pipeline.py add tracking module time info (#7627)

* pipeline.py add tracking module time
fix infer_cfg_vehicle_violation.yml for examples

* pipeline examples infer_cfg_vehicle_violation.yml delete irrelevant config

* fix infer_cfg_vehicle_violation.yml enable

deploy/pipeline/config/examples/infer_cfg_vehicle_violation.yml

@@ -24,7 +24,8 @@ VEHICLE_RETROGRADE:
   frame_len: 8
   sample_freq: 7
   enable: True
-  filter_horizontal_flag: False
-  deviation: 23
+  filter_horizontal_flag: True
+  keep_right_flag: True
+  deviation: 45
   move_scale: 0.01
-  fence_line: []   #[x1,y1,x2,y2] y2>y1.
+  fence_line: [570,  163, 1030,  752]   #[x1,y1,x2,y2] y2>y1.

deploy/pipeline/config/tracker_config.yml

-# config of tracker for MOT SDE Detector, use 'JDETracker' as default.
+# config of tracker for MOT SDE Detector, use 'OCSORTTracker' as default, 'JDETracker' here is just BYTETracker.
 # The tracker of MOT JDE Detector (such as FairMOT) is exported together with the model.
 # Here 'min_box_area' and 'vertical_ratio' are set for pedestrian, you can modify for other objects tracking.
 
-type: OCSORTTracker # choose one tracker in ['JDETracker', 'OCSORTTracker']
+type: BOTSORTTracker # choose one tracker in ['JDETracker', 'OCSORTTracker', 'DeepSORTTracker','BOTSORTTracker']
+# When using for MTMCT(Multi-Target Multi-Camera Tracking), you should modify to 'DeepSORTTracker'
 
 
-# BYTETracker
+# just as BYTETracker, used for FairMOT in PP-Tracking project and for ByteTrack in PP-Humanv1 project
 JDETracker:
   use_byte: True
   det_thresh: 0.3
@@ -16,6 +17,7 @@ JDETracker:
   vertical_ratio: 0 # 1.6 for pedestrian
 
 
+# used for OC-SORT in PP-Humanv2 project and PP-Vehicle project
 OCSORTTracker:
   det_thresh: 0.4
   max_age: 30
@@ -23,6 +25,31 @@ OCSORTTracker:
   iou_threshold: 0.3
   delta_t: 3
   inertia: 0.2
-  vertical_ratio: 0
   min_box_area: 0
+  vertical_ratio: 0
   use_byte: False
+  use_angle_cost: False
+
+
+# used for DeepSORT and MTMCT in PP-Tracking project
+DeepSORTTracker:
+  input_size: [64, 192] # An unique operation to scale the sub-image of the selected detected boxes to a fixed size
+  min_box_area: 0
+  vertical_ratio: -1
+  budget: 100
+  max_age: 70
+  n_init: 3
+  metric_type: cosine
+  matching_threshold: 0.2
+  max_iou_distance: 0.9
+
+BOTSORTTracker:
+  track_high_thresh: 0.3
+  track_low_thresh: 0.2
+  new_track_thresh: 0.4
+  match_thresh: 0.7
+  track_buffer: 30
+  min_box_area: 0
+  camera_motion: False
+  cmc_method: 'sparseOptFlow' # only camera_motion is True,
+                              # sparseOptFlow | files (Vidstab GMC) | orb | ecc

deploy/pipeline/pipeline.py

@@ -530,6 +530,7 @@ class PipePredictor(object):
         else:
             self.predict_image(input)
         self.pipe_timer.info()
+        self.mot_predictor.det_times.tracking_info(average=True)
 
     def predict_image(self, input):
         # det
@@ -747,7 +748,8 @@ class PipePredictor(object):
                 res = self.mot_predictor.predict_image(
                     [copy.deepcopy(frame_rgb)],
                     visual=False,
-                    reuse_det_result=reuse_det_result)
+                    reuse_det_result=reuse_det_result,
+                    frame_count=frame_id)
 
                 # mot output format: id, class, score, xmin, ymin, xmax, ymax
                 mot_res = parse_mot_res(res)

deploy/pptracking/python/mot_utils.py

@@ -271,6 +271,40 @@ class Timer(Times):
                 format(preprocess_time * 1000, inference_time * 1000,
                        postprocess_time * 1000))
 
+    def tracking_info(self, average=True):
+        pre_time = self.preprocess_time_s.value()
+        infer_time = self.inference_time_s.value()
+        post_time = self.postprocess_time_s.value()
+        track_time = self.tracking_time_s.value()
+
+        total_time = pre_time + infer_time + post_time
+        if self.with_tracker:
+            total_time = total_time + track_time
+        total_time = round(total_time, 4)
+        print(
+            "------------------ Tracking Module Time Info ----------------------"
+        )
+
+        preprocess_time = round(pre_time / max(1, self.img_num),
+                                4) if average else pre_time
+        postprocess_time = round(post_time / max(1, self.img_num),
+                                 4) if average else post_time
+        inference_time = round(infer_time / max(1, self.img_num),
+                               4) if average else infer_time
+        tracking_time = round(track_time / max(1, self.img_num),
+                              4) if average else track_time
+
+        if self.with_tracker:
+            print(
+                "preprocess_time(ms): {:.2f}, inference_time(ms): {:.2f}, postprocess_time(ms): {:.2f}, tracking_time(ms): {:.2f}".
+                format(preprocess_time * 1000, inference_time * 1000,
+                       postprocess_time * 1000, tracking_time * 1000))
+        else:
+            print(
+                "preprocess_time(ms): {:.2f}, inference_time(ms): {:.2f}, postprocess_time(ms): {:.2f}".
+                format(preprocess_time * 1000, inference_time * 1000,
+                       postprocess_time * 1000))
+
     def report(self, average=False):
         dic = {}
         pre_time = self.preprocess_time_s.value()