Python+OpenCV读取json文件并显示

import os import json import cv2 underwater_classes = ['holothurian', 'echinus', 'scallop', 'starfish'] root_dir = '/media/chiyukunpeng/CHENPENG01/contest/underwater_object_detection20200827/data/test-A-image/' images = [os.path.join(root_dir, f) for f in os.listdir(root_dir)] label_file = "/media/chiyukunpeng/CHENPENG01/contest/underwater_object_detection20200827/results/cas_x101.bbox.json" test_json = json.load(open(label_file, 'r')) raw_label_file = "/media/chiyukunpeng/CHENPENG01/contest/underwater_object_detection20200827/data/annotations/test-A.json" test_json_raw = json.load(open(raw_label_file, "r")) # convert image_id to image_filename imgid2name = {} for imageinfo in test_json_raw['images']: imgid = imageinfo['id'] imgid2name[imgid] = imageinfo['file_name'] for anno in test_json: img_id = anno['image_id'] filename = imgid2name[img_id] anno['image_id'] = filename for image in images: img = cv2.imread(image) image = str(os.path.basename(image)) for anno in test_json: if anno['image_id'] == image: xmin, ymin, xmax, ymax = int(anno['bbox'][0]),int(anno['bbox'][1]),int(anno['bbox'][2]),int(anno['bbox'][3]) confidence = round(anno['score'], 2) class_id = int(anno['category_id']) class_name = underwater_classes[class_id-1] cv2.rectangle(img, (xmin, ymin), (xmax, ymax), (0, 255, 0), 3) cv2.putText(img, str(confidence), (xmin, ymin - 20), cv2.FONT_HERSHEY_COMPLEX, 3, (255, 0, 0), 3) cv2.putText(img, str(class_name), (xmin+20, ymin - 20), cv2.FONT_HERSHEY_COMPLEX, 3, (255, 0, 0), 3) cv2.imshow('img', img) cv2.waitKey(0)