// Tencent is pleased to support the open source community by making ncnn available.
//
// Copyright (C) 2021 THL A29 Limited, a Tencent company. All rights reserved.
//
// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
// in compliance with the License. You may obtain a copy of the License at
//
// https://opensource.org/licenses/BSD-3-Clause
//
// Unless required by applicable law or agreed to in writing, software distributed
// under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
// CONDITIONS OF ANY KIND, either express or implied. See the License for the
// specific language governing permissions and limitations under the License.
#include "yolo.h"
#include <opencv2/core/core.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include "cpu.h"
static float fast_exp(float x) {
union {
uint32_t i;
float f;
} v{};
v.i = (1 << 23) * (1.4426950409 * x + 126.93490512f);
return v.f;
}
static float sigmoid(float x) {
return 1.0f / (1.0f + fast_exp(-x));
}
static float intersection_area(const Object &a, const Object &b) {
cv::Rect_<float> inter = a.rect & b.rect;
return inter.area();
}
static void qsort_descent_inplace(std::vector<Object> &face_objects, int left, int right) {
int i = left;
int j = right;
float p = face_objects[(left + right) / 2].prob;
while (i <= j) {
while (face_objects[i].prob > p)
i++;
while (face_objects[j].prob < p)
j--;
if (i <= j) {
// swap
std::swap(face_objects[i], face_objects[j]);
i++;
j--;
}
}
// #pragma omp parallel sections
{
// #pragma omp section
{
if (left < j) qsort_descent_inplace(face_objects, left, j);
}
// #pragma omp section
{
if (i < right) qsort_descent_inplace(face_objects, i, right);
}
}
}
static void qsort_descent_inplace(std::vector<Object> &face_objects) {
if (face_objects.empty())
return;
qsort_descent_inplace(face_objects, 0, face_objects.size() - 1);
}
static void nms_sorted_bboxes(const std::vector<Object> &face_objects, std::vector<int> &picked,
float nms_threshold) {
picked.clear();
const int n = face_objects.size();
std::vector<float> areas(n);
for (int i = 0; i < n; i++) {
areas[i] = face_objects[i].rect.width * face_objects[i].rect.height;
}
for (int i = 0; i < n; i++) {
const Object &a = face_objects[i];
int keep = 1;
for (int j = 0; j < (int) picked.size(); j++) {
const Object &b = face_objects[picked[j]];
// intersection over union
float inter_area = intersection_area(a, b);
float union_area = areas[i] + areas[picked[j]] - inter_area;
// float IoU = inter_area / union_area
if (inter_area / union_area > nms_threshold)
keep = 0;
}
if (keep)
picked.push_back(i);
}
}
static void
generate_grids_and_stride(const int target_w, const int target_h, std::vector<int> &strides,
std::vector<GridAndStride> &grid_strides) {
for (int stride: strides) {
int num_grid_w = target_w / stride;
int num_grid_h = target_h / stride;
for (int g1 = 0; g1 < num_grid_h; g1++) {
for (int g0 = 0; g0 < num_grid_w; g0++) {
GridAndStride gs;
gs.grid0 = g0;
gs.grid1 = g1;
gs.stride = stride;
grid_strides.push_back(gs);
}
}
}
}
static void generate_proposals(std::vector<GridAndStride> grid_strides, const ncnn::Mat &pred,
float prob_threshold, std::vector<Object> &objects) {
const int num_points = grid_strides.size();
//识别种类数
const int num_class = 43;
const int reg_max_1 = 16;
for (int i = 0; i < num_points; i++) {
const float *scores = pred.row(i) + 4 * reg_max_1;
// find label with max score
int label = -1;
float score = -FLT_MAX;
for (int k = 0; k < num_class; k++) {
float confidence = scores[k];
if (confidence > score) {
label = k;
score = confidence;
}
}
float box_prob = sigmoid(score);
if (box_prob >= prob_threshold) {
ncnn::Mat bbox_pred(reg_max_1, 4, (void *) pred.row(i));
{
ncnn::Layer *softmax = ncnn::create_layer("Softmax");
ncnn::ParamDict pd;
pd.set(0, 1); // axis
pd.set(1, 1);
softmax->load_param(pd);
ncnn::Option opt;
opt.num_threads = 1;
opt.use_packing_layout = false;
softmax->create_pipeline(opt);
softmax->forward_inplace(bbox_pred, opt);
softmax->destroy_pipeline(opt);
delete softmax;
}
float pred_ltrb[4];
for (int k = 0; k < 4; k++) {
float dis = 0.f;
const float *dis_after_sm = bbox_pred.row(k);
for (int l = 0; l < reg_max_1; l++) {
dis += l * dis_after_sm[l];
}
pred_ltrb[k] = dis * grid_strides[i].stride;
}
float pb_cx = (grid_strides[i].grid0 + 0.5f) * grid_strides[i].stride;
float pb_cy = (grid_strides[i].grid1 + 0.5f) * grid_strides[i].stride;
float x0 = pb_cx - pred_ltrb[0];
float y0 = pb_cy - pred_ltrb[1];
float x1 = pb_cx + pred_ltrb[2];
float y1 = pb_cy + pred_ltrb[3];
Object obj;
obj.rect.x = x0;
obj.rect.y = y0;
obj.rect.width = x1 - x0;
obj.rect.height = y1 - y0;
obj.label = label;
obj.prob = box_prob;
objects.push_back(obj);
}
}
}
Yolo::Yolo() {
blob_pool_allocator.set_size_compare_ratio(0.f);
workspace_pool_allocator.set_size_compare_ratio(0.f);
}
int
Yolo::load(const char *model_type, int target_size, const float *mean_values,
const float *norm_values, bool use_gpu) {
}
int
Yolo::load(AAssetManager *mgr, const char *model_type, int _target_size, const float *_mean_values,
const float *_norm_values, bool use_gpu) {
yolo.clear();
blob_pool_allocator.clear();
workspace_pool_allocator.clear();
ncnn::set_cpu_powersave(2);
ncnn::set_omp_num_threads(ncnn::get_big_cpu_count());
yolo.opt = ncnn::Option();
#if NCNN_VULKAN
yolo.opt.use_vulkan_compute = use_gpu;
#endif
yolo.opt.num_threads = ncnn::get_big_cpu_count();
yolo.opt.blob_allocator = &blob_pool_allocator;
yolo.opt.workspace_allocator = &workspace_pool_allocator;
char param_path[256];
char model_path[256];
//拼接模型名（路径）
sprintf(param_path, "yolov8%s.param", model_type);
sprintf(model_path, "yolov8%s.bin", model_type);
__android_log_print(ANDROID_LOG_DEBUG, "ncnn", "param_path %s", param_path);
__android_log_print(ANDROID_LOG_DEBUG, "ncnn", "model_path %s", model_path);
yolo.load_param(mgr, param_path);
yolo.load_model(mgr, model_path);
target_size = _target_size;
mean_values[0] = _mean_values[0];
mean_values[1] = _mean_values[1];
mean_values[2] = _mean_values[2];
norm_values[0] = _norm_values[0];
norm_values[1] = _norm_values[1];
norm_values[2] = _norm_values[2];
return 0;
}
void Yolo::setNativeCallback(JavaVM *vm, jobject input, jlong nativeObjAddr, jobject pJobject) {
javaVM = vm;
/**
* JNIEnv不支持跨线程调用
* */
JNIEnv *env;
vm->AttachCurrentThread(&env, nullptr);
//此时input转为output
j_output = env->NewGlobalRef(input);
j_mat_addr = nativeObjAddr;
j_callback = env->NewGlobalRef(pJobject);
}
int
Yolo::detect(const cv::Mat &rgb, std::vector<Object> &objects, float prob_threshold,
float nms_threshold) {
int width = rgb.cols;
int height = rgb.rows;
// pad to multiple of 32
int w = width;
int h = height;
float scale = 1.f;
if (w > h) {
scale = (float) target_size / w;
w = target_size;
h = h * scale;
} else {
scale = (float) target_size / h;
h = target_size;
w = w * scale;
}
ncnn::Mat in = ncnn::Mat::from_pixels_resize(rgb.data, ncnn::Mat::PIXEL_RGB2BGR, width, height,
w, h);
// pad to target_size rectangle
int w_pad = (w + 31) / 32 * 32 - w;
int h_pad = (h + 31) / 32 * 32 - h;
ncnn::Mat in_pad;
ncnn::copy_make_border(in, in_pad, h_pad / 2, h_pad - h_pad / 2, w_pad / 2, w_pad - w_pad / 2,
ncnn::BORDER_CONSTANT, 0.f);
in_pad.substract_mean_normalize(0, norm_values);
ncnn::Extractor ex = yolo.create_extractor();
ex.input("images", in_pad);
std::vector<Object> proposals;
ncnn::Mat out;
ex.extract("output", out);
std::vector<int> strides = {8, 16, 32}; // might have stride=64
std::vector<GridAndStride> grid_strides;
generate_grids_and_stride(in_pad.w, in_pad.h, strides, grid_strides);
generate_proposals(grid_strides, out, prob_threshold, proposals);
// sort all proposals by score from highest to lowest
qsort_descent_inplace(proposals);
// apply nms with nms_threshold
std::vector<int> picked;
nms_sorted_bboxes(proposals, picked, nms_threshold);
int count = picked.size();
objects.resize(count);
for (int i = 0; i < count; i++) {
objects[i] = proposals[picked[i]];
// adjust offset to original unpadded
float x0 = (objects[i].rect.x - (w_pad / 2)) / scale;
float y0 = (objects[i].rect.y - (h_pad / 2)) / scale;
float x1 = (objects[i].rect.x + objects[i].rect.width - (w_pad / 2)) / scale;
float y1 = (objects[i].rect.y + objects[i].rect.height - (h_pad / 2)) / scale;
// clip
x0 = std::max(std::min(x0, (float) (width - 1)), 0.f);
y0 = std::max(std::min(y0, (float) (height - 1)), 0.f);
x1 = std::max(std::min(x1, (float) (width - 1)), 0.f);
y1 = std::max(std::min(y1, (float) (height - 1)), 0.f);
objects[i].rect.x = x0;
objects[i].rect.y = y0;
objects[i].rect.width = x1 - x0;
objects[i].rect.height = y1 - y0;
}
// sort objects by area
struct {
bool operator()(const Object &a, const Object &b) const {
return a.rect.area() > b.rect.area();
}
} objects_area_greater;
std::sort(objects.begin(), objects.end(), objects_area_greater);
/**
* 回调给Java/Kotlin层
* */
JNIEnv *env;
javaVM->AttachCurrentThread(&env, nullptr);
jclass callback_clazz = env->GetObjectClass(j_callback);
jclass output_clazz = env->GetObjectClass(j_output);
jmethodID j_method_id = env->GetMethodID(
callback_clazz, "onDetect", "(Ljava/util/ArrayList;)V"
);
jclass list_clazz = env->FindClass("java/util/ArrayList");
jmethodID arraylist_init = env->GetMethodID(list_clazz, "<init>", "()V");
jmethodID arraylist_add = env->GetMethodID(list_clazz, "add", "(Ljava/lang/Object;)Z");
jobject arraylist_obj = env->NewObject(list_clazz, arraylist_init);
for (int i = 0; i < count; i++) {
auto item = objects[i];
jfieldID type = env->GetFieldID(output_clazz, "type", "I");
env->SetIntField(j_output, type, item.label);
jfieldID position = env->GetFieldID(output_clazz, "position", "[F");
float array[4];
array[0] = item.rect.x;
array[1] = item.rect.y;
array[2] = item.rect.width;
array[3] = item.rect.height;
jfloatArray rectArray = env->NewFloatArray(4);
env->SetFloatArrayRegion(rectArray, 0, 4, array);
env->SetObjectField(j_output, position, rectArray);
jfieldID prob = env->GetFieldID(output_clazz, "prob", "F");
env->SetFloatField(j_output, prob, item.prob);
//add
env->CallBooleanMethod(arraylist_obj, arraylist_add, j_output);
}
//回调
env->CallVoidMethod(j_callback, j_method_id, arraylist_obj);
/**
* Mat数据。
* <br>-----------------------------------------------<br>
* 通过内存地址赋值。Java层传入Mat对象内存地址，再通过C++给此地址赋值，Java即可得到内存地址的Mat矩阵数据
* */
__android_log_print(ANDROID_LOG_DEBUG, "ncnn", "nativeObjAddr %lld", j_mat_addr);
auto *res = (cv::Mat *) j_mat_addr;
res->create(rgb.rows, rgb.cols, rgb.type());
memcpy(res->data, rgb.data, rgb.rows * rgb.step);
return 0;
}
int Yolo::draw(cv::Mat &rgb, const std::vector<Object> &objects) {
static const char *class_names[] = {
"tripod", "tee", "person",
"shut-off valve", "hazard signs", "pressure tester",
"pressure gauge", "reflective clothing", "respirator masks",
"throat foil", "round-headed water gun", "safety signs",
"helmet", "security identification", "safety ropes",
"intercom", "pointed water gun", "switch",
"alarm device", "joint", "construction street signs",
"gas detectors", "hoses", "hose_rectangle",
"flow-meter", "fire hydrant box", "fire extinguisher",
"lighting equipment", "flame-out protection", "exposed wires",
"circuit diagram", "cordon", "regulator",
"length adjuster", "stickers", "across wires",
"road cones", "hose", "filter",
"distribution box", "long-shank valves", "valve", "ducts"
};
static const unsigned char colors[19][3] = {
{54, 67, 244},
{99, 30, 233},
{176, 39, 156},
{183, 58, 103},
{181, 81, 63},
{243, 150, 33},
{244, 169, 3},
{212, 188, 0},
{136, 150, 0},
{80, 175, 76},
{74, 195, 139},
{57, 220, 205},
{59, 235, 255},
{7, 193, 255},
{0, 152, 255},
{34, 87, 255},
{72, 85, 121},
{158, 158, 158},
{139, 125, 96}
};
int color_index = 0;
for (const auto &obj: objects) {
const unsigned char *color = colors[color_index % 19];
color_index++;
cv::Scalar cc(color[0], color[1], color[2]);
cv::rectangle(rgb, obj.rect, cc, 2);
char text[256];
sprintf(text, "%s", class_names[obj.label]);
// sprintf(text, "%s %.1f%%", class_names[obj.label], obj.prob * 100);
int baseLine = 0;
cv::Size label_size = cv::getTextSize(text, cv::FONT_HERSHEY_SIMPLEX, 0.5, 1, &baseLine);
int x = obj.rect.x;
int y = obj.rect.y - label_size.height - baseLine;
if (y < 0)
y = 0;
if (x + label_size.width > rgb.cols)
x = rgb.cols - label_size.width;
cv::Size size = cv::Size(label_size.width, label_size.height + baseLine);
cv::Rect rc = cv::Rect(cv::Point(x, y), size);
cv::rectangle(rgb, rc, cc, -1);
cv::Scalar text_scalar = (color[0] + color[1] + color[2] >= 381)
? cv::Scalar(0, 0, 0)
: cv::Scalar(255, 255, 255);
cv::putText(rgb, text,
cv::Point(x, y + label_size.height),
cv::FONT_HERSHEY_SIMPLEX,
0.5,
text_scalar, 1
);
}
return 0;
}
