Android系统下OpenCV的人脸检测模块的设计  

Design of face detection module based on OpenCV in Android system

作　　者：公衍宇[1] 郭琦[1] 于超[2] 

机构地区：[1]河北工业大学信息工程学院,天津300380 [2]河北工业大学机械工程学院,天津300380

出　　处：《电子设计工程》2012年第20期52-54,共3页Electronic Design Engineering

摘　　要：针对解决OpenCV人脸检测模块在Android平台编译和移植的问题,提出一种利用JNI技术(Java Native Interface)调用OpenCV以及采用Android NDK(Native Development Kit)生成共享库的目标检测方法。文中从分析利用Android NDK编译Android平台所需要的OpenCV静态库的问题入手,详细阐述了利用JNI调用OpenCV相关函数的具体步骤。经过多次试验,证明该人脸检测模块的平均检测时间为1 280 ms,具有较高的检测速度和检测精度。Aiming at the problem of the OpenCV face detection module compiling and transplant in the Android platform,a JNI technology（Java Native-Interface） call OpenCV and Android NDK（Native Development Kit） to generate a shared library object detection method is proposed.Starting from analysis of compile the OpenCV static libraries needed by Android platform use Android NDK,Implementation steps of using JNI to call the relative functions of OpenCV are explained in detail.After numerous experiments to prove that the face detection module,the average detection time 1 280 ms,a high detection speed and detection accuracy.

关 键 词：ANDROID OPENCV 人脸检测 JNI NDK 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]