本文主要是介绍YOLOV----- ONNX 推理过程、可视化图片、保存检测到的目标，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！
一、代码
import os
import cv2
import numpy as np
import onnxruntime
import timeCLASSES = ['person'] # coco80类别class YOLOV5():def __init__(self, onnxpath):self.onnx_session = onnxruntime.InferenceSession(onnxpath)self.input_name = self.get_input_name()self.output_name = self.get_output_name()# -------------------------------------------------------# 获取输入输出的名字# -------------------------------------------------------def get_input_name(self):input_name = []for node in self.onnx_session.get_inputs():input_name.append(node.name)return input_namedef get_output_name(self):output_name = []for node in self.onnx_session.get_outputs():output_name.append(node.name)return output_name# -------------------------------------------------------# 输入图像# -------------------------------------------------------def get_input_feed(self, img_tensor):input_feed = {}for name in self.input_name:input_feed[name] = img_tensorreturn input_feed# -------------------------------------------------------# 1.cv2读取图像并resize# 2.图像转BGR2RGB和HWC2CHW# 3.图像归一化# 4.图像增加维度# 5.onnx_session 推理# -------------------------------------------------------# def inference(self, img_path):# img = cv2.imread(img_path)# or_img = cv2.resize(img, (640, 640))# img = or_img[:, :, ::-1].transpose(2, 0, 1) # BGR2RGB和HWC2CHW# img = img.astype(dtype=np.float32)# img /= 255.0# img = np.expand_dims(img, axis=0)# input_feed = self.get_input_feed(img)# pred = self.onnx_session.run(None, input_feed)[0]# return pred, or_img# def inference(self, img_path):# try:# img = cv2.imread(img_path)# or_img = cv2.resize(img, (640, 640))# img = or_img[:, :, ::-1].transpose(2, 0, 1) # BGR2RGB和HWC2CHW# img = img.astype(dtype=np.float32) / 255.0# img = np.expand_dims(img, axis=0)# input_feed = self.get_input_feed(img)# pred = self.onnx_session.run(None, input_feed)[0]# return pred, or_img# except Exception as e:# print(f"Error during inference: {e}")# return None, Nonedef inference(self, img):try:img = cv2.imread(img)or_img = img.copy()resized_img = cv2.resize(img, (640, 640)) # 根据模型要求进行resizeresized_img = resized_img[:, :, ::-1].transpose(2, 0, 1) # BGR2RGB和HWC2CHWresized_img = resized_img.astype(dtype=np.float32) / 255.0resized_img = np.expand_dims(resized_img, axis=0)input_feed = self.get_input_feed(resized_img)pred = self.onnx_session.run(None, input_feed)[0]return pred, or_img, resized_imgexcept Exception as e:print(f"Error during inference: {e}")return None, None, None# dets: array [x,6] 6个值分别为x1,y1,x2,y2,score,class
# thresh: 阈值
def nms(dets, thresh):x1 = dets[:, 0]y1 = dets[:, 1]x2 = dets[:, 2]y2 = dets[:, 3]# -------------------------------------------------------# 计算框的面积# 置信度从大到小排序# -------------------------------------------------------areas = (y2 - y1 + 1) * (x2 - x1 + 1)scores = dets[:, 4]keep = []index = scores.argsort()[::-1]while index.size > 0:i = index[0]keep.append(i)# -------------------------------------------------------# 计算相交面积# 1.相交# 2.不相交# -------------------------------------------------------x11 = np.maximum(x1[i], x1[index[1:]])y11 = np.maximum(y1[i], y1[index[1:]])x22 = np.minimum(x2[i], x2[index[1:]])y22 = np.minimum(y2[i], y2[index[1:]])w = np.maximum(0, x22 - x11 + 1)h = np.maximum(0, y22 - y11 + 1)overlaps = w * h# -------------------------------------------------------# 计算该框与其它框的IOU，去除掉重复的框，即IOU值大的框# IOU小于thresh的框保留下来# -------------------------------------------------------ious = overlaps / (areas[i] + areas[index[1:]] - overlaps)idx = np.where(ious <= thresh)[0]index = index[idx + 1]return keepdef xywh2xyxy(x):# [x, y, w, h] to [x1, y1, x2, y2]y = np.copy(x)y[:, 0] = x[:, 0] - x[:, 2] / 2y[:, 1] = x[:, 1] - x[:, 3] / 2y[:, 2] = x[:, 0] + x[:, 2] / 2y[:, 3] = x[:, 1] + x[:, 3] / 2return ydef filter_box(org_box, conf_thres, iou_thres): # 过滤掉无用的框# -------------------------------------------------------# 删除为1的维度# 删除置信度小于conf_thres的BOX# -------------------------------------------------------org_box = np.squeeze(org_box)conf = org_box[..., 4] > conf_thresbox = org_box[conf == True]# -------------------------------------------------------# 通过argmax获取置信度最大的类别# -------------------------------------------------------cls_cinf = box[..., 5:]cls = []for i in range(len(cls_cinf)):cls.append(int(np.argmax(cls_cinf[i])))all_cls = list(set(cls))# -------------------------------------------------------# 分别对每个类别进行过滤# 1.将第6列元素替换为类别下标# 2.xywh2xyxy 坐标转换# 3.经过非极大抑制后输出的BOX下标# 4.利用下标取出非极大抑制后的BOX# -------------------------------------------------------output = []for i in range(len(all_cls)):curr_cls = all_cls[i]curr_cls_box = []curr_out_box = []for j in range(len(cls)):if cls[j] == curr_cls:box[j][5] = curr_clscurr_cls_box.append(box[j][:6])curr_cls_box = np.array(curr_cls_box)# curr_cls_box_old = np.copy(curr_cls_box)curr_cls_box = xywh2xyxy(curr_cls_box)curr_out_box = nms(curr_cls_box, iou_thres)for k in curr_out_box:output.append(curr_cls_box[k])output = np.array(output)return outputdef draw(image, box_data):# -------------------------------------------------------# 取整，方便画框# -------------------------------------------------------boxes = box_data[..., :4].astype(np.int32)scores = box_data[..., 4]classes = box_data[..., 5].astype(np.int32)for box, score, cl in zip(boxes, scores, classes):top, left, right, bottom = boxprint('class: {}, score: {}'.format(CLASSES[cl], score))print('box coordinate left,top,right,down: [{}, {}, {}, {}]'.format(top, left, right, bottom))cv2.rectangle(image, (top, left), (right, bottom), (0, 0, 255), 2) #红色#cv2.rectangle(image, (top, left), (right, bottom), (255, 0, 0), 2) #蓝色cv2.putText(image, '{0} {1:.2f}'.format(CLASSES[cl], score),(top, left),cv2.FONT_HERSHEY_SIMPLEX,0.6, (0, 0, 255), 2)#可能存在多个人体，提取置信度最高的人体框
def extract_person(image, box_data):image = np.squeeze(image, axis=0) # 形状变为 (3, 640, 640)# 转换为 (H, W, C) 格式image = np.transpose(image, (1, 2, 0))print(f"Resized image shape: {image.shape}")cv2.imshow('read Image',image)cv2.waitKey(0)cv2.destroyAllWindows()boxes = box_data[..., :4].astype(np.int32)scores = box_data[..., 4]classes = box_data[..., 5].astype(np.int32)if len(boxes) > 0:max_index = np.argmax(scores)#top, left, right, bottom = boxes[max_index]left,top, right, bottom = boxes[max_index]# 打印原始坐标print(f"Original Box coordinates: top={top}, left={left}, right={right}, bottom={bottom}")# 提取目标区域person = image[top:bottom, left:right]print(f"Extracted person image size: {person.shape}")cv2.imshow('extracted person',person)cv2.waitKey(0)cv2.destroyAllWindows()# 保存图像cv2.imwrite('extract.png', person*255)return personif __name__ == "__main__":onnx_path = r'E:\detect_person\person.onnx'model = YOLOV5(onnx_path)output, or_img, resize_img = model.inference(r"G:\depth_detect\huang2\huang1.png")outbox = filter_box(output, 0.35, 0.35)if len(outbox) > 0:#原图画框可视化#draw(or_img, outbox)#提取目标区域extract = extract_person(resize_img,outbox)else:print("No objects detected.")这篇关于YOLOV----- ONNX 推理过程、可视化图片、保存检测到的目标的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！
