职称：高级工程师/博士生导师

系所：射线所（EFM Group）

办公室：校本部东区环化楼6楼

E-mail：njulegao@163.com; gaosong@shu.edu.cn

（本课题组长期招收博士、博士后和科研助理）

个人简介

致力于环境人工智能与智能感知的交叉研究，聚焦大气污染来源解析与减污降碳环境效应，主要研究方向包括：多源大数据预测与污染溯源、机器视觉智能感知、具身智能与视嗅融合、管控决策智能体。革新VOCs在线监测技术，形成工业园区大气污染智慧监管新模式。主持或参与各类省部级课题10余项，现承担多项国家科技重大专项，负责机器视觉与智能感知。团队建有高性能GPU计算集群，支撑视觉模型训练与数据仿真，并开展生成式人工智能国际科研合作与技术开发。曾在上海市环境监测中心开展新型监测技术研究与标准化。入选上海环保系统专业技术领军人才，兼任VOCs专委会常委；任《环境科学研究》青年编委、《大气与环境光学学报》编委。在JHM、JCP等发表SCI论文30余篇，授权专利3项，制定标准10余项；牵头获上海市科技进步二等奖3项、生态环境部科技二等奖1项。长期担任Environment International、Journal of Cleaner Production、Environmental Technology & Innovation、Analytical Chemistry等期刊审稿人。ORCID：https://orcid.org/0000-0002-9627-1651。

研究方向

   1. 大气污染来源解析与减污降碳环境效应：VOCs、恶臭、温室气体等

   2. 新型环境监测技术研发应用与标准化：光学，传感器，质谱

   3. 环境人工智能与多源大数据挖掘：机器视觉、预测溯源、生成式AI

   4. 视嗅融合智能感知与管控决策智能体：机器狗溯源，路径规划

代表性中文论文

1. 基于受限记忆相关机制的红外光学气体成像视频气云分割[J]. 光学学报, 2026

2. 挥发性有机物无组织泄漏红外成像智能算法应用研究[J]. 中国环境监测, 2026

3. 基于人工智能的长三角地区工业园区甲烷排放特征与来源解析[J]. 环境科学研究, 2026

4. 恶臭气体监测技术研究进展[J]. 自然杂志, 2025

5. 我国重点区域环境大气VOCs监测体系现状及发展方向[J]. 环境科学研究, 2023

6. 工业园区VOCs光离子化气体检测技术适用性研究[J].环境科学研究，2023

7. 典型化工集中区环境空气SVOCs污染特征及来源解析[J].环境科学,2021

8. 合成树脂行业挥发性有机物排放成分谱及影响[J].中国环境科学,2020

9. 上海某石化园区周边区域VOCs污染特征及健康风险[J].环境科学,2018.

10. 工业区恶臭污染自动监控体系设计[J].中国环境监测,2018.

Name: Song Gao

Professional Title: Senior Engineer / Doctoral Supervisor

e-mail: njulegao@163.com；gaosong@shu.edu.cn

Research fields:

       1. Source Apportionment of Air Pollution and Environmental Effects of Pollution and Carbon Reduction: VOCs, Odors, Greenhouse Gases, etc.

2. Development, Application, and Standardization of Novel Environmental Monitoring Technologies: Optics, Sensors, Mass Spectrometry.

3. Environmental AI and Multi-Source Big Data Mining: Machine Vision, Prediction and Source Tracing, Generative AI.

4. Visual-Olfactory Fusion Intelligent Perception and Control Decision-Making Agent: Robot Dog-Based Source Tracing, Path Planning. (Our research group is continuously recruiting PhD students, postdoctoral fellows, and research assistants.)

Gao Song, Senior Engineer and Doctoral Supervisor. He completed his undergraduate and master's studies at Nanjing University and received his Ph.D. from Fudan University.

Dedicated to interdisciplinary research in environmental AI and intelligent perception, with a focus on air pollution source apportionment and the environmental effects of pollution and carbon reduction. Main research directions include multi-source big data prediction and pollution source tracing, machine vision-based intelligent perception, embodied visual–olfactory fusion, and environmental decision-making agent. He has developed key technologies in dynamic graph convolutional networks, spatiotemporal evolution modeling, and related areas. His work focuses on solving the challenges of trace VOCs and odor observation, innovating online NMHC monitoring technology, developing high-throughput reactive organic compound observation technology, and establishing a new model for intelligent online atmospheric supervision of VOCs pollution in industrial parks. He has led or participated in more than 10 major research projects and is currently undertaking several National Science and Technology Major Projects, responsible for machine vision and intelligent sensing. His team has built a high-performance GPU computing cluster to support vision model training and data simulation, and carries out international research collaboration and technology development. He formerly worked at the Shanghai Environmental Monitoring Center, conducting research on and standardization of new environmental monitoring technologies. He was selected as a leading professional and technical talent in Shanghai's environmental protection system and a "Three-Five" talent in environmental monitoring of the MEE. He concurrently serves as a standing committee member of the Professional Committee on Prevention and Control of VOCs Pollution. He serves as a young editorial board member of Research of Environmental Sciences and an editorial board member of the Journal of Atmospheric and Environmental Optics. He teaches the graduate professional practice base course "Environmental Big Data and Artificial Intelligence" and the doctoral course "Introduction to AI Applications in the Environment." He has published more than 30 SCI papers in journals such as JHM and JCP, holds 3 granted patents, and has contributed to the formulation of more than 10 standards. He has received the Second Prize of the Shanghai Science and Technology Progress Award three times and the Second Prize of the Science and Technology Award of the Ministry of Ecology and Environment once. He has long served as a reviewer for journals such as Environment International, Journal of Cleaner Production, Environmental Technology & Innovation, and Analytical Chemistry. ORCID: https://orcid.org/0000-0002-9627-1651.

RESEARCH AND DEVELOPMENT INTERESTS

1. Big Data Analytics, Artificial Intelligence, and Machine Vision

We apply supervised and unsupervised learning to mine environmental data, assess data quality, and support environmental management. Our machine vision research uses infrared imaging, deep learning, graph convolutional networks, and spatiotemporal modeling to detect, localize, and quantify gas clouds, concentrations, and emission rates in complex backgrounds. It also enables small-target recognition for smoking detection and rapid warning. In Hangzhou Bay, we developed an integrated system for air pollutant prediction and pollution-source transport identification, combining big data, machine learning, online monitoring, and machine vision. We further conduct deep-learning-based time-series and regional ozone prediction and extend this vision framework to AI gas-cloud imaging for exhaust emission estimation.

2. Novel Environmental Monitoring Technologies and Standardization

We have innovatively established an online VOCs monitoring system for industrial parks, promoting the application of sensors, chromatography, mass spectrometry, and optics in VOC monitoring. We have made significant breakthroughs in monitoring technology research and promoted the upgrading of monitoring methods. We also apply AI-enabled monitoring technology to identify environmental pollution problems. Our research includes new gas detection technologies, such as electronic noses and infrared imaging, with the aim of promoting technology application and standardization through intelligent recognition.

3. Gaseous Pollution: Machine Vision-Based Identification of Fugitive Emissions

We apply machine vision, including infrared imaging, deep learning, and spatiotemporal modeling, to identify fugitive atmospheric emissions in complex industrial environments. Our research focuses on detecting, recognizing, localizing, and quantifying gas clouds, concentrations, and emission rates, thereby enabling the identification and quantification of fugitive VOCs and odorous emissions.

4. Visual–Olfactory Fusion Intelligent Perception and Decision-Making Agents

We develop embodied visual–olfactory perception and decision-making agents for pollution source tracing and control. Research includes robot dog-based source tracing, AI-based observation path optimization, and autonomous path planning, enabling intelligent monitoring route scheduling and adaptive decision-making.

RESEARCH EXPERIENCE

1. 2026, Machine Vision for Intelligent Identification and Quantification of VOC Leakage Using OGI

2. 2026, AI-Agent-Driven Dynamic Optimization and Effectiveness Verification for Emergency Support and Control. NSTMP

3. 2025, Research and Development of Intelligent Identification Technologies and Equipment Based on AI and Machine Vision for Environmental Monitoring in Key Industries. NSTMP

4. 2025, AI-Driven Dynamic Perception Systems for Pollution Sources in Industrial Clusters and Smart Law Enforcement and Supervision. NSTMP

5. 2025, AI reshapes a green future. SIEP,

6. 2022, VOCs LiDAR and Portable Mass Spectrometry Monitoring Technology Research and Development and Application Demonstration. National Science and Technology Major Project (NSTMP)

7. 2022, Principal investigator for a major municipal project on online perception and intelligent early warning of hospital epidemic risks. SIEP.

8. 2017, Emergency Early Warning Evaluation Technology and Demonstration Research for Sudden Air Pollution Accidents. NSTMP.

9. 2016, Development and application of online measurement system for ambient atmospheric organic matter based on multi-ion source time-of-flight mass spectrometry technology.

10. 2014, National Environmental Protection Public Welfare Scientific Research "Research on the Emission Characteristics and Control Countermeasures of VOCs in Typical Chemical Parks".

RESEARCH AWARDS

1.2023, "Research and Application of key Technologies for waste Gas Collection, purification, Monitoring and Control in Rubber factories" won the second prize for scientific and technological progress in Shanghai.

2. 2020, won the second prize of Shanghai Science and Technology Progress Award as the second author of "Key Technologies and Applications of Intelligent Supervision and Traceability of Odor Pollution in Industrial Parks"

3. 2019, won the second prize of Environmental Protection Science and Technology Award of the Ministry of Ecology and Environment as the first author of "Research and Application of Key Technology for On-line Monitoring of Air Feature Pollution in Chemical Concentration Areas"

4. 2018, the first "National VOCs Monitoring and Governance Innovation Achievements" outstanding youth for innovative scientific and technological achievements

5. 2017, Technical leader in environmental protection system in Shanghai

6. 2008, "Establishment of Shanghai Ambient Air Quality Prediction and Forecast System and its Application in High Pollution Day Early Warning Linkage" and won the second prize of Shanghai Science and Technology Progress Award

PUBLICATION

1. Insight into VOCs source profiles by machine learning: Role of commonalities in synergistic pollution controls, Journal of Hazardous Materials.

2. Data-driven machine learning quantifies ozone transport in the Hangzhou Bay urban cluster, Frontiers of Environmental Science & Engineering.

3. Calibration innovations to enhance the accuracy of proton-transfer-reaction mass spectrometry for VOCs measurements, Atmospheric Environment.

4. Identification of atmospheric emerging contaminants from industrial emissions: A case study of halogenated hydrocarbons emitted by the pharmaceutical industry, Environment International.

5. Prediction and explanation for ozone variability using cross-stacked ensemble learning model, Science of The Total Environment.

6. Multi-Scenario Validation and Assessment of a Particulate Matter Sensor Monitor Optimized by Machine Learning Methods, Sensors.

7. Obtaining accurate non-methane hydrocarbon data for ambient air in urban areas: comparison of non-methane hydrocarbon data between indirect and direct methods, Atmos. Meas. Tech., 16,5709-5723.

8. Role of garbage classification in air pollution improvement of a municipal solid waste disposal base, Journal of Cleaner Production, 423(2023),138737.

9. Prediction and cause investigation of ozone based on a double-stage attention mechanism recurrent neural network[J]. Front. Environ. Sci. Eng. 2023, 17(2): 21.

10. New understanding of source profiles: Example of the coating industry[J]. Journal of Cleaner Production, 357(2022):132025.

STANDARD

National Standard

1. Calibration Regulations for Continuous Automatic Monitoring System of Non-Methane Total Hydrocarbons in Ambient Air.

2. Technical Specification for Automatic Monitoring of Odor Pollutants in Ambient Air

3. Method for determination of performance of industrial organic waste gas purification devices(GB/T 40200-2021)

Local Standard and Group Standard

1. Criteria for evaluating air quality in indoor public places for children aged 0-6（T/SICCA 018—2023）

2. Technical requirements and monitoring specifications of infrared Optical Gas Imager(OGI) for VOC leakage detection（T/ACEF 095—2023）

3. Technical requirements and monitoring specifications of portable photoionization detector(PID) for VOCs（T/ACEF 096—2023）

4. Technical Specifications for Grid Monitoring System for VOCs in the Yangtze River Delta Industrial Park (DB31/T1098-2022)

5. Technical Specifications for Navigation Monitoring of VOCs in the Yangtze River Delta Ecological Green Integrated Development Demonstration Zone (DB31/T 310002-2021)

6. Technical Specifications for Online Monitoring of Organic Sulfur in Ambient Air (DB31/T1089-2018).

7. Technical Specifications for Online Monitoring of Non-Methane Total Hydrocarbons in Ambient Air (DB31/T1090-2018).

8. Odor Pollutant Emission Standards(DB311025-2016)