完成第一篇 darknet环境配置后,可基于python进行YOLOv4的inference,查看YOLOv4的检测结果。需要用到的库有 darknet目录下的
darknet.py
文件,以及编译出来的
libdarknet.so
。
1. 需要用的的库
import os
import cv2
import numpy as np
import random
import darknet
需要导入的库主要是opencv-python和darknet,darknet即darknet.py
文件。
2. 加载网络
加载网络需要读取cfg文件,weights文件,data文件和names文件,如果缺少文件,会报读取失败的错误。yolov4.cfg
和coco.data
文件可以在darknet文件夹下的cfg文件夹中找到,coco.names
文件可以在data文件夹中找到,yolov4.weights
文件可以在这篇文章中介绍的网盘中下载。
netMain = None
metaMain = None
altNames = None
configPath = "./cfg/yolov4.cfg"
weightPath = "./weights/yolov4.weights"
metaPath = "./data/coco.data"
if not os.path.exists(configPath):
raise ValueError("Invalid config path `" + os.path.abspath(configPath)+"`")
if not os.path.exists(weightPath):
raise ValueError("Invalid weight path `" + os.path.abspath(weightPath)+"`")
if not os.path.exists(metaPath):
raise ValueError("Invalid data file path `" + os.path.abspath(metaPath)+"`")
if netMain is None:
netMain = darknet.load_net_custom(configPath.encode("ascii"), weightPath.encode("ascii"), 0, 1) # batch size = 1
if metaMain is None:
metaMain = darknet.load_meta(metaPath.encode("ascii"))
if altNames is None:
try:
with open(metaPath) as metaFH:
metaContents = metaFH.read()
import re
match = re.search("names *= *(.*)$", metaContents, re.IGNORECASE | re.MULTILINE)
if match:
result = match.group(1)
else:
result = None
try:
if os.path.exists(result):
with open(result) as namesFH:
namesList = namesFH.read().strip().split("\n")
altNames = [x.strip() for x in namesList]
except TypeError:
pass
except Exception:
pass
3. 加载图片
首先通过opencv读取图片,读取图片之后需要将图片由BGR格式转为RGB格式,接着resize到608x608
大小并将图片转为darknet的IMAGE类型,输入到网络之后,检测结果保存在detections里面。
image_name = './data/dog.jpg'
src_img = cv2.imread(image_name)
bgr_img = src_img[:, :, ::-1]
height, width = bgr_img.shape[:2]
rsz_img = cv2.resize(bgr_img, (darknet.network_width(netMain), darknet.network_height(netMain)),
interpolation=cv2.INTER_LINEAR)
darknet_image, _ = darknet.array_to_image(rsz_img)
detections = darknet.detect_image(netMain, metaMain, darknet_image, thresh=0.25)
4. 可视化结果
需要先将检测结果转为x1y1x2y2的形式,再使用opencv的rectangle函数画图。
# convert xywh to xyxy
def convert_back(x, y, w, h):
xmin = int(round(x - (w / 2)))
xmax = int(round(x + (w / 2)))
ymin = int(round(y - (h / 2)))
ymax = int(round(y + (h / 2)))
return xmin, ymin, xmax, ymax
# Plotting functions
def plot_one_box(x, img, color=None, label=None, line_thickness=None):
# Plots one bounding box on image img
tl = line_thickness or round(0.001 * max(img.shape[0:2])) + 1 # line thickness
color = color or [random.randint(0, 255) for _ in range(3)]
c1, c2 = (int(x[0]), int(x[1])), (int(x[2]), int(x[3]))
cv2.rectangle(img, c1, c2, color, thickness=tl)
if label:
tf = max(tl - 1, 1) # font thickness
t_size = cv2.getTextSize(label, 0, fontScale=tl / 3, thickness=tf)[0]
c2 = c1[0] + t_size[0], c1[1] - t_size[1] - 3
cv2.rectangle(img, c1, c2, color, -1) # filled
cv2.putText(img, label, (c1[0], c1[1] - 2), 0, tl / 3, [225, 255, 255], thickness=tf, lineType=cv2.LINE_AA)
random.seed(1)
colors = [[random.randint(0, 255) for _ in range(3)] for _ in range(metaMain.classes)]
for detection in detections:
x, y, w, h = detection[2][0], \
detection[2][1], \
detection[2][2], \
detection[2][3]
conf = detection[1]
x *= width / darknet.network_width(netMain)
w *= width / darknet.network_width(netMain)
y *= height / darknet.network_height(netMain)
h *= height / darknet.network_height(netMain)
xyxy = np.array([x - w / 2, y - h / 2, x + w / 2, y + h / 2])
label = detection[0].decode()
index = altNames.index(label)
label = f'{label} {conf:.2f}'
plot_one_box(xyxy, src_img, label=label, color=colors[index % metaMain.classes])
cv2.imwrite('result.jpg', src_img)
最终结果保存在result.jpg
图片里,结果如下图所示。
结果显示图片中dog的置信度为0.98,bicycle的置信度为0.92,truck的置信度为0.92,pottedplant的置信度没有显示,通过运行代码可知为0.34。通过darknet命令./darknet detector test cfg/coco.data cfg/yolov4.cfg yolov4.weights -thresh 0.25 ./data/dog.jpg
运行的结果如下,结果基本一致。
data/dog.jpg: Predicted in 21.041000 milli-seconds.
bicycle: 92%
dog: 98%
truck: 92%
pottedplant: 33%