kuangxuan.py

import os
import numpy as np
import cv2
import cv2 as cv
import argparse

weightsPath = "D:/git/work/keras-yolo3/yolov3.weights"
configPath = "D:/git/work/keras-yolo3/yolov3.cfg"
labelsPath = "D:/git/work/keras-yolo3/model_data/coco_classes.txt"
rootdir = "D:/git/work/keras-yolo3/images"  # 图像读取地址
savepath = "D:/git/work/keras-yolo3/kuangxuanimages"  # 图像保存地址

# 初始化一些参数
LABELS = open(labelsPath).read().strip().split("\n")  # 物体类别
#print(LABELS)
COLORS = np.random.randint(0, 255, size=(len(LABELS), 3), dtype="uint8")  # 颜色
#print(COLORS)

filelist = os.listdir(rootdir)  # 打开对应的文件夹
total_num = len(filelist)  # 得到文件夹中图像的个数
#print(total_num)
# 如果输出的文件夹不存在，创建即可
if not os.path.isdir(savepath):
    os.makedirs(savepath)

for (dirpath, dirnames, filenames) in os.walk(rootdir):
    for filename in filenames:
        # 必须将boxes在遍历新的图片后初始化
        boxes = []
        confidences = []
        classIDs = []
        net = cv2.dnn.readNetFromDarknet(configPath, weightsPath)
        path = os.path.join(dirpath, filename)
        image = cv.imread(path)
        (H, W) = image.shape[:2]
        # 得到 YOLO需要的输出层
        #print(H, W)
        ln = net.getLayerNames()
        ln = [ln[i[0] - 1] for i in net.getUnconnectedOutLayers()]
        # 从输入图像构造一个blob，然后通过加载的模型，给我们提供边界框和相关概率
        blob = cv2.dnn.blobFromImage(image, 1 / 255.0, (416, 416), swapRB=True, crop=False)
        net.setInput(blob)
        layerOutputs = net.forward(ln)
        # 在每层输出上循环
        for output in layerOutputs:
            # 对每个检测进行循环
            for detection in output:
                scores = detection[5:]
                classID = np.argmax(scores)
                confidence = scores[classID]
                # 过滤掉那些置信度较小的检测结果
                if confidence > 0.5:
                    # 框后接框的宽度和高度
                    box = detection[0:4] * np.array([W, H, W, H])
                    (centerX, centerY, width, height) = box.astype("int")
                    # 边框的左上角
                    x = int(centerX - (width / 2))
                    y = int(centerY - (height / 2))
                    # 更新检测出来的框
                    # 批量检测图片注意此处的boxes在每一次遍历的时候要初始化，否则检测出来的图像框会叠加
                    boxes.append([x, y, int(width), int(height)])
                    confidences.append(float(confidence))
                    classIDs.append(classID)
        # 极大值抑制
        idxs = cv2.dnn.NMSBoxes(boxes, confidences, 0.2, 0.3)
        #print(confidence)
        k = -1
        if len(idxs) > 0:
            # for k in range(0,len(boxes)):
            for i in idxs.flatten():
                (x, y) = (boxes[i][0], boxes[i][1])
                (w, h) = (boxes[i][2], boxes[i][3])
                # 在原图上绘制边框和类别
                color = [int(c) for c in COLORS[classIDs[i]]]
                # image是原图，     左上点坐标， 右下点坐标， 颜色， 画线的宽度
                cv2.rectangle(image, (x, y), (x + w, y + h), color, 2)
                text = "{}: {:.4f}".format(LABELS[classIDs[i]], confidences[i])
                # 各参数依次是：图片，添加的文字，左上角坐标(整数)，字体，        字体大小，颜色，字体粗细
                cv2.putText(image, text, (x, y - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.5, color, 2)
                # 图像裁剪注意坐标要一一对应
                # 图片裁剪 裁剪区域【Ly:Ry,Lx:Rx】
                cut = image[y:(y + h), x:(x + w)]
                # boxes的长度即为识别出来的车辆个数，利用boxes的长度来定义裁剪后车辆的路径名称
                if k < len(boxes):
                    k = k + 1
                # 从字母a开始每次+1
                t = chr(ord("a") + k)
                # 写入文件夹，这块写入的时候不支持int（我也不知道为啥），所以才用的字母
                cv.imwrite(savepath + "/" + filename.split(".")[0] + "_" + t + ".jpg", cut)