whether the image of the sun is aligned with the collector of the tower type solar power generation system can directly judge whether the focusing of the heliostat is accurate, the detection is very intuitive and convenient, the detection time is greatly saved, and the efficiency of rectification is improved; and the application has the advantages of simple structure, low cost and convenient large-scale application.

The invention provides an irregular sun image center extraction method and system for sun imaging load accurate registration. The method comprises the steps of preliminarily determining the position of a sun image center; determining grid lines of the sun image; determining a sun image boundary area corresponding to each grid line; extracting pixel coordinates of intersection points of the grid lines and the sun contour; and eliminating outliers through iterative operation, and calculating the center of the sun image. The irregular sun image center extraction method provided by the invention can adapt to different sun forms, iterative operation is carried out based on multiple grid lines, the calculation result is stable, and the recognition precision is high.

The application discloses a kind of solar image feature enhancement method and system based on space-frequency domain feature learning.The method comprises: obtaining solar image training set;The image to be processed is input into the solar feature enhancement model based on space-frequency domain feature collaborative learning, and the feature enhancement result is obtained, the space-frequency domain collaborative enhancement loss function is constructed based on feature enhancement result and high-quality reference image, and the solar feature enhancement model is trained based on space-frequency domain collaborative enhancement loss function;Obtain the solar image to be tested, input into the solar feature enhancement model trained, and obtain the solar image after feature enhancement.The application can effectively improve the feature expression capability of solar image under complex imaging conditions, and has significant advantages in overall radiation consistency maintenance, small structure detail enhancement in solar surface area, noise suppression and quality improvement of structural degradation image.

The invention discloses a full-solar-surface sun image quality classification method based on deep learning, and the method comprises the steps: carrying out the classification of obtained full-solar-surface sun images, obtaining multi-quality-class full-solar-surface sun images with balanced proportions, and constructing a data set; performing zooming and center cutting operation on each image in the data set in sequence to obtain a data set under a preset size; randomly dividing the data set in the preset size into a training set, a verification set and a test set according to a preset proportion; constructing a deep learning model; carrying out training and hyper-parameter optimization on the deep learning model according to the training set and the verification set, and storing the model with the best performance on the test set as a full-solar-surface sun image quality classification model; and inputting the full-solar-surface sun image to be classified as a full-solar-surface sun image quality classification model to obtain a full-solar-surface sun image quality classification result. According to the method, the shot H alpha wave band full-solar-surface sun image is classified according to the cloud layer coverage degree with excellent performance, and subsequent manual screening is not needed due to the full-automatic processing mode and the high accuracy rate of the method.

The application provides a sun image-based star sensor installation correction method and system, which comprises the following steps: making a satellite continuously collect sun full-disk images in a steady-state sun-seeking mode; adopting a fitting method to perform edge detection and sun center extraction, and performing multi-frame averaging on the centroid coordinates; determining a satellite correction angle according to the sun full-disk images and the relationship between a sun space telescope phase plane coordinate system and a satellite body coordinate system, so as to ensure that the sun images are in the center of a field of view; and calculating a star sensor installation matrix correction value according to the satellite correction angle, so as to realize sun image-based star sensor installation correction. The application can meet the needs of future sun exploration missions, provides a good technical means for high-precision sun observation of a space science satellite, and has engineering application value.

The invention discloses a solar dynamics observation data pre-training method based on multi-scale contrast learning, and belongs to the technical field of intelligent image processing. The method comprises the following steps: preprocessing SDO observation data, and constructing a pre-training framework combining an auto-encoder and contrast learning, including reconstruction loss and three levels of contrast targets: class mark alignment of global time indexes, cross-modal patch alignment of consistent positions, and spatial discrimination constraints in images at the same moment; signature logarithm intensity transformation and synchronous geometric enhancement are adopted, CNN / ViT backbones are compatible, and multi-granularity characterization with an invariable output mode and physical perception is achieved. The model trained by the method can be used as a unified visual backbone of solar image analysis, supports channel translation and missing measurement interpolation, event classification and forecast, segmentation and retrieval and other applications, reduces the labeling and computing power requirements, improves the precision and robustness, and is suitable for multi-task migration and rapid landing deployment in solar physical research and space weather services.

The invention discloses a polarization / hypersensitivity / inertial navigation tight combination modeling and adaptive fusion method, and belongs to the technical field of bionic polarization integrated navigation. According to the method, polarized light intensity measured by a polarization sensor and sun image coordinates measured by a solar sensor are used as measurement and are subjected to tight combination modeling with an inertial measurement unit, and an expansion state equation including incident light intensity and degree of polarization and a nonlinear measurement equation are established. A self-adaptive fusion strategy is designed based on a dual credibility discrimination function of a polarization degree and a distance between a sun image coordinate and a sun-sensitive principal point coordinate, and a measurement mode of a polarization sensor, a sun-sensitive sensor or a combination of the polarization sensor and the sun-sensitive principal point is dynamically selected. And estimating a state quantity in real time through unscented Kalman filtering, and correcting a navigation error. The original light intensity information measured by the polarization sensor and the original sun image coordinate information measured by the solar sensor are directly used for measurement, so that the nonlinear transfer effect is inhibited. According to the invention, the orientation precision and environmental adaptability of the integrated navigation system in a complex environment are improved.

The invention discloses a solar activity multi-class event detection and classification method. The invention provides a method for automatically detecting and classifying multiple types of solar activity events by combining solar multi-modal image data and a deep learning method and mining spatial features of the multiple types of solar activity events. The method comprises the following steps: collecting a solar multi-modal image and preprocessing the solar multi-modal image; detecting a plurality of types of solar activity events by using a plurality of traditional solar image detection algorithms and generating labels; constructing a data set based on the multi-modal sun image and the label; establishing a solar activity multi-class event detection and classification model based on a deep learning mode, and carrying out learning training around a data set; and detecting and classifying the activity events involved in the input sun image, and outputting a detection and classification result. According to the invention, the sun activity events involved in the sun image can be effectively detected and classified, an end-to-end processing mode greatly reduces the degree of manual intervention, and an automatic operation mode improves the real-time processing capability for coping with massive sun observation data.

The application relates to the technical field of attitude measurement, in particular to a method and device for determining an attitude based on feature matching of a sun image, wherein the method comprises the following steps: obtaining a target sun image, then extracting sunspot features meeting preset stability conditions from the target sun image, combining bidirectional consistency constraints, geometric consistency constraints, local gray consistency constraints and sub-pixel correction constraints to obtain sunspot matching feature point pairs meeting preset matching conditions, then constructing unit direction observation vector sets of two groups of target sun images, constructing a centralized direction vector set, and generating a relative attitude change amount of a spacecraft corresponding to the target sun image. Thus, the problem that related technologies use other vector information for attitude compensation, which is large in calculation amount, insufficient in real-time performance and difficult to meet the high real-time performance and high robustness requirements of an embedded space platform is solved.