package com.dahua.netsdk.lib.callback.impl;
|
|
import com.dahua.netsdk.lib.NetSDKLib;
|
import com.dahua.netsdk.lib.ToolKits;
|
import com.dahua.netsdk.lib.Utils;
|
import com.sun.jna.Pointer;
|
|
import java.io.File;
|
import java.nio.charset.Charset;
|
|
/**
|
* @author ： 47040
|
* @since ： Created in 2020/7/18 10:51
|
*/
|
public class DefaultAnalyseTaskResultCallBack implements NetSDKLib.fAnalyseTaskResultCallBack {
|
|
private static DefaultAnalyseTaskResultCallBack singleInstance;
|
private File picturePath;
|
private static NetSDKLib.DEV_EVENT_FACERECOGNITION_INFO msg=new NetSDKLib.DEV_EVENT_FACERECOGNITION_INFO();
|
|
private DefaultAnalyseTaskResultCallBack() {
|
picturePath = new File("./AnalyzerPicture/");
|
if (!picturePath.exists()) {
|
picturePath.mkdir();
|
}
|
}
|
|
public static DefaultAnalyseTaskResultCallBack getSingleInstance() {
|
if (singleInstance == null) {
|
singleInstance = new DefaultAnalyseTaskResultCallBack();
|
}
|
return singleInstance;
|
}
|
|
@Override
|
public int invoke(NetSDKLib.LLong lAttachHandle, Pointer pstAnalyseTaskResult, Pointer pBuf, int dwBufSize, Pointer dwUser) {
|
long start=System.currentTimeMillis();
|
NetSDKLib.NET_CB_ANALYSE_TASK_RESULT_INFO task = new NetSDKLib.NET_CB_ANALYSE_TASK_RESULT_INFO();
|
ToolKits.GetPointerData(pstAnalyseTaskResult, task); // 从指针获取智能事件分析信息
|
|
for (int i = 0; i < task.nTaskResultNum; i++) {
|
// nTaskID 和主动推送时的一一对应
|
System.out.println(task.stuTaskResultInfos[i].nTaskID);
|
// 实际的事件个数
|
for (int j = 0; j < task.stuTaskResultInfos[i].nEventCount; j++) {
|
NetSDKLib.NET_SECONDARY_ANALYSE_EVENT_INFO info = task.stuTaskResultInfos[i].stuEventInfos[j];
|
// 事件类型 对应 EM_ANALYSE_EVENT_TYPE
|
System.out.println("type:" + info.emEventType);
|
switch (info.emEventType) {
|
//特征提取
|
case NetSDKLib.EM_ANALYSE_EVENT_TYPE.EM_ANALYSE_EVENT_FEATURE_ABSTRACT: {
|
NetSDKLib.DEV_EVENT_FEATURE_ABSTRACT_INFO stInfo = new NetSDKLib.DEV_EVENT_FEATURE_ABSTRACT_INFO();
|
ToolKits.GetPointerDataToStruct(info.pstEventInfo, 0, stInfo);
|
|
System.out.println("nChannelID:" + stInfo.nChannelID); // 通道号
|
System.out.println("nAction:" + stInfo.nAction); // 0:脉冲 1:开始 2:停止
|
System.out.println("emClassType:" + stInfo.emClassType); // 智能事件所属大类 参考 EM_CLASS_TYPE
|
|
for (int k = 0; k < stInfo.nFeatureNum; k++) {
|
// 错误码、特征版本号
|
System.out.println("FeatureVectorList[" + k + "].FeatureErrCode:" + stInfo.stuFeatureVectorList[k].emFeatureErrCode); // 错误码
|
System.out.println("FeatureVectorList[" + k + "].szFeatureVersion:" + new String(stInfo.stuFeatureVectorList[k].szFeatureVersion).trim()); // 特征版本版本号
|
// 这里的特征是设备用于比对的二进制数据，不是图片，具体内容请咨询设备研发
|
System.out.println("FeatureVectorList[" + k + "].stuFeatureVector.dwOffset:" + stInfo.stuFeatureVectorList[k].stuFeatureVector.dwOffset); // 人脸小图特征值在二进制数据块中的偏移
|
System.out.println("FeatureVectorList[" + k + "].stuFeatureVector.dwLength:" + stInfo.stuFeatureVectorList[k].stuFeatureVector.dwLength); // 人脸小图特征值长度，单位:字节
|
// 人脸抓拍角度、质量数据
|
System.out.println("FeatureVectorList[" + k + "].stuFaceAttribute.nAngle1:" + stInfo.stuFeatureVectorList[k].stuFaceAttribute.nAngle[0]); // 人脸抓拍角度,三个角度分别是：仰俯角,偏航角,翻滚角；默认值[999,999,999]表示无此数据
|
System.out.println("FeatureVectorList[" + k + "].stuFaceAttribute.nAngle2:" + stInfo.stuFeatureVectorList[k].stuFaceAttribute.nAngle[1]); // 偏航角
|
System.out.println("FeatureVectorList[" + k + "].stuFaceAttribute.nAngle3:" + stInfo.stuFeatureVectorList[k].stuFaceAttribute.nAngle[2]); // 翻滚角
|
System.out.println("FeatureVectorList[" + k + "].stuFeatureVector.nFaceAlignScore:" + stInfo.stuFeatureVectorList[k].stuFaceAttribute.nFaceAlignScore); // 人脸对齐得分分数,取值范围 0~10000，-1为无效值
|
System.out.println("FeatureVectorList[" + k + "].stuFeatureVector.nFaceQuality:" + stInfo.stuFeatureVectorList[k].stuFaceAttribute.nFaceQuality); // 人脸抓拍质量分数,取值范围 0~10000
|
}
|
break;
|
}
|
//目标识别
|
case NetSDKLib.EM_ANALYSE_EVENT_TYPE.EM_ANALYSE_EVENT_FACE_RECOGNITION: {
|
/*NetSDKLib.DEV_EVENT_FACERECOGNITION_INFO stInfo = new NetSDKLib.DEV_EVENT_FACERECOGNITION_INFO();
|
ToolKits.GetPointerDataToStruct(info.pstEventInfo, 0, stInfo);
|
System.out.println("objectType: " + new String(stInfo.stuObject.szObjectType, Charset.forName(Utils.getPlatformEncode())));
|
System.out.println("人脸匹配到的候选对象数量:" + stInfo.nCandidateNum + ",人员信息,相似度:" + stInfo.stuCandidates[0].bySimilarity);*/
|
//msg读字节
|
|
msg.getPointer().write(0,info.pstEventInfo.getByteArray(0,msg.size()),0,msg.size());
|
//读取nCandidateNum属性
|
int nCandidateNum=(int)msg.readField("nCandidateNum");
|
//获取stuCandidates结构体偏移量
|
int offset=msg.fieldOffset("stuCandidates");
|
//获取CANDIDATE_INFO结构体大小
|
int size=msg.stuCandidates[0].size();
|
for (int m = 0; m < nCandidateNum; m++) {
|
//按照偏移量写入stuCandidates数组
|
msg.stuCandidates[m].getPointer().write(0,info.pstEventInfo.getByteArray(offset+size*m, size),0,size);
|
//读取相似度
|
byte similarity=(byte)msg.stuCandidates[m].readField("bySimilarity");
|
//读取stPersonInfo结构体对象
|
msg.stuCandidates[m].readField("stPersonInfo");
|
msg.stuCandidates[m].stPersonInfo.read();
|
//人员唯一标示(证件号码,工号,或其他编号)
|
String personId=new String(msg.stuCandidates[m].stPersonInfo.szID,Charset.forName(Utils.getPlatformEncode())).trim();
|
//人员唯一标识符,首次由服务端生成,区别于ID字段
|
String personUId=new String(msg.stuCandidates[m].stPersonInfo.szUID,Charset.forName(Utils.getPlatformEncode())).trim();
|
//
|
System.out.println("id:"+personId+",uid: "+personUId+",similarity: "+similarity);
|
}
|
msg.readField("stuObject");
|
//保存图片
|
//下发的图片
|
String picture = picturePath + "/my-" + System.currentTimeMillis() + ".jpg";
|
ToolKits.savePicture(pBuf, 0, msg.stuObject.stPicInfo.dwFileLenth, picture);
|
//匹配的图片
|
picture = picturePath + "/search-" + System.currentTimeMillis() + ".jpg";
|
ToolKits.savePicture(pBuf, msg.stuObject.stPicInfo.dwFileLenth, dwBufSize, picture);
|
break;
|
}
|
default: {
|
System.out.println("default");
|
break;
|
}
|
}
|
}
|
|
System.out.println(new String(task.stuTaskResultInfos[i].szFileID, Charset.forName(Utils.getPlatformEncode())).trim());
|
}
|
long end=System.currentTimeMillis();
|
System.out.println("cost millions: "+(end-start));
|
return 0;
|
}
|
}
|
