ppvehicle plate license

deploy/pphuman/config/infer_cfg_pphuman.yml

@@ -9,7 +9,7 @@ DET:
   batch_size: 1
 
 MOT:
-  model_dir: pipeline_model//mot_ppyoloe_s_36e_ppvehicle/
+  model_dir: output_inference/mot_ppyoloe_l_36e_pipeline/
   tracker_config: deploy/pphuman/config/tracker_config.yml
   batch_size: 1
   basemode: "idbased"

deploy/pphuman/config/infer_cfg_ppvehicle.yml

 crop_thresh: 0.5
 attr_thresh: 0.5
 kpt_thresh: 0.2
-visual: False
+visual: True
 warmup_frame: 50
 
 DET:
@@ -9,7 +9,7 @@ DET:
   batch_size: 1
 
 MOT:
-  model_dir: pipeline_model//mot_ppyoloe_s_36e_ppvehicle/
+  model_dir: output_inference/mot_ppyoloe_s_36e_ppvehicle/
   tracker_config: deploy/pphuman/config/tracker_config.yml
   batch_size: 1
   basemode: "idbased"
@@ -17,14 +17,14 @@ MOT:
 
 VEHICLE_PLATE:
   det_algorithm: "DB"
-  det_model_dir: "output/ch_PP-OCRv3_det_infer/"
+  det_model_dir: output_inference/ch_PP-OCRv3_det_infer/
   det_limit_side_len: 480
   det_limit_type: "max"
   rec_algorithm: "SVTR_LCNet"
-  rec_model_dir: "output/ch_PP-OCRv3_rec_infer/"
+  rec_model_dir: output_inference/ch_PP-OCRv3_rec_infer/
   rec_image_shape: [3, 48, 320]
   rec_batch_num: 6
-  word_dict_path: "deploy/pphuman/ppvehicle/rec_word_dict.txt"
+  word_dict_path: deploy/pphuman/ppvehicle/rec_word_dict.txt
   basemode: "idbased"
   enable: True
 
@@ -34,39 +34,8 @@ ATTR:
   basemode: "idbased"
   enable: False
 
-VIDEO_ACTION:
-  model_dir: output_inference/ppTSM
-  batch_size: 1
-  frame_len: 8
-  sample_freq: 7
-  short_size: 340
-  target_size: 320
-  basemode: "videobased"
-  enable: False
-
-SKELETON_ACTION:
-  model_dir: output_inference/STGCN
-  batch_size: 1
-  max_frames: 50
-  display_frames: 80
-  coord_size: [384, 512]
-  basemode: "skeletonbased"
-  enable: False
-
-ID_BASED_DETACTION:
-  model_dir: output_inference/detector
-  batch_size: 1
-  basemode: "idbased"
-  enable: False
-
-ID_BASED_CLSACTION:
-  model_dir: output_inference/classification
-  batch_size: 1
-  basemode: "idbased"
-  enable: False
-
 REID:
-  model_dir: output_inference/reid_model/
+  model_dir: output_inference/vehicle_reid_model/
   batch_size: 16
   basemode: "idbased"
   enable: False

deploy/pphuman/pipeline.py

@@ -45,7 +45,7 @@ from python.action_utils import KeyPointBuff, ActionVisualHelper
 from pipe_utils import argsparser, print_arguments, merge_cfg, PipeTimer
 from pipe_utils import get_test_images, crop_image_with_det, crop_image_with_mot, parse_mot_res, parse_mot_keypoint
 from python.preprocess import decode_image, ShortSizeScale
-from python.visualize import visualize_box_mask, visualize_attr, visualize_pose, visualize_action
+from python.visualize import visualize_box_mask, visualize_attr, visualize_pose, visualize_action, visualize_vehicleplate
 
 from pptracking.python.mot_sde_infer import SDE_Detector
 from pptracking.python.mot.visualize import plot_tracking_dict
@@ -618,7 +618,7 @@ class PipePredictor(object):
                                                   center_traj)  # visualize
                         writer.write(im)
                         if self.file_name is None:  # use camera_id
-                            cv2.imshow('PPHuman', im)
+                            cv2.imshow('PPHuman&&PPVehicle', im)
                             if cv2.waitKey(1) & 0xFF == ord('q'):
                                 break
                     continue
@@ -630,7 +630,6 @@ class PipePredictor(object):
                 if self.with_vehicleplate:
                     platelicense = self.vehicleplate_detector.get_platelicense(
                         crop_input)
-                    print("find plate license:{}".format(platelicense))
                     self.pipeline_res.update(platelicense, 'vehicleplate')
 
                 if self.with_attr:
@@ -828,6 +827,13 @@ class PipePredictor(object):
             image = visualize_attr(image, attr_res, boxes)
             image = np.array(image)
 
+        vehicleplate_res = result.get('vehicleplate')
+        if vehicleplate_res:
+            boxes = mot_res['boxes'][:, 1:]
+            image = visualize_vehicleplate(image, vehicleplate_res['plate'],
+                                           boxes)
+            image = np.array(image)
+
         kpt_res = result.get('kpt')
         if kpt_res is not None:
             image = visualize_pose(

deploy/pphuman/ppvehicle/vehicle_plate.py

@@ -499,10 +499,7 @@ class PlateRecognizer(object):
                 plate_images = get_rotate_crop_image(image_list[idx], box)
                 plate_texts = self.textrecognizer.predict_text([plate_images])
                 plate_pcar_list.append(plate_texts)
-                # print("plate text:{}".format(plate_texts))
             plate_text_list.append(plate_pcar_list)
-            # newimg = draw_boxes(image_list[idx], boxes_pcar)
-            # cv2.imwrite("vehicle_plate.jpg", newimg)
         return self.check_plate(plate_text_list)
 
     def check_plate(self, text_list):
@@ -513,11 +510,10 @@ class PlateRecognizer(object):
         ]
         plate_all = {"plate": []}
         for text_pcar in text_list:
-            platelicense = None
+            platelicense = ""
             for text_info in text_pcar:
                 text = text_info[0][0][0]
                 if len(text) > 2 and text[0] in simcode and len(text) < 10:
-                    # print("text:{} length:{}".format(text, len(text)))
                     platelicense = text
             plate_all["plate"].append(platelicense)
         return plate_all
@@ -529,7 +525,6 @@ def main():
     img_list = get_test_images(FLAGS.image_dir, FLAGS.image_file)
     for img in img_list:
         image = cv2.imread(img)
-        # image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
         results = detector.get_platelicense([image])
 
 

deploy/python/visualize.py

@@ -403,3 +403,38 @@ def visualize_action(im,
             thickness=text_thickness)
 
     return im
+
+
+def visualize_vehicleplate(im, results, boxes=None):
+    if isinstance(im, str):
+        im = Image.open(im)
+        im = np.ascontiguousarray(np.copy(im))
+        im = cv2.cvtColor(im, cv2.COLOR_RGB2BGR)
+    else:
+        im = np.ascontiguousarray(np.copy(im))
+
+    im_h, im_w = im.shape[:2]
+    text_scale = max(0.5, im.shape[0] / 3000.)
+    text_thickness = 1
+
+    line_inter = im.shape[0] / 40.
+    for i, res in enumerate(results):
+        if boxes is None:
+            text_w = 3
+            text_h = 1
+        else:
+            box = boxes[i]
+            text = res
+            if text == "":
+                continue
+            text_w = int(box[2])
+            text_h = int(box[5])
+            text_loc = (text_w, text_h)
+            cv2.putText(
+                im,
+                text,
+                text_loc,
+                cv2.FONT_ITALIC,
+                text_scale, (0, 255, 255),
+                thickness=text_thickness)
+    return im