paddledet 训练旋转目标检测 ppyoloe-r 训练自己的数据集

Posted 2022-12-15 东东就是我

1.数据转换
labelme2coco，原来是labelme标注的points 通过opencv转为4个坐标

# encoding=utf-8
import argparse
import collections
import datetime
import glob
import json
import os
import os.path as osp
import sys
import uuid
import cv2
import imgviz
import numpy as np

import labelme

try:
    import pycocotools.mask
except ImportError:
    print("Please install pycocotools:\\n\\n    pip install pycocotools\\n")
    sys.exit(1)


def main():
    input_dir='G:/customer/visionary_s_3d_dete/zhixi'
    output_dir='dataset/zhixi'
    os.makedirs(output_dir)
    os.makedirs(osp.join(output_dir, "JPEGImages"))

    data = dict(
        images=[
            # license, url, file_name, height, width, date_captured, id
        ],
        annotations=[
            # segmentation, area, iscrowd, image_id, bbox, category_id, id
        ],
        categories=[
            # supercategory, id, name
        ],
    )
    data["categories"].append(
        dict(supercategory=None, id=0, name='sack',)
    )

    out_ann_file = osp.join(output_dir, "annotations.json")
    label_files = glob.glob(osp.join(input_dir, "*.json"))
    for image_id, filename in enumerate(label_files):
        print("Generating dataset from:", filename)

        label_file = labelme.LabelFile(filename=filename)

        base = osp.splitext(osp.basename(filename))[0]
        out_img_file = osp.join(output_dir, "JPEGImages", base + ".jpg")

        img = labelme.utils.img_data_to_arr(label_file.imageData)
        imgviz.io.imsave(out_img_file, img)
        data["images"].append(
            dict(
                license=0,
                url=None,
                file_name=osp.relpath(out_img_file, osp.dirname(out_ann_file)),
                height=img.shape[0],
                width=img.shape[1],
                date_captured=None,
                id=image_id,
            )
        )

        masks =   # for area
        segmentations = collections.defaultdict(list)  # for segmentation
        for shape in label_file.shapes:
            points = shape["points"]
            label = shape["label"]
            group_id = shape.get("group_id")
            shape_type = shape.get("shape_type")
            mask = labelme.utils.shape_to_mask(
                img.shape[:2], points, shape_type
            )

            if group_id is None:
                group_id = uuid.uuid1()

            instance = (label, group_id)

            if instance in masks:
                masks[instance] = masks[instance] | mask
            else:
                masks[instance] = mask

            if shape_type == "rectangle":
                (x1, y1), (x2, y2) = points
                x1, x2 = sorted([x1, x2])
                y1, y2 = sorted([y1, y2])
                points = [x1, y1, x2, y1, x2, y2, x1, y2]
            if shape_type == "circle":
                (x1, y1), (x2, y2) = points
                r = np.linalg.norm([x2 - x1, y2 - y1])
                # r(1-cos(a/2))<x, a=2*pi/N => N>pi/arccos(1-x/r)
                # x: tolerance of the gap between the arc and the line segment
                n_points_circle = max(int(np.pi / np.arccos(1 - 1 / r)), 12)
                i = np.arange(n_points_circle)
                x = x1 + r * np.sin(2 * np.pi / n_points_circle * i)
                y = y1 + r * np.cos(2 * np.pi / n_points_circle * i)
                points = np.stack((x, y), axis=1).flatten().tolist()
            elif shape_type=="polygon":
                points=np.float32(points)
                rect = cv2.minAreaRect(points)  # 最小外接矩形
                points = cv2.boxPoints(rect).flatten()
                points=points.tolist()
                # points = np.asarray(box).flatten().tolist()

            segmentations[instance].append(points)
        segmentations = dict(segmentations)

        for instance, mask in masks.items():
            mask = np.asfortranarray(mask.astype(np.uint8))
            mask = pycocotools.mask.encode(mask)
            area = float(pycocotools.mask.area(mask))
            bbox = pycocotools.mask.toBbox(mask).flatten().tolist()

            data["annotations"].append(
                dict(
                    id=len(data["annotations"]),
                    image_id=image_id,
                    category_id=0,
                    segmentation=segmentations[instance],
                    area=area,
                    bbox=bbox,
                    iscrowd=0,
                )
            )
    with open(out_ann_file, "w") as f:
        json.dump(data, f)


if __name__ == "__main__":
    main()

2.修改data.yml文件
这个路径要根据自己实际修改，实在不行就debug看看哪里出问题

metric: RBOX
num_classes: 15

TrainDataset:
  !COCODataSet
    image_dir:
    anno_path: annotations/annotations.json
    dataset_dir: ../dataset/zhixi
    data_fields: ['image', 'gt_bbox', 'gt_class', 'is_crowd', 'gt_poly']

EvalDataset:
  !COCODataSet
    image_dir:
    anno_path: annotations/annotations.json
    dataset_dir: ../dataset/zhixi
    data_fields: ['image', 'gt_bbox', 'gt_class', 'is_crowd', 'gt_poly']

TestDataset:
  !ImageFolder
    anno_path: annotations/annotations.json
    dataset_dir: ../dataset/zhixi

3.训练报错
RuntimeError: (PreconditionNotMet) The third-party dynamic library (cublas64_102.dll;cublas64_10.dll) that Paddle depends on is not configured correctly. (error code is 126)

解决：路径为C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v10.0\\bin

在bin路径下将cublas64_100.dll重命名为cublas64_10.dll

在bin路径下将cusolver64_100重命名为cusolver64_10
————————————————
版权声明：本文为CSDN博主「李伯爵的指间沙」的原创文章，遵循CC 4.0 BY-SA版权协议，转载请附上原文出处链接及本声明。
原文链接：https://blog.csdn.net/m0_37690102/article/details/123474171