基于图像和IMU传感器的生物行为分析系统设计

doi:10.16180/j.cnki.issn1007-7820.2022.04.005

摘要/Abstract

摘要：

在小鼠行为学实验中,仅使用数字图像处理技术无法对小鼠行为进行全面检测和分析。针对这一问题,文中提出基于图像和IMU传感器结合的生物行为检测方法。该方法在Python编程环境下,利用数字图像处理技术对小鼠运动视频进行检测跟踪,获得小鼠的运动行为学参数。同时,使用本实验室设计的无线蓝牙微型IMU传感器测量小鼠的三轴加速度、三轴角速度、三轴磁力,对测量数据使用扩展卡尔曼滤波算法进行解算以获得小鼠的姿态等体态变化信息。实验结果表明,该系统能较好地自动分析出小鼠的运动行为信息和姿态信息,并能在生物行为分析系统界面上实时显示出行为信息。

关键词: 小鼠, 行为学实验, 数字图像处理技术, 运动行为学参数, IMU传感器, 扩展卡尔曼滤波, 姿态信息, 实时显示

Abstract:

In order to solve the problem that the digital image processing technology alone cannot detect and analyze the behavior of mice in behavioral experiments, a biological behavior detection method based on the combination of image and IMU sensor is proposed in this study. In Python programming environment, this method uses digital image processing technology to detect and track the motion video of mice, and obtain the motion behavior parameters of mice. Meanwhile, the wireless bluetooth micro IMU sensor designed by the laboratory is used to measure the three-axis acceleration, three-axis angular velocity and three-axis magnetic force of mice. The extended Kalman filter algorithm is used to solve the measured data to obtain the posture and other posture change information of mice. The experimental results show that the system can automatically analyze the motor behavior and posture information of mice, and can display the behavior information on the interface of biological behavior analysis system in real time.

Key words: mice, behavioral experiment, digital image processing technology, motor behavior parameters, IMU sensor, extended Kalman filtering, attitude information, real-time display

中图分类号:

TP212.9

杨孙运,阚秀. 基于图像和IMU传感器的生物行为分析系统设计[J]. 电子科技, 2022, 35(4): 28-34.

Sunyun YANG,Xiu KAN. Design of Biological Behavior Analysis System Based on Vision and IMU Sensors[J]. Electronic Science and Technology, 2022, 35(4): 28-34.

图/表 11

图1

图2

图3

图4

图5

图6

图7

表1

图8

图9

图10

参考文献 19

[1]	Weissbrod A, Shapiro A, Vasserman G, et al. Automated long-term tracking and social behavioural phenotyping of animal colonies within a semi-natural environment[J]. Nat Communications, 2013, 4(1):1-10.
[2]	刁泽浩. 自由活体动物的动态跟踪算法研究[D]. 保定:河北大学, 2018.
	Diao Zehao, Research on dynamic tracking algorithm for free living animals[D]. Baoding:Hebei University, 2018.
[3]	Hong W Z, Kennedy A, Zelikowsky M, et al. Automated measurement of mouse social behaviors using depth sensing, video tracking, and machine learning[J]. Proceedings of the National Academy of Sciences, 2015, 112(38):E5351-E5360.
[4]	Noldus L P, Spink A J, Tegelenbosch R A. EthoVision: a versatile video tracking system for automation of behavioral experiments[J]. Behavior Research Methods, Instruments, & Computers: A Journal of the Psychonomic Society, 2001, 33(3):398-414.
[5]	Rodriguez A, Zhang H Q, Klaminder J, et al. ToxId: an efficient algorithm to solve occlusions when tracking multiple animals[J]. Scientific Reports, 2017, 7(1):186-204. doi: 10.1038/s41598-017-00286-6
[6]	Altulea A H, Jalab H A, Ibrahim R W. Fractional holder mean-based image segmentation for mouse behavior analysis in conditional place preference test[J]. Signal Image and Video Processing, 2020, 14(1):135-142. doi: 10.1007/s11760-019-01533-1
[7]	Salem G, Krynitsky J, Hayes M, et al. Three-dimensional pose estimation for laboratory mouse from monocular images[J]. IEEE Transactions on Image Processing, 2019, 28(9):4273-4287. doi: 10.1109/TIP.2019.2908796
[8]	白桂峰. 基于MEMS惯性传感器的人体步态分析关键技术研究[D]. 太原:中北大学, 2020.
	Bai Guifeng. Research on key technologies of human gait analysis based on MEMS inertial sensor[D]. Taiyuan:North University of China, 2020.
[9]	郝福明. 基于微惯性传感器的生猪异常行为监测[D]. 太原:太原理工大学, 2018.
	Hao Fuming. Abnormal behavior monitoring of pigs based on micro inertial sensor[D]. Taiyuan:Taiyuan University of Technology, 2018.
[10]	倪力. 基于智能穿戴及信息融合的山羊健康状态识别[D]. 合肥:安徽农业大学, 2017.
	Ni Li. Goat health status identification based on smart wear and information fusion[D]. Hefei:Anhui Agricultural University, 2017.
[11]	Barnich O, Van Droogenbroeck M. ViBe: a universal background subtraction algorithm for video sequences[J]. IEEE Transactions on Image Processing, 2011, 20(6):1709-1724. doi: 10.1109/TIP.2010.2101613
[12]	王春江, 李鹏. 基于ZYNQ的运动目标检测系统设计[J]. 电子科技, 2020, 33(5):82-86.
	Wang Chunjiang, Li Peng. Design of moving target detection system based on ZYNQ[J]. Electronic Science and Technology, 2020, 33(5):82-86.
[13]	郭迎春, 杨飞飞, 师硕. 基于自适应的ViBe运动目标检测方法[J]. 控制工程, 2019, 26(9):1703-1711.
	Guo Yingchun, Yang Feifei, Shi Shuo. Moving object detection methods based on adaptive ViBe[J]. Control Engineering of China, 2019, 26(9):1703-1711.
[14]	王春丹, 谢红薇, 李亚旋, 等. 融合改进的三帧差分和ViBe算法的运动目标检测[J]. 计算机工程与应用, 2020, 56(13):199-203.
	Wang Chundan, Xie Hongwei, Li Yaxuan, et al. Motion object detection with improved three-frame difference and ViBe algorithm[J]. Computer Engineering and Applications, 2020, 56(13):199-203.
[15]	卢艳军, 陈雨荻, 张太宁, 等. 一种实用的四旋翼飞行器姿态融合算法研究[J]. 电光与控制, 2020, 27(8):84-89.
	Lu Yanjun, Chen Yudi, Zhang Taining, et al. A practical quadrotor attitude fusion algorithm[J]. Electronics and Control, 2020, 27(8):84-89.
[16]	米刚, 田增山, 金悦, 等. 基于MIMU和磁力计的姿态更新算法研究[J]. 传感技术学报, 2015, 28(1):43-48.
	Mi Gang, Tian Zengshan, Jin Yue, et al. MIMU update algorithm based on the posture and magnetometer[J]. Chinese Journal of Sensors and Actuators, 2015, 28(1):43-48.
[17]	秦永元. 惯性导航[M]. 北京: 科学出版社, 2006.
	Qin Yongyuan. Inertial navigation[M]. Beijing: Science Press, 2006.
[18]	Sabatini A M. Quaternion-based extended Kalman filter for determining orientation by inertial and magnetic sensing[J]. IEEE Transactions on Biomedical Engineering, 2006, 53(7):1346-1356. pmid: 16830938
[19]	石青. 基于位姿调节的模糊自适应EKF组合导航算法[J]. 电子设计工程, 2018, 26(10):79-83.
	Shi Qing. Fuzzy adaptive extended Kalman filter based on position and attitude factors in INS/GPS/CNS integrated navigation system[J]. Electronic Design Engineering, 2018, 26(10):79-83.

实验时间	运动距离 /像素	平均速度 /像素·s^-1	中央停留时间比例	周围停留时间比例
0~3 min	7 604	25.35	0.94%	99.06%
3~6 min	17 307	57.69	1.89%	98.11%
6~9 min	29 062	96.87	2.64%	97.36%