Abstract:AI and deep learning techniques are beginning to play an increasing role in astronomy as a necessary tool to deal with the data avalanche. Here we describe an application for finding resolved Planetary Nebulae (PNe) in crowded, wide-field, narrow-band H-alpha survey imagery in the Galactic plane. PNe are important to study late stage of stellar evolution of low to intermediate-mass stars. However, the confirmed ~3800 Galactic PNe fall far short of the numbers expected. Traditional visual searching for resolved PNe is time-consuming due to the large data size and areal coverage of modern astronomical surveys, especially those taken in narrow-band filters highlighting emission nebulae. To test and facilitate more objective, reproducible, efficient and reliable trawls for PNe candidates we have developed a new, deep learning algorithm. In this paper, we applied the algorithm to several H-alpha digital surveys (e.g. IPHAS and VPHAS+). The training and validation dataset was built with true PNe from the HASH database. After transfer learning, it was then applied to the VPHAS+ survey. We examined 979 out of 2284 survey fields with each survey field covering 1 * 1 deg^2. With a sample of 454 PNe from the IPHAS as our validation set, our algorithm correctly identified 444 of these objects (97.8%), with only 16 explicable 'false' positives. Our model returned ~20,000 detections, including 2637 known PNe and many other kinds of catalogued non-PNe such as HII regions. A total of 815 new high-quality PNe candidates were found, 31 of which were selected as top-quality targets for subsequent optical spectroscopic follow-up. Representative preliminary confirmatory spectroscopy results are presented here to demonstrate the effectiveness of our techniques with full details to be given in paper-II.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the problem of automatically identifying resolved planetary nebulae (PNe) in wide - field narrow - band H𝛼 survey data within the Galactic plane. Specifically, the paper attempts to solve the following key problems: 1. **Insufficient number of existing PNe**: - The currently confirmed number of Galactic planetary nebulae is approximately 3,800, which is far lower than the theoretically expected number (6,600 to 45,000). This indicates that traditional visual search methods are inefficient and incomplete. 2. **Limitations of traditional search methods**: - Traditional visual search methods are very time - consuming when dealing with large amounts of data and complex environments, and are easily influenced by subjective factors, resulting in low discovery efficiency and a high false - positive rate. - Automated image analysis techniques also have limitations when dealing with complex nebula morphologies and low - surface - brightness targets. 3. **Improving search efficiency and accuracy**: - The paper proposes a new deep - learning - based algorithm to screen out potential planetary nebula candidates from large - scale astronomical survey data in a more efficient, objective, and reliable manner. - This algorithm is particularly suitable for dealing with nebula targets in dense star fields, such as those in the galactic nucleus region. 4. **Verifying the effectiveness of the new algorithm**: - Researchers trained and verified the new algorithm using IPHAS and VPHAS+ survey data, demonstrating its high accuracy (97.8%) in identifying known PNe and discovering a large number of new high - quality PNe candidates. Through these efforts, the paper aims to develop a new method that can systematically detect and classify planetary nebulae, thereby providing more comprehensive and reliable sample data for astronomical research. ### Formulas and model structure To achieve the above goals, the paper proposes a Mask R - CNN model based on Swin - Transformer. The following are the key parts of the model and their formula representations: #### Feature extraction module The feature extraction module uses Swin - Transformer, which reduces computational complexity through window partitioning and shifted - window mechanisms. The specific steps are as follows: 1. **Input feature projection**: The input feature vector \( X \) is projected into three vectors through linear transformation: Query (\( Q \)), Key (\( K \)), and Value (\( V \)). \[ Q = XW_Q, \quad K = XW_K, \quad V = XW_V \] 2. **Attention calculation**: The correlation between Query and Key is calculated by dot product and normalized by the softmax function to obtain the attention map \( A \). \[ A = \text{softmax}\left(\frac{QK^T}{\sqrt{d_k}}\right) \] 3. **Output feature**: The final output feature is obtained by weighted summation. \[ Y = AV \] where \( d_k \) is the dimension of the Key vector, and \( W_Q \), \( W_K \), and \( W_V \) are learnable weight matrices. #### Multi - scale feature pyramid network (FPN) The multi - scale feature pyramid network is used to generate feature maps of different scales in order to better capture different details of the target. #### Region proposal network (RPN) The RPN module is used to generate bounding boxes that may contain the target. #### ROI Align The ROI Align module is used to extract the feature parts corresponding to the bounding boxes from the feature map. #### Classification and regression branches The classification branch is used to predict the target category, and the regression branch is used to precisely locate the target position. Through the collaborative work of these modules, the model can efficiently identify and classify planetary nebulae in complex astronomical images.

AI Techniques for Uncovering Resolved Planetary Nebula Candidates from Wide-field VPHAS+ Survey Data

First Discovery and Confirmation of PN Candidates Found from AI and Deep Learning Techniques Applied to VPHAS+ Survey Data

Hunting for extremely faint planetary nebulae in the SDSS spectroscopic database

Detection and Classification of Astronomical Targets with Deep Neural Networks in Wide Field Small Aperture Telescopes

Pulsar candidate identification using advanced transformer-based models

Deep Neural Network closed-loop with raw data for optical resident space object detection

Application of Deep Learning Methods Combined with Physical Background in Wide Field of View Imaging Atmospheric Cherenkov Telescopes

Prospects on high-energy source searches based on pattern recognition: Object detection in the H.E.S.S. Galactic Plane Survey and catalogue cross-matches

A Semi-Automated Computational Approach for Infrared Dark Cloud Localization: A Catalog of Infrared Dark Clouds

Identification of planetary and proto-planetary nebulae candidates through akari infrared photometry

PSF-based Analysis for Detecting Unresolved Wide Binaries

Searching for Quasi-Periodic Eruptions using Machine Learning

Discovering Ca II Absorption Lines With a Neural Network

Spitzer IRAC observations of newly-discovered planetary nebulae from the Macquarie-AAO-Strasbourg H-alpha Planetary Nebula Project

A Two-Stage Deep Learning Detection Classifier for the ATLAS Asteroid Survey

deep PACO: Combining statistical models with deep learning for exoplanet detection and characterization in direct imaging at high contrast

DeepGhostBusters: Using Mask R-CNN to Detect and Mask Ghosting and Scattered-Light Artifacts from Optical Survey Images

An Image Quality Evaluation and Masking Algorithm Based On Pretrained Deep Neural Networks

Point spread function estimation for wide field small aperture telescopes with deep neural networks and calibration data

A transient search using combined human and machine classifications

An Image Quality Evaluation and Masking Algorithm Based On Pre-trained Deep Neural Networks