first in the java side, my codes looks like:
private String path="/mnt/sdcard/";
InitFeature(width,height,path);
public native void InitFeature(int width, int height,String path);
then in the jni side, it's:
//passed from the java part and release after InitFreature
const char* nPath;
//to save the globle path
char* g_path;
JNIEXPORT void JNICALL Java_org_opencv_samples_tutorial6_Tutorial2Activity_InitFeature(JNIEnv* env, jobject,jint width,jint height,jstring path);
JNIEXPORT void JNICALL Java_org_opencv_samples_tutorial6_Tutorial2Activity_InitFeature(JNIEnv* env, jobject,jint width,jint height,jstring path)
{
//jsize path_size;
nPath=env->GetStringUTFChars(path,NULL);
//path_size=env->GetArrayLength(path);
LOGD("path: %s \n",nPath);
int length=strlen(nPath);
LOGD("length: %d \n",length);
g_path=new char[length+1];
memcpy(g_path,nPath,length+1);
LOGD("path_2: %s \n",g_path);
LOGD("length: %d \n",strlen(g_path));
char l_path[128];
strcpy(l_path,g_path);
strcat(l_path,"color.jpg");
LOGD("path_3: %s \n",l_path);
LOGD("length: %d \n",sizeof(l_path));
m_width=width;
m_height=height;
center.x=float (m_width/2.0+0.5);//float (Img_tmp->width/2.0+0.5);
center.y=float (m_width/2.0+0.5);//float (Img_tmp->height/2.0+0.5);
env->ReleaseStringUTFChars(path,nPath);
}
since I have different native calls, one to initiate features(shown here) and others to process every frame, and after you env->ReleaseStringUTFChars(path,nPath); the string would be invisible to jni part. I have the copy the string to a global char* g_path;
and a little sample is here as well, the file path is "/mnt/sdcard/color.jpg" and check those logs.
then you can use imread() to get this jpg.
I use other libs like libjpg, so I am not showing the codes here.