基于Kinect的哺乳期母猪姿态识别算法的研究
|
摘要
[目的]哺乳期母猪的姿态是其母性的外在表现,为监测母猪在哺乳期的哺乳行为,提出了一种基于Kinect的无接触式母猪姿态识别算法。[方法]使用Kinect 2.0采集位于限位栏内哺乳期小梅山母猪的深度数据。先通过姿态预分类将母猪的姿态分为站姿与卧姿,而后针对卧姿,使用基于DBSCAN(density-based spatial clustering of applications with noise)密度聚类算法计算母猪高度信息的簇数,通过比较簇的个数将卧姿分为侧卧与趴卧;针对站姿,使用基于脊背线提取的识别算法,将脊背线分成前后2段,通过比较前后2段脊背线的平均高度将站姿分为站立与坐立。[结果]比较人眼观察结果与算法识别结果,该算法识别站立、坐立的准确率分别为94.3%、92.6%,趴卧识别准确率为84.2%,侧卧姿态为93.7%。[结论]提出了一种无接触式的哺乳期母猪姿态识别算法,为母猪哺乳能力的评判与健康状况的分析提供了技术支持。
[Objectives]Sow posture is an important welfare indicator with regards to the maternal behavior. This paper proposes a non-contact sow posture recognition algorithm based on Kinect to monitor the maternal behavior of sows during the lactating period.[Methods]Kinect 2.0 was used to collect depth data from Xiaomeishan sows in a farrowing pen. Firstly,the posture of the sow was divided into two categories: standing and lying. Based on the DBSCAN( density-based spatial clustering of applications with noise)density clustering algorithm,the cluster number of the height information of the sow was calculated. The lying posture was grouped into lateral lying and the sternal lying depending on the number of the clusters. For standing posture,a recognition algorithm based on dorsal extraction was implemented,the back line was divided into two segments by the average of the height data of the two back lines,and standing posture was further divided into standing and sitting posture. [Results]Compared the results of algorithm recognition to human eye observation,the proposed system achieved an accuracy of 94.3% for standing,92.6% for sitting,84.2% for sternal lying and 93.7% for lateral lying. [Conclusions]In this paper,a non-contact behavior monitoring method was proposed,which can provide reliable technological support for the evaluation of sow health status and a stockman support system.
[Objectives]Sow posture is an important welfare indicator with regards to the maternal behavior. This paper proposes a non-contact sow posture recognition algorithm based on Kinect to monitor the maternal behavior of sows during the lactating period.[Methods]Kinect 2.0 was used to collect depth data from Xiaomeishan sows in a farrowing pen. Firstly,the posture of the sow was divided into two categories: standing and lying. Based on the DBSCAN( density-based spatial clustering of applications with noise)density clustering algorithm,the cluster number of the height information of the sow was calculated. The lying posture was grouped into lateral lying and the sternal lying depending on the number of the clusters. For standing posture,a recognition algorithm based on dorsal extraction was implemented,the back line was divided into two segments by the average of the height data of the two back lines,and standing posture was further divided into standing and sitting posture. [Results]Compared the results of algorithm recognition to human eye observation,the proposed system achieved an accuracy of 94.3% for standing,92.6% for sitting,84.2% for sternal lying and 93.7% for lateral lying. [Conclusions]In this paper,a non-contact behavior monitoring method was proposed,which can provide reliable technological support for the evaluation of sow health status and a stockman support system.
引文
[1]陶源栋,沈明霞,刘龙申,等.基于Kinect的母猪呼吸频率测定算法[J].南京农业大学学报,2017,40(5):921-927. DOI:10.7685/jnau.201701032.Tao Y D,Shen M X,Liu L S,et al. Study on measurement algorithm of sow respiratory frequency based on Kinect[J]. Journal of NanjingAgricultural University,2017,40(5):921-927(in Chinese with English abstract).
[2]郁厚安,高云,黎煊,等.动物行为监测的研究进展——以舍养商品猪为例[J].中国畜牧杂志,2015,51(20):66-70,75.Yu H A,Gao Y,Li X,et al. Research review of animal behavior monitoring technologies:commercial pigs as realistic example[J]. Chinese Journal of Animal Science,2015,51(20):66-70,75(in Chinese with English abstract).
[3]刘龙申,沈明霞,姚文,等.基于加速度传感器的母猪产前行为特征采集与分析[J].农业机械学报,2013,44(3):192-196,191.Liu L S,Shen M X,Yao W,et al. Acquisition and analysis of sows’behavior before farrowing based on acceleration sensor[J]. Transactions of the Chinese Society for Agricultural Machinery,2013,44(3):192-196,191(in Chinese with English abstract).
[4]汪开英,赵晓洋,何勇.畜禽行为及生理信息的无损监测技术研究进展[J].农业工程学报,2017,33(20):197-209.Wang K Y,Zhao X Y,He Y,Review on noninvasive monitoring technology of poultry behavior and physiological information[J]. Transactions of the Chinese Society of Agricultural Engineering,2017,33(20):197-209(in Chinese with English abstract).
[5] Cornou C,Lundbye-Christensen S. Classifying sows’activity types from acceleration patterns:an application of the Multi-Process Kalman Filter[J].Applied Animal Behaviour Science,2008,111(3/4):262-273.
[6]宋迪迪,王修远,屈敏直,等.基于SVM-HMM混合模型的生猪姿态识别方法研究[J].电子世界,2017(19):9-11.Song D D,Wang X Y,Qu M Z,et al. Research on pig posture identification method based on SVM-HMM hybrid model[J]. Electronics World,2017(19):9-11(in Chinese with English abstract).
[7]李哲,田建艳,郑晟,等.基于MPU6050和HMC5883L的猪的姿态检测[J].江苏农业科学,2016,44(2):434-437.Li Z,Tian J Y,Zhen S,et al. Posture detection of pigs based on MPU6050 and HMC5883L[J]. Jiangsu Agricultural Sciences,2016,44(2):434-437(in Chinese).
[8]闫丽,沈明霞,姚文,等.基于MPU6050传感器的哺乳期母猪姿态识别方法[J].农业机械学报,2015,46(5):279-285.Yan L,Shen M X,Yao W,et al. Recognition method of lactating sows’posture based on sensor MPU6050[J]. Transactions of the Chinese Society for Agricultural Machinery,2015,46(5):279-285(in Chinese with English abstract).
[9] Lao F,Brown-Brandl T,Stinn J P,et al. Automatic recognition of lactating sow behaviors through depth image processing[J]. Computers and Electronics in Agriculture,2016,125:56-62.
[10]张光跃,刘龙申,沈明霞,等.基于超声波的母猪产前行为监测系统设计[J].中国农业大学学报,2017,22(8):109-115.Zhang G Y,Liu L S,Shen M X,et al. Monitoring system design of sow’s antenatal behavior based on ultrasonic[J]. Journal of China Agricultural University,2017,22(8):109-115(in Chinese with English abstract).
[11] Yang L,Zhang L Y,Dong H W,et al. Evaluating and improving the depth accuracy of Kinect for Windows v2[J]. IEEE Sensors Journal,2015,15(8):4275-4285.
[12]邢军.小梅山猪生长发育性状的观察[J].中国畜禽种业,2007(10):35-37.Xing J. Observation of growth and development characters of Xiaomeishan pig[J]. The Chinese Livestock and Poultry Breeding,2007(10):35-37(in Chinese).
[13]邢军,陈军,吴井生,等.小梅山猪种质特性保存与利用的研究进展[J].中国猪业,2013,8(S1):79-81.Xing J,Chen J,Wu J S,et al. Research progress on preservation and utilization of germplasm properties of Xiaomeishan pig[J]. China Swine Industry,2013,8(S1):79-81(in Chinese).
[14]凌天星,丁威,陈军,等.小梅山猪种质资源的保护[J].中国畜禽种业,2006(5):42-45.Ling T X,Ding W,Chen J,et al. Protection of germplasm resources of Xiaomeishan pig[J]. The Chinese Livestock and Poultry Breeding,2006(5):42-45(in Chinese).
[15]闫丽,沈明霞,谢秋菊,等.哺乳母猪高危动作识别方法研究[J].农业机械学报,2016,47(1):266-272.Yan L,Shen M X,Xie Q J,et al. Research on recognition method of lactating sows’dangerous body movement[J]. Transactions of the Chinese Society for Agricultural Machinery,2016,47(1):266-272(in Chinese with English abstract).
[16] Lachat E,Macher H,Mittet M A,et al. First experiences with Kinect v2 sensor for close range 3D modelling[C]//The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences. Avila,Spain,2015.
[17] Song W B,Le A V,Yun S,et al. Depth completion for kinect v2 sensor[J]. Multimedia Tools&Applications,2017,76(3):4357-4380.
[18]伍育红.聚类算法综述[J].计算机科学,2015,42(6A):491-499,524.Wu Y H.General overview on clustering algorithms[J]. Computer Science,2015,42(6A):491-499,524(in Chinese with English abstract).
[19]金建国.聚类方法综述[J].计算机科学,2014,41(11A):288-293.Jin J G.Review of clustering method[J]. Computer Science,2014,41(11A):288-293(in Chinese with English abstract).
[20]李卓,杜晓冬,毛涛涛,等.基于深度图像的猪体尺检测系统[J].农业机械学报,2016,47(3):311-318.Li Z,Du X D,Mao T T,et al. Pig dimension detection system based on depth image[J]. Transactions of the Chinese Society for Agricultural Machinery,2016,47(3):311-318(in Chinese with English abstract).
[21]刘同海,滕光辉,付为森,等.基于机器视觉的猪体体尺测点提取算法与应用[J].农业工程学报,2013,29(2):161-168.Liu T H,Teng G H,Fu W S,et al. Extraction algorithms and applications of pig body size measurement points based on computer vision[J].Transactions of the Chinese Society of Agricultural Engineering,2013,29(2):161-168(in Chinese with English abstract).
[2]郁厚安,高云,黎煊,等.动物行为监测的研究进展——以舍养商品猪为例[J].中国畜牧杂志,2015,51(20):66-70,75.Yu H A,Gao Y,Li X,et al. Research review of animal behavior monitoring technologies:commercial pigs as realistic example[J]. Chinese Journal of Animal Science,2015,51(20):66-70,75(in Chinese with English abstract).
[3]刘龙申,沈明霞,姚文,等.基于加速度传感器的母猪产前行为特征采集与分析[J].农业机械学报,2013,44(3):192-196,191.Liu L S,Shen M X,Yao W,et al. Acquisition and analysis of sows’behavior before farrowing based on acceleration sensor[J]. Transactions of the Chinese Society for Agricultural Machinery,2013,44(3):192-196,191(in Chinese with English abstract).
[4]汪开英,赵晓洋,何勇.畜禽行为及生理信息的无损监测技术研究进展[J].农业工程学报,2017,33(20):197-209.Wang K Y,Zhao X Y,He Y,Review on noninvasive monitoring technology of poultry behavior and physiological information[J]. Transactions of the Chinese Society of Agricultural Engineering,2017,33(20):197-209(in Chinese with English abstract).
[5] Cornou C,Lundbye-Christensen S. Classifying sows’activity types from acceleration patterns:an application of the Multi-Process Kalman Filter[J].Applied Animal Behaviour Science,2008,111(3/4):262-273.
[6]宋迪迪,王修远,屈敏直,等.基于SVM-HMM混合模型的生猪姿态识别方法研究[J].电子世界,2017(19):9-11.Song D D,Wang X Y,Qu M Z,et al. Research on pig posture identification method based on SVM-HMM hybrid model[J]. Electronics World,2017(19):9-11(in Chinese with English abstract).
[7]李哲,田建艳,郑晟,等.基于MPU6050和HMC5883L的猪的姿态检测[J].江苏农业科学,2016,44(2):434-437.Li Z,Tian J Y,Zhen S,et al. Posture detection of pigs based on MPU6050 and HMC5883L[J]. Jiangsu Agricultural Sciences,2016,44(2):434-437(in Chinese).
[8]闫丽,沈明霞,姚文,等.基于MPU6050传感器的哺乳期母猪姿态识别方法[J].农业机械学报,2015,46(5):279-285.Yan L,Shen M X,Yao W,et al. Recognition method of lactating sows’posture based on sensor MPU6050[J]. Transactions of the Chinese Society for Agricultural Machinery,2015,46(5):279-285(in Chinese with English abstract).
[9] Lao F,Brown-Brandl T,Stinn J P,et al. Automatic recognition of lactating sow behaviors through depth image processing[J]. Computers and Electronics in Agriculture,2016,125:56-62.
[10]张光跃,刘龙申,沈明霞,等.基于超声波的母猪产前行为监测系统设计[J].中国农业大学学报,2017,22(8):109-115.Zhang G Y,Liu L S,Shen M X,et al. Monitoring system design of sow’s antenatal behavior based on ultrasonic[J]. Journal of China Agricultural University,2017,22(8):109-115(in Chinese with English abstract).
[11] Yang L,Zhang L Y,Dong H W,et al. Evaluating and improving the depth accuracy of Kinect for Windows v2[J]. IEEE Sensors Journal,2015,15(8):4275-4285.
[12]邢军.小梅山猪生长发育性状的观察[J].中国畜禽种业,2007(10):35-37.Xing J. Observation of growth and development characters of Xiaomeishan pig[J]. The Chinese Livestock and Poultry Breeding,2007(10):35-37(in Chinese).
[13]邢军,陈军,吴井生,等.小梅山猪种质特性保存与利用的研究进展[J].中国猪业,2013,8(S1):79-81.Xing J,Chen J,Wu J S,et al. Research progress on preservation and utilization of germplasm properties of Xiaomeishan pig[J]. China Swine Industry,2013,8(S1):79-81(in Chinese).
[14]凌天星,丁威,陈军,等.小梅山猪种质资源的保护[J].中国畜禽种业,2006(5):42-45.Ling T X,Ding W,Chen J,et al. Protection of germplasm resources of Xiaomeishan pig[J]. The Chinese Livestock and Poultry Breeding,2006(5):42-45(in Chinese).
[15]闫丽,沈明霞,谢秋菊,等.哺乳母猪高危动作识别方法研究[J].农业机械学报,2016,47(1):266-272.Yan L,Shen M X,Xie Q J,et al. Research on recognition method of lactating sows’dangerous body movement[J]. Transactions of the Chinese Society for Agricultural Machinery,2016,47(1):266-272(in Chinese with English abstract).
[16] Lachat E,Macher H,Mittet M A,et al. First experiences with Kinect v2 sensor for close range 3D modelling[C]//The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences. Avila,Spain,2015.
[17] Song W B,Le A V,Yun S,et al. Depth completion for kinect v2 sensor[J]. Multimedia Tools&Applications,2017,76(3):4357-4380.
[18]伍育红.聚类算法综述[J].计算机科学,2015,42(6A):491-499,524.Wu Y H.General overview on clustering algorithms[J]. Computer Science,2015,42(6A):491-499,524(in Chinese with English abstract).
[19]金建国.聚类方法综述[J].计算机科学,2014,41(11A):288-293.Jin J G.Review of clustering method[J]. Computer Science,2014,41(11A):288-293(in Chinese with English abstract).
[20]李卓,杜晓冬,毛涛涛,等.基于深度图像的猪体尺检测系统[J].农业机械学报,2016,47(3):311-318.Li Z,Du X D,Mao T T,et al. Pig dimension detection system based on depth image[J]. Transactions of the Chinese Society for Agricultural Machinery,2016,47(3):311-318(in Chinese with English abstract).
[21]刘同海,滕光辉,付为森,等.基于机器视觉的猪体体尺测点提取算法与应用[J].农业工程学报,2013,29(2):161-168.Liu T H,Teng G H,Fu W S,et al. Extraction algorithms and applications of pig body size measurement points based on computer vision[J].Transactions of the Chinese Society of Agricultural Engineering,2013,29(2):161-168(in Chinese with English abstract).