new

single_people_detect_track.py

@@ -459,7 +459,7 @@ def triangulate_one_point_from_multiple_views_linear(
     proj_matrices: Float[Array, "N 3 4"],
     points: Num[Array, "N 2"],
     confidences: Optional[Float[Array, "N"]] = None,
-    conf_threshold: float = 0.2,
+    conf_threshold: float = 0.4,  # 0.2
 ) -> Float[Array, "3"]:
     """
     Args:
@@ -473,7 +473,6 @@ def triangulate_one_point_from_multiple_views_linear(
     assert len(proj_matrices) == len(points)
     N = len(proj_matrices)
     # 置信度加权DLT
-    # 置信度加权DLT
     if confidences is None:
         weights = jnp.ones(N, dtype=jnp.float32)
     else:
@@ -932,18 +931,18 @@ def filter_camera_port(detections: list[Detection]):
 
 
 # 相机内外参路径
-CAMERA_PATH = Path("/home/admin/Documents/ActualTest_WeiHua/camera_params")
+CAMERA_PATH = Path("/home/admin/Documents/ActualTest_WeiHua/camera_params/")
 # 所有机位的相机内外参
 AK_CAMERA_DATASET: ak.Array = get_camera_params(CAMERA_PATH)
 
 # 2d检测数据路径
-DATASET_PATH = Path("/home/admin/Documents/ActualTest_WeiHua/Test_Video")
+DATASET_PATH = Path("/home/admin/Documents/ActualTest_WeiHua/Segment_1/")
 # 指定机位的2d检测数据
 camera_port = [5602, 5603, 5604, 5605]
 KEYPOINT_DATASET = get_camera_detect(DATASET_PATH, camera_port, AK_CAMERA_DATASET)
 
 # 获取一段完整的跳跃片段
-FRAME_INDEX = [i for i in range(700, 1600)]  # 552， 1488
+FRAME_INDEX = [i for i in range(700, 1600)]  # Segement_1:(700, 1600)
 KEYPOINT_DATASET = get_segment(camera_port, FRAME_INDEX, KEYPOINT_DATASET)
 
 
@@ -974,8 +973,10 @@ all_3d_kps: dict[str, list] = {}
 
 tracking_initialized = False
 lost_frame_count = 0
-lost_frame_threshold = 12  # 0.5秒，假设20fps
+lost_frame_threshold = 12  # 0.5秒
 
+# 丢失目标帧数计数器
+loss_track_count = 0
 
 # ===================== 主循环：逐帧处理检测与跟踪 =====================
 while True:
@@ -984,7 +985,8 @@ while True:
         # 获取下一个时间步的所有相机检测结果
         detections = next(sync_gen)
         # 过滤低置信度的检测，提升后续三角化和跟踪的准确性
-        detections = filter_keypoints_by_scores(detections, threshold=0.5)
+        detections = filter_keypoints_by_scores(detections, threshold=0.2)
+        # detections = get_filter_detections(detections) # 伞降跳台时使用
     except StopIteration:
         # 检测数据读取完毕，退出主循环
         break
@@ -1026,7 +1028,7 @@ while True:
         lost_frame_count += 1  # 丢失帧数+1
         # 进一步完善退出条件：
         # 1. 连续丢失阈值帧后才退出
-        # 2. 若丢失时，最后一次检测到的时间与当前帧时间间隔超过1秒，才彻底退出
+        # 2. 若丢失时，最后一次检测到的时间与当前帧时间间隔超过0.5秒，才彻底退出
         last_tracking = None
         if global_tracking_state.trackings:
             last_tracking = list(global_tracking_state.trackings.values())[0]
@@ -1058,6 +1060,10 @@ while True:
         tracking_detection = []  # 存储每个跟踪目标在各相机下最优匹配的检测
         temp_matrix = []  # 打印用：每个相机的最大相似度
         for camera_name in affinities.keys():
+            # indices_T:表示匹配到检测的tracking的索引(在tracking列表中的下标)
+            # indices_D:表示匹配到tracking的detection的索引(在detections列表中的下标)
+            indices_T = affinities[camera_name].indices_T.item()
+            indices_D = affinities[camera_name].indices_D.item()
             camera_matrix = jnp.array(affinities[camera_name].matrix).flatten()
             detection_index = jnp.argmax(camera_matrix).item()  # 取最大相似度的检测索引
             if isnan(camera_matrix[detection_index].item()):
@@ -1065,11 +1071,11 @@ while True:
             temp_matrix.append(
                 f"{camera_name} : {camera_matrix[detection_index].item()}"
             )
+            match_tracking = affinities[camera_name].trackings[indices_T]
             # 选取相似度大于阈值的检测目标更新跟踪状态
             # if camera_matrix[detection_index].item() > affinities_threshold:
-            tracking_detection.append(
-                affinities[camera_name].detections[detection_index]
-            )
+            # if match_tracking == element_tracking:
+            tracking_detection.append(affinities[camera_name].detections[indices_D])
         print("affinities matrix:", temp_matrix)
         # 只有匹配到足够多的检测目标时才更新跟踪（如多于2个相机）
         if len(tracking_detection) > 2:
@@ -1082,13 +1088,17 @@ while True:
                 "update tracking:",
                 global_tracking_state.trackings.values(),
             )
-        # 不再在else分支里删除tracking，只用lost_frame_count判定
+        else:
+            loss_track_count += 1
+        # ======如果单帧数据量不够，考虑如何更新跟踪=====
 
 # 对每一个3d目标进行滑动窗口平滑处理
 smoothed_points = smooth_3d_keypoints(all_3d_kps, window_size=5)
 
+print("Tracking completed, total loss frames processed:", count)
+
 # 将结果保存到json文件中
-with open("samples/Test_WeiHua.json", "wb") as f:
+with open("samples/Test_WeiHua_Segment_1.json", "wb") as f:
     f.write(orjson.dumps(smoothed_points))
 # 输出每个3d目标的维度
 for element_3d_kps_id in smoothed_points.keys():