在Tensorflow Object Detection API中避免重叠的边界框可以通过非最大抑制（Non-Maximum Suppression, NMS）来实现。NMS是一种常用的边界框去重技术，用于在目标检测中排除重叠的边界框。

以下是一个使用Tensorflow Object Detection API进行目标检测，并应用非最大抑制的代码示例：

import numpy as np
import tensorflow as tf
from object_detection.utils import visualization_utils as vis_util
from object_detection.utils import label_map_util

# 加载标签映射文件和模型
PATH_TO_LABELS = 'path/to/label_map.pbtxt'
PATH_TO_MODEL = 'path/to/frozen_inference_graph.pb'
NUM_CLASSES = 90

# 加载标签映射文件
label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES, use_display_name=True)
category_index = label_map_util.create_category_index(categories)

# 加载模型到内存
detection_graph = tf.Graph()
with detection_graph.as_default():
    od_graph_def = tf.GraphDef()
    with tf.gfile.GFile(PATH_TO_MODEL, 'rb') as fid:
        serialized_graph = fid.read()
        od_graph_def.ParseFromString(serialized_graph)
        tf.import_graph_def(od_graph_def, name='')

# 定义非最大抑制函数
def apply_nms(boxes, scores, iou_threshold=0.5):
    selected_indices = tf.image.non_max_suppression(boxes, scores, max_output_size=boxes.shape[0], iou_threshold=iou_threshold)
    selected_boxes = tf.gather(boxes, selected_indices)
    selected_scores = tf.gather(scores, selected_indices)
    return selected_boxes, selected_scores

# 运行目标检测
with detection_graph.as_default():
    with tf.Session(graph=detection_graph) as sess:
        # 获取输入和输出张量
        image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
        detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
        detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')
        detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')
        num_detections = detection_graph.get_tensor_by_name('num_detections:0')

        # 读取输入图像
        image = cv2.imread('path/to/input/image.jpg')
        image_expanded = np.expand_dims(image, axis=0)

        # 进行目标检测
        (boxes, scores, classes, num) = sess.run(
            [detection_boxes, detection_scores, detection_classes, num_detections],
            feed_dict={image_tensor: image_expanded})

        # 应用非最大抑制
        selected_boxes, selected_scores = sess.run(apply_nms(boxes[0], scores[0]))

        # 可视化结果
        vis_util.visualize_boxes_and_labels_on_image_array(
            image,
            np.squeeze(selected_boxes),
            np.squeeze(selected_classes).astype(np.int32),
            np.squeeze(selected_scores),
            category_index,
            use_normalized_coordinates=True,
            line_thickness=8)

        # 保存结果图像
        cv2.imwrite('path/to/output/image.jpg', image)

在上述代码中，首先加载了标签映射文件和模型，然后定义了一个apply_nms函数来应用非最大抑制。接下来，使用Tensorflow Object Detection API进行目标检测，然后调用apply_nms函数对检测结果进行去重。最后，使用vis_util.visualize_boxes_and_labels_on_image_array函数将结果可视化，并保存结果图像。

请确保将代码中的'path/to/label_map.pbtxt'和'path/to/frozen_inference_graph.pb'替换为正确的标签映射文件和模型的路径。另外，还需安装所需的依赖库，如numpy、tensorflow和opencv-python。