Add RTMP stream function with FFMPEG

Makefile

@@ -7,6 +7,8 @@ OPENMP=0
 LIBSO=0
 ZED_CAMERA=0
 ZED_CAMERA_v2_8=0
+STREAM=0
+FFMPEG=0
 
 # set GPU=1 and CUDNN=1 to speedup on GPU
 # set CUDNN_HALF=1 to further speedup 3 x times (Mixed-precision on Tensor Cores) GPU: Volta, Xavier, Turing, Ampere, Ada and higher
@@ -94,7 +96,21 @@ NVCC=nvcc
 OPTS=-Ofast
 LDFLAGS= -lm -pthread
 COMMON= -Iinclude/ -I3rdparty/stb/include
-CFLAGS=-Wall -Wfatal-errors -Wno-unused-result -Wno-unknown-pragmas -fPIC -rdynamic
+CFLAGS=-Wall -Wfatal-errors -Wno-unused-result -Wno-unknown-pragmas -fPIC -fpermissive -rdynamic
+
+ifeq ($(STREAM), 1)
+COMMON+= -DSTREAM
+CFLAGS+= -DSTREAM
+LDFLAGS+= `pkg-config --libs libavformat libavcodec libavutil libswscale 2>/dev/null`
+COMMON+= `pkg-config --cflags libavformat libavcodec libavutil libswscale 2>/dev/null`
+endif
+
+ifeq ($(FFMPEG), 1)
+COMMON+= -DFFMPEG
+CFLAGS+= -DFFMPEG
+LDFLAGS+= `pkg-config --libs libswresample libswscale libavutil libavcodec libavformat 2>/dev/null`
+COMMON+= `pkg-config --cflags libswresample libswscale libavutil libavcodec libavformat 2>/dev/null`
+endif
 
 ifeq ($(DEBUG), 1)
 #OPTS= -O0 -g
@@ -172,6 +188,12 @@ ifeq ($(GPU), 1)
 LDFLAGS+= -lstdc++
 OBJ+=convolutional_kernels.o activation_kernels.o im2col_kernels.o col2im_kernels.o blas_kernels.o crop_layer_kernels.o dropout_layer_kernels.o maxpool_layer_kernels.o network_kernels.o avgpool_layer_kernels.o
 endif
+ifeq ($(STREAM), 1)
+OBJ+=stream.o streamer.o
+endif
+ifeq ($(FFMPEG), 1)
+OBJ+=image_ffmpeg.o
+endif
 
 OBJS = $(addprefix $(OBJDIR), $(OBJ))
 DEPS = $(wildcard src/*.h) Makefile include/darknet.h

src/demo.c

@@ -15,6 +15,10 @@
 #include <sys/time.h>
 #endif
 
+#ifdef FFMPEG
+#include "image_ffmpeg.h"
+#endif
+
 #ifdef OPENCV
 
 #include "http_stream.h"
@@ -54,6 +58,9 @@ static const int thread_wait_ms = 1;
 static volatile int run_fetch_in_thread = 0;
 static volatile int run_detect_in_thread = 0;
 
+#ifdef FFMPEG
+static int input_is_stream = 0;
+#endif
 
 void *fetch_in_thread(void *ptr)
 {
@@ -65,10 +72,16 @@ void *fetch_in_thread(void *ptr)
             this_thread_yield();
         }
         int dont_close_stream = 0;    // set 1 if your IP-camera periodically turns off and turns on video-stream
-        if (letter_box)
+        if (letter_box){
             in_s = get_image_from_stream_letterbox(cap, net.w, net.h, net.c, &in_img, dont_close_stream);
-        else
+        }else{
+#ifdef FFMPEG
+            if (input_is_stream) in_s = get_image_from_ffmpeg_stream_resize(&in_img, net.w, net.h, net.c);
+            else in_s = get_image_from_stream_resize(cap, net.w, net.h, net.c, &in_img, dont_close_stream);
+#else
             in_s = get_image_from_stream_resize(cap, net.w, net.h, net.c, &in_img, dont_close_stream);
+#endif
+        }
         if (!in_s.data) {
             printf("Stream closed.\n");
             custom_atomic_store_int(&flag_exit, 1);
@@ -178,6 +191,10 @@ void demo(char *cfgfile, char *weightfile, float thresh, float hier_thresh, int
 
     if(filename){
         printf("video file: %s\n", filename);
+#ifdef FFMPEG
+        open_video_stream(filename);
+        input_is_stream = 1;
+#endif
         cap = get_capture_video_stream(filename);
         demo_skip_frame = is_live_stream(filename);
     }else{
@@ -358,7 +375,6 @@ void demo(char *cfgfile, char *weightfile, float thresh, float hier_thresh, int
                 write_frame_cv(output_video_writer, show_img);
                 printf("\n cvWriteFrame \n");
             }
-
             while (custom_atomic_load_int(&run_detect_in_thread)) {
                 if(avg_fps > 50) this_thread_yield();
                 else this_thread_sleep_for(thread_wait_ms);   // custom_join(detect_thread, 0);
@@ -406,6 +422,9 @@ void demo(char *cfgfile, char *weightfile, float thresh, float hier_thresh, int
                 start_time = get_time_point();
             }
         }
+#ifdef FFMPEG
+        av_pkt_unref();
+#endif
     }
     printf("input video stream closed. \n");
     if (output_video_writer) {

src/detector.c

@@ -9,6 +9,14 @@
 #include "demo.h"
 #include "option_list.h"
 
+#ifdef STREAM
+#include "stream.h"
+#endif
+
+#ifdef FFMPEG
+#include "image_ffmpeg.h"
+#endif
+
 #ifndef __COMPAR_FN_T
 #define __COMPAR_FN_T
 typedef int (*__compar_fn_t)(const void*, const void*);
@@ -158,7 +166,6 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
     args.mosaic_bound = net.mosaic_bound;
     args.contrastive = net.contrastive;
     args.contrastive_jit_flip = net.contrastive_jit_flip;
-    args.contrastive_color = net.contrastive_color;
     if (dont_show && show_imgs) show_imgs = 2;
     args.show_imgs = show_imgs;
 
@@ -1979,6 +1986,15 @@ void run_detector(int argc, char **argv)
     int ext_output = find_arg(argc, argv, "-ext_output");
     int save_labels = find_arg(argc, argv, "-save_labels");
     char* chart_path = find_char_arg(argc, argv, "-chart", 0);
+
+    int stream_bitrate = find_int_arg(argc, argv, "-stream_bitrate", 0);
+    int stream_frame_width = find_int_arg(argc, argv, "-stream_width", 0);
+    int stream_frame_height = find_int_arg(argc, argv, "-stream_height", 0);
+    int stream_gop_size = find_int_arg(argc, argv, "-stream_gop", 0);
+    int stream_fps = find_int_arg(argc, argv, "-stream_fps", 0);
+    char *stream_addr = find_char_arg(argc, argv, "-stream_address", 0);
+    char *stream_profile = find_char_arg(argc, argv, "-stream_profile", "high444");
+
     // While training, decide after how many epochs mAP will be calculated. Default value is 4 which means the mAP will be calculated after each 4 epochs
     int mAP_epochs = find_int_arg(argc, argv, "-mAP_epochs", 4);
     if (argc < 4) {
@@ -2028,17 +2044,26 @@ void run_detector(int argc, char **argv)
         int it_num = 100;
         draw_object(datacfg, cfg, weights, filename, thresh, dont_show, it_num, letter_box, benchmark_layers);
     }
-    else if (0 == strcmp(argv[2], "demo")) {
+    else if (0 == strcmp(argv[2], "demo") || 0 == strcmp(argv[2], "stream")) {
         list *options = read_data_cfg(datacfg);
         int classes = option_find_int(options, "classes", 20);
         char *name_list = option_find_str(options, "names", "data/names.list");
         char **names = get_labels(name_list);
         if (filename)
             if (strlen(filename) > 0)
                 if (filename[strlen(filename) - 1] == 0x0d) filename[strlen(filename) - 1] = 0;
-        demo(cfg, weights, thresh, hier_thresh, cam_index, filename, names, classes, avgframes, frame_skip, prefix, out_filename,
-            mjpeg_port, dontdraw_bbox, json_port, dont_show, ext_output, letter_box, time_limit_sec, http_post_host, benchmark, benchmark_layers, json_file_output);
 
+        if (0 == strcmp(argv[2], "demo")){
+            demo(cfg, weights, thresh, hier_thresh, cam_index, filename, names, classes, avgframes, frame_skip, prefix, out_filename,
+            mjpeg_port, dontdraw_bbox, json_port, dont_show, ext_output, letter_box, time_limit_sec, http_post_host, benchmark, benchmark_layers, json_file_output);
+        }
+#ifdef STREAM
+        if (0 == strcmp(argv[2], "stream")){
+            stream(cfg, weights, thresh, hier_thresh, cam_index, filename, names, classes, avgframes, frame_skip, prefix, out_filename,
+            mjpeg_port, dontdraw_bbox, json_port, dont_show, ext_output, letter_box, time_limit_sec, http_post_host, benchmark, benchmark_layers,
+            stream_bitrate, stream_addr, stream_frame_width, stream_frame_height, stream_profile, stream_gop_size, stream_fps);
+        }
+#endif
         free_list_contents_kvp(options);
         free_list(options);
     }