众多煤矿事故调查分析表明，生产过程中的人为失误始终是导致事故发生的首要原因。人的不安全状态是产生不安全行为的前提，对矿工不安全状态实现岗前识别和控制能够保证矿工在生产过程中处于良好的状态，从而最大程度上减少人为失误。近年来计算机模式识别技术的快速发展，为人的行为及状态信息的识别检测提供了新的思路和发展方向。但目前将该领域技术应用于矿工状态检测方面鲜有研究，生产实际中的应用很少。本文针对不安全状态中较为重要的不安全情绪状态开展研究，运用行为心理学、计算机模式识别技术等领域知识，建立了一种矿工岗前不安全情绪状态的辨识检测方法，旨在实现矿工不安全状态的早发现、早控制，从而为减少不安全行为及人为失误提供新的方法和技术支撑。本文的主要研究工作和结论如下：
      （1）矿工岗前不安全状态理论研究。通过对不安全行为等相关理论的分析，确定了矿工不安全状态的概念及内涵。以不安全状态中较为重要的不安全情绪状态为研究对象，通过对不安全情绪的产生机理、不安全行为的情绪作用进行分析，并依据基本情绪理论，将矿工不安全情绪划分为岗前不安全情绪状态和生产过程中不安全情绪状态，两者在情绪类型和产生机理上存在区别。确定了矿工岗前不安全情绪状态包括愤怒、悲伤、恐惧、厌恶四种。
     （2）矿工岗前不安全情绪状态实验研究。以注意力集中能力和反应能力作为不安全状态的测量指标，采用实验测量和主观情绪量表的方法采集了矿工在不同情绪下的注意力能力、反应能力指标以及身体动作姿态图片，并应用SPSS对实验数据进行分析处理。实验结果表明，四种不安全情绪状态对矿工注意力能力和反应能力存在负向影响，负向影响程度从低到高依次为厌恶、愤怒、恐惧、悲伤。依据不同情绪类型对矿工注意力和反应能力的影响程度，将矿工岗前情绪状态分为3级：1级为很差，包括悲伤、恐惧、愤怒三种情绪；2级为较差，包括厌恶情绪；3级为正常，包括中性和高兴情绪。
     （3）矿工岗前不安全情绪状态识别算法研究。通过人体建模将人体简化为骨骼关节点的形式，并应用Openpose算法完成了人体关节点提取。综合考虑岗前检测实际情况，对动作与情绪关系进行了分析，确定将重心、胸腔弯曲度、头部弯曲度作为表征情绪的身体动作姿态参数，并基于Python确立了三项参数的计算方法。通过对实验采集到的图片样本进行特征提取与处理，建立了情绪与动作姿态的映射关系。构建了基于人体动作识别的矿工岗前不安全情绪状态辨识算法，经实例验证整体准确率达到76%，可应用于矿工岗前情绪状态检测。

    The investigation and analysis of numerous coal mine accidents show that human error in the production process is always the primary cause of accidents. People's unsafe state is the premise of producing unsafe behavior, the identification and control of the unsafe state can ensure the miners to be in a good condition in the production process, so as to reduce human error to the greatest extent. In recent years, the rapid development of computer pattern recognition technology provides a new way of thinking and development direction for the identification and detection of human behavior and state information. However, there is little research on the application of this technology in miner condition detection, and it is rarely used in practice. This paper studied the relatively important unsafe emotional state in the unsafe state. This paper established an identification and detection method for the unsafe emotional state of miners before they work, using behavioral psychology, computer pattern recognition technology and other fields of knowledge, aiming at realizing the early detection and early control of the unsafe state of miners. So as to provide new methods and technical support for reducing unsafe behavior and human error. The main research work and conclusions of this paper are as follows:
(1) Study on Theory of Unsafe State of Miners Before Working.Through the analysis of unsafe act and other related theory, puts forward the concept of the unsafe state for miners: individuals due to external factors or their own factors, attention ability, reaction ability, risk perception, cognitive level, active ability and so on in a poor stability, lead to individual in poor stability, high sensitivity, easy to produce unsafe behavior of a situation. The research object is unsafe emotional state which is more important in unsafe state, through the mechanism of unsafe the mood, emotion effect of unsafe behavior is analyzed, and based on the theory of basic emotions, miners' unsafe emotions is divided into pre-work unsafe state and unsafe emotional state in the course of production, there are differences between emotion types and mechanism. The unsafe emotional states of miners before work were determined, including anger, sadness, fear and disgust.
(2) Experimental Study on Unsafe Emotional State of Miners Before Working. To miners pre-work unsafe emotional state, the ability of concentration and reaction were used as the measurement indexes of unsafe state, adopt the method of experimental measurements and subjective emotional scale collected subjects ability under different emotional attention, responsiveness index and body gesture images, used SPSS to analyze the experimental data processing. The experimental results show that four unsafe emotional states have negative influence on the attention ability and reaction ability of miners, and negative influence degree from low to high is disgust, anger, fear and sadness. Accroding to the influence degree of different emotional types on the attention and reaction ability of miners, the pre-work emotional state of miners was divided into three levels. Level 1 was very poor, including sadness, fear and anger. Level 2 is poor, including disgust. Level 3 is normal, including neutral and happy emotions.
(3) Research on Recognition Algorithm of Unsafe Emotional State of Miners Before Working. The human body is simplified into the form of bone nodes through human modeling, and human nodes are extracted by using Openpose algorithm. Taking the actual situation of pre-work inspection into comprehensive consideration, the relationship between movement and emotion was analyzed, the center of gravity, chest curvature and head curvature were taken as the body motion posture parameters to represent emotion, and the calculation method of the three parameters was established based on Python. Through the feature extraction and processing of the image samples collected in experiment, mapping relationship between emotion and gesture was established. Based on Python, the identification algorithm of miners' pre-work unsafe emotional state based on human motion recognition is constructed. Based on human action recognition, this paper constructs the recognition algorithm of miners' unsafe emotional state, and the overall accuracy reaches 76% through the example verification, which can be applied to the detection of miners' emotional state before work.

[1]Chen H, Liu J, Long R Y, et al. Research on 10-year Tendency of China Coal Mine Accidents and the Characteristics of Human Factors [J]. Safety Science, 2012, 50:745-750.

[2]柳茹林, 程卫民, 于岩斌等. 基于HFACS-MI改进模型的煤矿重特大事故人因分析[J]. 煤矿安全, 2017,48(08),250-253.

[3]沈剑,李红霞.基于扎根理论的主动安全型矿工素质特征及行为驱动因素研究[J]. 西安科技大学学报,2018,38(05):852-859.

[4]马亚静. 矿工情绪与安全问题研究[J]. 中国煤炭, 2007, 33(1): 60-62.

[5]唐健. 情绪与安全研究 [J]. 电力安全技术, 2013, 5(7): 8-10.

[6]李乃文, 郭利霞. 基于Multi-agent的矿工情绪稳定性模型构建[J]. 中国安全生产科学技术, 2014, 10(12):172-177.

[7]李焕. 情绪与不安全行为关系实验研究[D]. 西安科技大学. 2015.

[8]田水承, 杨鹏飞, 李磊等. 矿工不良情绪影响因素及干预对策研究[J]. 矿业安全与环保, 2016,43(6): 99-102.

[9]刘波. 安全情感文化对矿工不安全情绪的影响研究[D]. 西安科技大学, 2018.

[10]杨雪,杨娟,张瀚元. 基于系统动力学的矿工情绪演化机理研究[J].煤矿安全, 2021, 52(01):252-256.

[11]Thomson W M, Lawrence HP, Broadbent JM, et al. The impact of xerostomia on oral- health-related quality of life among younger adults[J]. Health&Quality of Life Out comes, 2006, 4(1): 86.

[12]代丽芳. 矿工情绪管理研究[D]. 辽宁工程技术大学, 2012.

[13]Waston D, Clark LA. Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 2013, 54(6): 1063-1073.

[14]惠璐. 不安全情绪对矿工风险决策影响的实验研究[D]. 西安科技大学, 2018.

[15]覃文波. 地铁施工不安全行为的情绪作用机理与实证研究[D]. 华中科技大学, 2019.

[16]李乃文, 姜秋敏. 矿工不安全心理的结构和测量[J]. 心理学探新, 2010, 30(3):91-96.

[17]栗继祖, 康立勋. 煤矿安全从业人员心理测试指标体系研究[J]. 安全与环境学报, 2005, 4(6):77-79.

[18]EUM Y. J., JANG E. H., PARK B. J., et al. Emotion recognition by ANS responses evoked by negative emotion[J]. Proceedings of the 2nd International Conference on Engineering and Industries, 2011:4 pp.-4 pp.

[19]CAPONECCHIA Carlo, SHEILS Ian. Perceptions of personal vulnerability to workplace hazards in the Australian construction industry[J]. Journal of Safety Research, 2011, 42(4): 253-258.

[20]李乾. 人脸情绪识别深度学习模型研究与应用[D]. 湘潭大学.2019.

[21]Ma L, Khorasani K. Facial expression recognition using constructive feed forward neural networks[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2004, 34(3):1588-1595.

[22]沈健. 基于脑电和人脸表情的双模态情绪识别系统[D]. 南京邮电大学, 2020.

[23]刘今越, 刘彦开, 贾晓辉, 郭士杰. 基于模型约束的人体姿态视觉识别算法研究[J].仪器仪表学报, 2020,41(04):208-217.

[24]D. A. Forsyth, M. M. Fleck. Body plans[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), San Juan, Puerto Rico, USA, 1997, 678-683.

[25]G.Mori, J.Malik. Estimating human body configurations using shape context matching[C]. European Conference on Computer Vision (ECCV), Copenhagen, Denmark, 2002, 666-680.

[26]X.Ren, A.C.Berg, J.Malik. Recovering human body configurations using pairwise constraints between parts[C]. IEEE International Conference on Computer Vision (ICCV), Beijing, China, 2005, 824-831.

[27]G. Hua, M.-H. Yang, Y. Wu. Learning to estimate human pose with data driven belief propagation[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), San Diego, CA,USA, 2005, 747-754.

[28] 冯晓月, 宋杰. 二维人体姿态估计研究进展[J]. 计算机科学, 2020, 47(11):128-136.

[29]M.A. Fischler, R. A. Elschlager. The representation and matching of pictorial structures[J]. IEEE Transactions on Computers, 1973, 100(1): 67-92.

[30]P. F. Felzenszwalb, D. P. Huttenlocher. Pictorial structures for object recognition[J]. International Journal of Computer Vision, 2005, 61(1): 55-79.

[31]M. Eichner, V. Ferrari, S. Zurich. Better appearance models for pictorial structures[C]. British Machine Vision Conference (BMVC), London, UK, 2009, 55.1-55.11.

[32]S. Johnson, M. Everingham. Combining discriminative appearance and segmentation cues for articulated human pose estimation[C]. Workshop of IEEE International Conference on Computer Vision (ICCV), Kyoto, Japan, 2009, 405-412.

[33]L. Bourdev, J. Malik. Poselets: Body part detectors trained using 3d human pose annotations[C].IEEE International Conference on Computer Vision (ICCV), Kyoto, Japan, 2009, 1365-1372.

[34]G. Gkioxari, P. Arbeláez, L. Bourdev, et al. Articulated pose estimation using discriminative armlet classifiers[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Portland, OR, USA, 2013, 3342-3349.

[35]L. Pishchulin, M. Andriluka, P. Gehler, et al. Poselet conditioned pictorial structures[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Portland, OR, USA, 2013, 588-595.

[36]彭帅, 黄宏博, 陈伟骏, 胡志颖, 袁铭阳. 基于卷积神经网络的人体姿态估计算法综述[J]. 北京信息科技大学学报(自然科学版), 2020, 35(03):59-66.

[37]Toshev A, Szegedy C. Deep Pose: Human pose estimation via deep neural networks[J]. [C]// Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 1653-1660.

[38]Wei S E, Ramakrishna V, Kanade T, et al. Convolutional pose machines [C]//Proceedings of the IEEE conference on Computer Vision and Pattern Recognition. 2016: 4724-4732.

[39]Umar I, Anton M. Pose Track: Joint multi-person pose estimation and tracking[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017: 2011-2020.

[40]Papandreou G , Zhu T , Kanazawa N , et al. Towards accurate multi-person pose estimation in the wild[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017: 4903-4911.

[41]Pishchulin L, Insafutdinov L, Tang L, et al. Deepcut: joint subset partition and labeling for multi person pose estimation[J]. Computer Vision and Pattern Recognition, 2016:34-50.

[42]Insafutdinov E, Andriluka M. Arttrack: articulated multi-person Tracking in the Wild[C] //Proceedings of the IEEE conference on Computer Vision and Pattern Recognition workshops, 2017:1293-1301.

[43]Zhe Cao. Realtime multi-person 2D pose estimation using part affinity fields[C] //Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition, 2017: 1302-1310.

[44]Coulson M. Attributing emotion to static body postures: Recognition accuracy, confusions, and viewpoint dependence[J]. Journal of Nonverbal Behavior, 2004, 28(2): 117-139.

[45]Vu H A, Yamazaki Y, Dong F, et al. Emotion recognition based on human gesture and speech information using RT middleware[C]//2011 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2011). IEEE, 2011: 787-791.

[46]Glowinski D, Camurri A, Volpe G, et al. Technique for automatic emotion recognition by body gesture analysis[C]//2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. IEEE, 2008: 1-6.

[47]Kessous L, Castellano G, Caridakis G. Multimodal emotion recognition in speech-based interaction using facial expression, body gesture and acoustic analysis[J]. Journal on Multimodal User Interfaces, 2010, 3(1-2): 33-48.

[48]Saha S, Datta S, Konar A, et al. A study on emotion recognition from body gestures using Kinect sensor[C]//2014 International Conference on Communication and Signal Processing. IEEE, 2014: 056-060.

[49]Glowinski D, Dael N, Camurri A, et al. Toward a minimal representation of affective gestures[J]. IEEE Transactions on Affective Computing, 2011, 2(2): 106-118.

[50]Gunes H, Piccardi M. Affect recognition from face and body: early fusion vs. late fusion[C]//2005 IEEE international conference on systems, man and cybernetics. IEEE, 2005: 3437-3443.

[51]Castellano G, Villalba S D, Camurri A. Recognising human emotions from body movement and gesture dynamics[C]//International Conference on Affective Computing and Intelligent Interaction. Springer, 2007: 71-82.

[52]Kosti R, Alvarez J M, Recasens A, et al. Emotion Recognition in Context[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2017.

[53]吴玉华. 矿井作业人员不安全行为特征规律分析[J]. 煤矿安全, 2011(12): 124-128.

[54]孟现飞, 丁恩杰, 刘全龙, 等. 危险源概念的重新界定及与隐患关系研究[J].中国安全科学学报, 2017, 27(04): 55-59.

[55]于言滔. 施工现场工人安全状态监测方法研究[D]. 清华大学, 2017.

[56]吴迪. 基于计算机视觉的施工人员安全状态监测技术研究[D]. 哈尔滨工业大学, 2019.

[57]Buck R. Prime theory: An integrated view of motivation and emotion[J]. Psychological Review, 1985, 92(92):389-413.

[58]Watson D, Tellegen A. Toward a consensual structure of mood[J]. Psychological Bulletin. 1985, 98(2): 219.

[59]Lazarus R S. Emotion and adaptation.[J]. Journal of Nervous & Mental Disease, 1991, 181(3):17-23.

[60]Campos J. Sicioemotional Development. Mussen P[J]. Handbook of Child Development, 1983(4):99-103.

[61]王振宏, 郭德俊, 刘海燕.情绪心理学[M].开明出版社, 2012.

[62]Lewis M, Haviland-Jones J M, Barrett L F. Handbook of Emotions, Third Edition[M]. Journal of Cognitive Psychotherapy, 2008.

[63]Plutchik R. Emotions, evolution,and adaptive processes:Feelings and emotions:the Loyola Symposium, 1970[C]. Academic Press.

[64]孔艳娜. 情绪研究方法[D]. 信阳师范学院, 2014.

[65]王璐. 基于脑电实验的情绪对矿工冒险行为影响研究[D]. 西安科技大学, 2016.

[66]田水承,郭方艺,杨鹏飞. 不良情绪对胶轮车驾驶员不安全行为的影响研究[J]. 矿业安全与环保, 2018,45(05):115-119+125.

[67]刘艳,刘鼎家,韩智攀. 基于动作识别的情绪提取方法研究[J]. 计算机工程, 2015, 41(05):300-305.

[68]王蔚,冯亚琴,杨再兴,王晓燕. 基于体感交互设备的人体重心计算方法[J]. 数据采集与处理, 2018,33(04):595-602.

[69]Jan V, Van D, Ruthger R, et al. Neural Correlates of Perceiving Emotional Faces and Bodies in Developmental Prosopagnosia: An Event-Related fMRI-Study[J]. PLoS ONE, 2008, 3(9):3195.

[70]Fredrickson B L . What Good Are Positive Emotions?[J]. Review of General Psychology, 1998, 2(3):300-319.