话不多说，想要搬移代码快速实现的直接看github：https://github.com/Frank1481906280/CV4Android
一、opencv的JNI配置可直接参考其他的blog，我贴出cmakelist，基本就行了。基本要求是要配置好NDK等，这些都可以在Android studio中完成。
# For more information about using CMake with Android Studio, read the
# documentation: https://d.android.com/studio/projects/add-native-code.html
# Sets the minimum version of CMake required to build the native library.
cmake_minimum_required(VERSION 3.4.1)
# ##################### OpenCV 环境 ############################
#设置OpenCV-android-sdk路径
set( OpenCV_DIR D:/opencv-3.4.6-android/OpenCV-android-sdk/sdk/native/jni )
find_package(OpenCV REQUIRED )
if(OpenCV_FOUND)
include_directories(${OpenCV_INCLUDE_DIRS})
message(STATUS "OpenCV library status:")
message(STATUS " version: ${OpenCV_VERSION}")
message(STATUS " libraries: ${OpenCV_LIBS}")
message(STATUS " include path: ${OpenCV_INCLUDE_DIRS}")
else(OpenCV_FOUND)
message(FATAL_ERROR "OpenCV library not found")
endif(OpenCV_FOUND)
# Creates and names a library, sets it as either STATIC
# or SHARED, and provides the relative paths to its source code.
# You can define multiple libraries, and CMake builds them for you.
# Gradle automatically packages shared libraries with your APK.
add_library( # Sets the name of the library.
native-lib
# Sets the library as a shared library.
SHARED
# Provides a relative path to your source file(s).
native-lib.cpp )
# Searches for a specified prebuilt library and stores the path as a
# variable. Because CMake includes system libraries in the search path by
# default, you only need to specify the name of the public NDK library
# you want to add. CMake verifies that the library exists before
# completing its build.
find_library( # Sets the name of the path variable.
log-lib
# Specifies the name of the NDK library that
# you want CMake to locate.
log )
# Specifies libraries CMake should link to your target library. You
# can link multiple libraries, such as libraries you define in this
# build script, prebuilt third-party libraries, or system libraries.
target_link_libraries( # Specifies the target library.
native-lib
${OpenCV_LIBS}
log
jnigraphics
# Links the target library to the log library
# included in the NDK.
${log-lib} )
二、关于常规图片的处理建议使用#include <android/bitmap.h>，这样可以直接将java层获取的bitmap直接转成OpenCV Mat类型。示例如下：
extern "C"
JNIEXPORT void JNICALL
Java_com_example_cv4android_MainActivity_pic2binary(JNIEnv *env, jobject thiz, jobject bitmap) {
// TODO: implement pic2binary()
AndroidBitmapInfo info; void *pixels;
CV_Assert(AndroidBitmap_getInfo(env, bitmap, &info) >= 0);
CV_Assert(info.format == ANDROID_BITMAP_FORMAT_RGBA_8888 || info.format == ANDROID_BITMAP_FORMAT_RGB_565);
CV_Assert(AndroidBitmap_lockPixels(env, bitmap, &pixels) >= 0);
CV_Assert(pixels); if (info.format == ANDROID_BITMAP_FORMAT_RGBA_8888)
{
Mat temp(info.height, info.width, CV_8UC4, pixels);
Mat gray;
cvtColor(temp, gray, COLOR_RGBA2GRAY);
adaptiveThreshold(gray,gray,255,1,0,5,10);
cvtColor(gray, temp, COLOR_GRAY2RGBA);
} else
{
Mat temp(info.height, info.width, CV_8UC2, pixels);
Mat gray;
cvtColor(temp, gray, COLOR_RGB2GRAY);
adaptiveThreshold(gray,gray,255,1,1,5,10);
cvtColor(gray, temp, COLOR_GRAY2RGB);
}
AndroidBitmap_unlockPixels(env, bitmap);
}
基本上只要转成了OpenCV所需要的Mat类型，你就可以使用任意的api对图像进行操作，但是我也踩过坑=。=（貌似使用bilateralFilter对图片进行双边滤波时候要求的type为CV_8UC3，但是我不管是从RGBA->RGB，或者直接CV_8UC4->CV_8UC3都不行）。
3、关于coco-mobilenet的加载，我用的是tensorflow下的pb模型。这一部分难点就是java中读取模型存放的路径转到jni中。
3.1首先我们先看JNI中是怎么用的：
extern "C"
JNIEXPORT void JNICALL
Java_com_example_cv4android_MainActivity_load2JNI_1SSD(JNIEnv *env, jobject thiz, jstring pbpath,
jstring configpath) {
// TODO: implement load2JNI_SSD()
const char *filePath_1 = env->GetStringUTFChars(pbpath, 0);
const char *filePath_2 = env->GetStringUTFChars(configpath, 0);
net=readNetFromTensorflow(filePath_1,filePath_2);
LOGE("加载分类器文件成功");
env->ReleaseStringUTFChars(pbpath, filePath_1);
env->ReleaseStringUTFChars(pbpath, filePath_2);
}
3.2java中怎么定义pbpath、configpath的
private void LoadModel2() {
String pbpath = getPath("frozen_inference_graph.pb", this);
String configpath = getPath("graph.pbtxt", this);
load2JNI_SSD(pbpath,configpath);
//net= Dnn.readNetFromTensorflow(pbpath,configpath);
flag2=true;
}
private native void load2JNI_SSD(String pbpath, String configpath);
private static String getPath(String file, Context context) {
AssetManager assetManager = context.getAssets();
BufferedInputStream inputStream = null;
try {
// Read data from assets.
inputStream = new BufferedInputStream(assetManager.open(file));
byte[] data = new byte[inputStream.available()];
inputStream.read(data);
inputStream.close();
// Create copy file in storage.
File outFile = new File(context.getFilesDir(), file);
FileOutputStream os = new FileOutputStream(outFile);
os.write(data);
os.close();
// Return a path to file which may be read in common way.
return outFile.getAbsolutePath();
} catch (IOException ex) {
Log.i("COCO-NET", "Failed to upload a file");
}
return "";
}
一目了然，我们将模型文件和config文件放置在assets目录下，通过读取assets目录下的文件获取路径传到我们的JNI中。
4、直接看我们是怎么使用这个模型对bitmap进行物体检测的！
extern "C"
JNIEXPORT jintArray JNICALL
Java_com_example_cv4android_MainActivity_SSD2detetct(JNIEnv *env, jobject thiz, jobject bitmap) {
// TODO: implement SSD2detetct()
vector<int> location_vec;
AndroidBitmapInfo info; void *pixels;
CV_Assert(AndroidBitmap_getInfo(env, bitmap, &info) >= 0);
CV_Assert(info.format == ANDROID_BITMAP_FORMAT_RGBA_8888 || info.format == ANDROID_BITMAP_FORMAT_RGB_565);
CV_Assert(AndroidBitmap_lockPixels(env, bitmap, &pixels) >= 0);
CV_Assert(pixels); if (info.format == ANDROID_BITMAP_FORMAT_RGBA_8888)
{
LOGE("分析识别");
Mat temp(info.height, info.width, CV_8UC4, pixels);
cvtColor(temp,temp,COLOR_RGBA2RGB);
int IN_WIDTH = 300;
int IN_HEIGHT = 300;
float WH_RATIO = (float)IN_WIDTH / IN_HEIGHT;
double IN_SCALE_FACTOR = 0.007843;
double MEAN_VAL = 127.5;
double THRESHOLD = 0.2;
//resize(temp,temp,Size(IN_HEIGHT,IN_WIDTH));
Mat blob=blobFromImage(temp,IN_SCALE_FACTOR,Size(IN_WIDTH,IN_HEIGHT),Scalar(MEAN_VAL,MEAN_VAL,MEAN_VAL),
false, false);
net.setInput(blob);
Mat detections =net.forward();
Mat detectionMat(detections.size[2], detections.size[3], CV_32F, detections.ptr<float>());
for(int i=0;i<detectionMat.rows;i++){
float confidence = detectionMat.at<float>(i, 2);
if (confidence > THRESHOLD)
{
size_t objectClass = (size_t)(detectionMat.at<float>(i, 1));
int tl_x = static_cast<int>(detectionMat.at<float>(i, 3) * temp.cols);
int tl_y = static_cast<int>(detectionMat.at<float>(i, 4) * temp.rows);
int br_x = static_cast<int>(detectionMat.at<float>(i, 5) * temp.cols);
int br_y = static_cast<int>(detectionMat.at<float>(i, 6) * temp.rows);
String label = format("%s: %.2f", classNames[objectClass], confidence);
location_vec.push_back(tl_x);
location_vec.push_back(tl_y);
location_vec.push_back(br_x);
location_vec.push_back(br_y);
classname.push_back(label);
LOGE("location: %d,%d,%d,%d\n",tl_x,tl_y,br_x,br_y);
LOGE("label: %s",label.c_str());
//rectangle(temp, Point(tl_x, tl_y), Point(br_x, br_y), Scalar(255,155,155),3);
//putText(temp, label, Point(tl_x, tl_y), FONT_HERSHEY_SIMPLEX, 0.5, Scalar(0, 255, 0));
}
}
//cvtColor(temp,temp,COLOR_RGB2RGBA);
} else
{
//基本为彩色图像，这里我就没写了，需要的可以自己更改
Mat temp(info.height, info.width, CV_8UC2, pixels);
}
AndroidBitmap_unlockPixels(env, bitmap);
int vecSize=location_vec.size();
if (vecSize == 0) return 0;
jintArray jarr = env->NewIntArray(vecSize);
//2.获取数组指针
jint *PCarr = env->GetIntArrayElements(jarr, NULL);
//3.赋值
int i = 0;
for(; i < vecSize; i++){
PCarr[i] = location_vec.at(i);
}
location_vec.clear();
//4.释放资源
env->ReleaseIntArrayElements(jarr, PCarr, 0);
//5.返回数组
return jarr;
}
在这个函数中，我直接return获取了物体的类别和坐标，然后发回给java层，通过cavan来绘制文字和矩形框，故这一部分就不写了！
总结：分享一下，免得以后自己忘记=- =。
网友评论