mmr/tkDNN
Archived
1
0
Fork 0
Code Issues 46 Pull Requests 12 Releases 4 Activity

Merge with cnet

Browse Source
This commit is contained in:
Fabio Bagni
2021-05-04 11:50:08 +02:00
parent 10f39d1055 d1ae1791d9
commit f8327e2dac
4 changed files with 29 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CMakeLists.txt
+3
View File
@@ -46,6 +46,9 @@ include_directories(${EIGEN3_INCLUDE_DIR})
find_package(OpenCV REQUIRED)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DOPENCV")
if(OpenCV_CUDA_VERSION)
add_compile_definitions(OPENCV_CUDACONTRIB)
endif()
# gives problems in cross-compiling, probably malformed cmake config
find_package(yaml-cpp REQUIRED)
include/tkDNN/CenternetDetection3DTrack.h
+2 -2
View File
@@ -151,7 +151,7 @@ public:
std::vector<std::vector<struct trackingRes>> tr_res;
std::vector<std::vector<struct trackingRes>> batchTracked;
std::vector<int> count_tr;
int track_id=0;
std::vector<int> track_id;
bool init_preprocessing();
@@ -161,7 +161,7 @@ public:
void pre_inf(const int bi);
void _get_additional_inputs();
cv::Mat transform_preds_with_trans(float x1, float x2);
void tracking(int bi);
void tracking(const int bi);
public:
tk::dnn::Network *pre_phase_net = nullptr;
src/CenternetDetection3DTrack.cpp
+23 -14
View File
@@ -14,12 +14,12 @@ bool CenternetDetection3DTrack::init(const std::string& tensor_path, const int n
confThreshold = conf_thresh;
inputCalibs = k_calibs;
tr_res.resize(nBatches);
count_tr.resize(nBatches, 0);
init_preprocessing();
init_pre_inf();
init_postprocessing();
init_visualization(n_classes);
count_tr.resize(nBatches, 0);
}
bool CenternetDetection3DTrack::init_preprocessing(){
@@ -203,6 +203,11 @@ bool CenternetDetection3DTrack::init_postprocessing(){
// Alloc array used in the kernel
checkCuda( cudaMalloc(&src_out, K *sizeof(float)) );
checkCuda( cudaMalloc(&ids_out, K *sizeof(int)) );
for(int bi=0; bi<nBatches; bi++){
track_id.push_back(0);
count_tr.push_back(0);
}
}
bool CenternetDetection3DTrack::init_visualization(const int n_classes){
@@ -293,7 +298,6 @@ void CenternetDetection3DTrack::pre_inf(const int bi){
void CenternetDetection3DTrack::preprocess(cv::Mat &frame, const int bi, const std::vector<cv::Size>& stream_size){
// -----------------------------------pre-process ------------------------------------------
batchTracked.clear();
cv::Size sz = originalSize[bi];
float scale = 1.0;
float new_height = sz.height * scale;
@@ -375,7 +379,7 @@ void CenternetDetection3DTrack::preprocess(cv::Mat &frame, const int bi, const s
cv::warpAffine(imageF, imageF, trans, cv::Size(dim.w, dim.h), cv::INTER_LINEAR );
cv::imshow("warp", imageF);
//cv::imshow("warp", imageF);
sz = imageF.size();
imageF.convertTo(imageF, CV_32FC3, 1/255.0);
@@ -418,8 +422,7 @@ cv::Mat CenternetDetection3DTrack::transform_preds_with_trans(float x1, float x2
return trans_out * target_coords;
}
void CenternetDetection3DTrack::tracking(int bi){
void CenternetDetection3DTrack::tracking(const int bi) {
float item_size[count_det];
int item_cl[count_det];
float dets[2*count_det];
@@ -495,7 +498,7 @@ void CenternetDetection3DTrack::tracking(int bi){
int tr_id = matched_indices[2*i];
int d_id = matched_indices[2*i+1];
// tr_res[bi][tr_id].det_res = det_res[d_id];
// tr_res[tr_id].det_res = det_res[d_id];
tr_res[bi][tr_id].det_res.score = det_res[d_id].score;
tr_res[bi][tr_id].det_res.cl = det_res[d_id].cl;
tr_res[bi][tr_id].det_res.ct = det_res[d_id].ct;
@@ -527,7 +530,6 @@ void CenternetDetection3DTrack::tracking(int bi){
tr_res[bi].clear();
count_tr[bi] = 0;
}
int old_count_tr = count_tr[bi];
if(count_tr[bi] != 0 && new_count_tr != count_tr[bi]) {
std::vector<struct trackingRes> new_tr_res;
@@ -535,7 +537,7 @@ void CenternetDetection3DTrack::tracking(int bi){
for(int i=0; i<count_tr[bi]; i++) {
if(unmatched_tracks[i]) {
struct trackingRes new_tr_res_;
// new_tr_res_new_det_res.det_res = tr_res[bi][i].det_res;
// new_tr_res_new_det_res.det_res = tr_res[i].det_res;
new_tr_res_.det_res.score = tr_res[bi][i].det_res.score;
new_tr_res_.det_res.cl = tr_res[bi][i].det_res.cl;
new_tr_res_.det_res.ct = tr_res[bi][i].det_res.ct;
@@ -587,17 +589,24 @@ void CenternetDetection3DTrack::tracking(int bi){
new_tr_res_.det_res.y = det_res[i].y;
new_tr_res_.det_res.z = det_res[i].z;
new_tr_res_.det_res.rot_y = det_res[i].rot_y;
new_tr_res_.tracking_id = track_id++;
new_tr_res_.tracking_id = track_id[bi]++;
new_tr_res_.age = 1;
new_tr_res_.active = 1;
new_tr_res_.color = rand() % 256;
tr_res[bi].push_back(new_tr_res_);
if(tr_res.size() <= bi) {
std::vector<struct trackingRes> v_new_tr_res_;
v_new_tr_res_.push_back(new_tr_res_);
tr_res.push_back(v_new_tr_res_);
}
else
tr_res[bi].push_back(new_tr_res_);
}
}
count_tr[bi] = count_tr_;
if(track_id==1000)
track_id=0;
if(track_id[bi]==1000)
track_id[bi]=0;
det_res.clear();
}
@@ -747,8 +756,8 @@ void CenternetDetection3DTrack::draw(std::vector<cv::Mat>& frames) {
int thickness = 2;
for(int bi=0; bi<frames.size(); ++bi) {
// draw dets
for(int i=0; i<batchTracked[bi].size(); i++) {
t = batchTracked[bi][i];
for(int i=0; tr_res.size() != 0 && i<tr_res[bi].size(); i++) {
t = tr_res[bi][i];
id = t.tracking_id;
txt = classesNames[t.det_res.cl-1]+'-'+std::to_string(id); //forse ha bisogno di cl-1
cv::Size text_size = getTextSize(txt, cv::FONT_HERSHEY_SIMPLEX, font_scale, thickness, &baseline);
src/NetworkRT.cpp
+1 -1
View File
@@ -140,7 +140,7 @@ NetworkRT::NetworkRT(Network *net, const char *name) {
engineRT = builderRT->buildEngineWithConfig(*networkRT, *configRT);
#else
//engineRT = builderRT->buildCudaEngine(*networkRT);
engineRT = std::shared_ptr<nvinfer1::ICudaEngine>(builderRT->buildCudaEngine(*networkRT));
engineRT = builderRT->buildCudaEngine(*networkRT);
#endif
if(engineRT == nullptr)
FatalError("cloud not build cuda engine")