The special geographical location of the polar regions increases the difficulty of modeling surface emissivity, thus the physical retrieval algorithms of the temperature and humidity profiles for microwave radiometers mainly focus on the regions between 60°S and 60°N. In this paper, the deep neural networks (DNN) and long short-term memory (LSTM) models are first implemented to retrieve atmospheric temperature and humidity profiles in real time from FY-3D/MWHTS in Arctic regions and are compared with the physical retrieval algorithm. The hyperparameters of the machine learning models are determined using the grid search and 10-fold cross-validation. Results show that, compared with the physical retrieval algorithm, the retrieval accuracies of the atmospheric temperature and humidity profiles of the DNN and LSTM models in June 2021 are higher over sea ice, and the maximum retrieval accuracies are improved by about 3.5 K and 42%. Over land, the retrieval accuracies of the atmospheric temperature profiles for the DNN and LSTM models in June 2021 are improved by about 5 K. The retrieved humidity results for these two models are not compared with the physical retrieval algorithm, which fails for the humidity profile retrieval over land. In addition, the retrieval results of the DNN-based and LSTM-based models using the independent validation data in February, April, and September are also evaluated over different surface types. The RMSEs of the retrieved temperature profiles for the two models are within 4 K, except for the near-surface, and the humidity profiles are within 25%, except for in February. The temperature profiles in September and the humidity profiles in February are somewhat reduced compared to other months because of the highly variable emissivity properties in autumn and winter. Overall results show that the machine learning method can well-evaluate the retrieval capability of FY-3D/MWHTS of the atmospheric temperature and humidity profiles in Arctic regions.
In response to the diverse types and large number of PCB surface defects, our paper proposes an improved YOLOv8-based method for PCB surface defect detection. First, a lightweight modification is performed by introducing RepGhostBottleNeck as the lightweight backbone network, which reduces the number of parameters in the training model. It should be noted that the term “lightweight” in this paper is relative to the original YOLOv8L baseline. Compared with extremely lightweight detectors, the model in this paper places greater emphasis on the balance between accuracy and efficiency. Additionally, an attention mechanism module and a small object detection head module are added to the backbone network. Furthermore, the loss function of the network is improved. Experimental results show that the improved model achieves an average mAP@0.5 of 0.976, demonstrating high-precision detection on the constructed dataset.
The statistical retrieval of atmospheric parameters will be greatly affected by the accuracy of the simulated brightness temperatures (BTs) derived from the radiative transfer model. However, it is challenging to further improve a physical-based radiative transfer model (RTM) developed based on the physical mechanisms of wave transmission through the atmosphere. We develop a deep neural network-based RTM (DNN-based RTM) to calculate the simulated BTs for the Microwave Temperature Sounder-II onboard the Fengyun-3D satellite under different weather conditions. The DNN-based RTM is compared in detail with the physical-based RTM in retrieving the atmospheric temperature profiles by the statistical retrieval scheme. Compared to the physical-based RTM, the DNN-based RTM can obtain higher accuracy for simulated BTs and enables the statistical retrieval scheme to achieve higher accuracy in temperature profile retrieval in clear, cloudy, and rainy sky conditions. Due to its ability to simulate microwave observations more accurately, the DNN-based RTM is valuable for the theoretical study of microwave remote sensing and the application of passive microwave observations.
The commonly used reference atmospheric profiles for the validation of retrieved atmospheric profiles for microwave sounders have bias compared with real atmospheric profile values, which is detrimental to the validation of the retrieval. Microwave sounder observations are the direct measurements of microwave radiation in atmospheric conditions and are a true representation of the status of the atmosphere. This paper proposed a validation method for the retrieved atmospheric temperature and atmospheric humidity profiles of the satellite-based microwave sounder using its own in-orbit observations. The validation experiments are performed both for the retrievals of the microwave temperature sounder-II (Xi’an Branch, China Academy of Space Technology, Xi’an, China. MWTS-II) and the microwave humidity and temperature sounder (National Space Science Center, Chinese Academy of Sciences, Beijing, China. MWHTS). The validation results show that the retrieved temperature profiles of MWTS-II have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure range of 3–30 hPa, and the accuracy of the rest of the pressure range is comparable between the profiles of ERA5 and the retrieved profiles. And the retrieved temperature profiles of MWHTS have higher accuracy compared to the temperature profiles of ERA5 in the atmospheric pressure level around 50 hPa and lower accuracy in the rest of the pressure levels. In addition, the retrieved humidity profiles of MWHTS have higher accuracy compared to the humidity profiles of ERA5 in the atmospheric pressure range of 350–925 hPa. The proposed validation method for the retrieved atmospheric temperature and atmospheric humidity profiles of MWHTS using its own observations is promising for improving the feasibility and reliability of the validation, and can be a good reference for the application of the satellite in-orbit observations and the optimization of the microwave sounders.
1 September 2020 in Journal of Physics: Conference Series
In order to improve the detection ability of chaotic system to detect weak periodic signal frequency, phase and amplitude, a chaotic system based on multiple correlation and phase-locked loop is established. Firstly, the signal to be measured is input into the detection frequency of chaotic system. Also, the reference signal of multiple correlation technology is constructed, and the signal to be tested is processed by multiple cross-correlation. Then, the signal processed by multiple correlation is input into the phase-locked loop to obtain the phase of the signal to be measured. Finally, the signal to be measured is input into the chaotic system again, and the amplitude of the signal to be measured is detected according to the change of the chaotic oscillator in critical periodic state and large-scale periodic state. The experimental results show that the detection system can detect the frequency, phase and amplitude of weak periodic signal, which is easy to operate and has high detection ability.
1 September 2020 in Journal of Physics: Conference Series
The selection of clear sky data in space-borne remote sensing data is very important for its data application. For FY-3D satellite microwave humidity and temperature sounder (MWHTS), an inversion system of atmospheric cloud water content by MWHTS is established based on neural network. The cloud water content inversion value is used to select clear sky data from MWHTS observation data. The experimental results show that FY-3D/MWHTS clear sky data selection method based on neural network can effectively select MWHTS observation data, thus improving the simulation brightness temperatures accuracy of MWHTS by radiative transfer model. This method can be used to select clear sky data by using space-borne observation data itself. It is easy to operate and has important practical value for climate change research, numerical weather forecast, etc., based on space-borne observation data.