RK3568 NPU RKNN（五）：RKNN-ToolKit-lite2板端推理

1、前言

本文仅记录本人学习过程，不具备教学指导意义。

2、目标

之前提到过，RKNN-Toolkit2-Lite2是RKNN-Toolkit2的阉割版，只保留了推理功能，可以直接运行在板卡上。本文目标将下载安装rknn-toolkit-lite2，使用野火提供的示例程序，体验 rknn-toolkit-lite2 在板卡端推理。

3、安装RKNN-ToolKit-lite2

这里使用的是ubuntu系统的板卡，以下命令都是在板卡端执行。

3.1、安装环境

#安装python工具，安装相关依赖和软件包等
sudo apt update
sudo apt-get install python3-dev python3-pip gcc
sudo apt install -y python3-opencv python3-numpy python3-setuptools

3.2、安装RKNN-ToolKit-lite2

# 获取 RKNN-ToolKit-lite2 工程文件
# 可以官网获取：https://github.com/airockchip/rknn-toolkit2/tree/master/rknn-toolkit-lite2
# 这里使用野火提供的
git clone https://gitee.com/LubanCat/lubancat_ai_manual_code.git# 安装 RKNN-ToolKit-lite2 软件工具包
# 我的python版本是3.8
pip3 install packages/rknn_toolkit_lite2-1.5.0-cp38-cp38-linux_aarch64.whl

3.3、验证

root@lubancat:~/lubancat_ai_manual_code/dev_env/rknn_toolkit_lite2# python3
Python 3.8.10 (default, Mar 18 2025, 20:04:55)
[GCC 9.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from rknnlite.api import RKNNLite
>>>

4、完整的测试程序

import urllib
import time
import sys
import numpy as np
import cv2
import platform
from rknnlite.api import RKNNLiteRK3566_RK3568_RKNN_MODEL = 'yolov5s_for_rk3566_rk3568.rknn'
RK3588_RKNN_MODEL = 'yolov5s_for_rk3588.rknn'
RK3562_RKNN_MODEL = 'yolov5s_for_rk3562.rknn'
IMG_PATH = './bus.jpg'OBJ_THRESH = 0.25
NMS_THRESH = 0.45
IMG_SIZE = 640CLASSES = ("person", "bicycle", "car", "motorbike ", "aeroplane ", "bus ", "train", "truck ", "boat", "traffic light","fire hydrant", "stop sign ", "parking meter", "bench", "bird", "cat", "dog ", "horse ", "sheep", "cow", "elephant","bear", "zebra ", "giraffe", "backpack", "umbrella", "handbag", "tie", "suitcase", "frisbee", "skis", "snowboard", "sports ball", "kite","baseball bat", "baseball glove", "skateboard", "surfboard", "tennis racket", "bottle", "wine glass", "cup", "fork", "knife ","spoon", "bowl", "banana", "apple", "sandwich", "orange", "broccoli", "carrot", "hot dog", "pizza ", "donut", "cake", "chair", "sofa","pottedplant", "bed", "diningtable", "toilet ", "tvmonitor", "laptop	", "mouse	", "remote ", "keyboard ", "cell phone", "microwave ","oven ", "toaster", "sink", "refrigerator ", "book", "clock", "vase", "scissors ", "teddy bear ", "hair drier", "toothbrush ")# decice tree for rk356x/rk3588
DEVICE_COMPATIBLE_NODE = '/proc/device-tree/compatible'def get_host():# get platform and device typesystem = platform.system()machine = platform.machine()os_machine = system + '-' + machineif os_machine == 'Linux-aarch64':try:with open(DEVICE_COMPATIBLE_NODE) as f:device_compatible_str = f.read()if 'rk3588' in device_compatible_str:host = 'RK3588'elif 'rk3562' in device_compatible_str:host = 'RK3562'else:host = 'RK3566_RK3568'except IOError:print('Read device node {} failed.'.format(DEVICE_COMPATIBLE_NODE))exit(-1)else:host = os_machinereturn hostdef sigmoid(x):return 1 / (1 + np.exp(-x))def xywh2xyxy(x):# Convert [x, y, w, h] to [x1, y1, x2, y2]y = np.copy(x)y[:, 0] = x[:, 0] - x[:, 2] / 2  # top left xy[:, 1] = x[:, 1] - x[:, 3] / 2  # top left yy[:, 2] = x[:, 0] + x[:, 2] / 2  # bottom right xy[:, 3] = x[:, 1] + x[:, 3] / 2  # bottom right yreturn ydef process(input, mask, anchors):anchors = [anchors[i] for i in mask]grid_h, grid_w = map(int, input.shape[0:2])box_confidence = sigmoid(input[..., 4])box_confidence = np.expand_dims(box_confidence, axis=-1)box_class_probs = sigmoid(input[..., 5:])box_xy = sigmoid(input[..., :2])*2 - 0.5col = np.tile(np.arange(0, grid_w), grid_w).reshape(-1, grid_w)row = np.tile(np.arange(0, grid_h).reshape(-1, 1), grid_h)col = col.reshape(grid_h, grid_w, 1, 1).repeat(3, axis=-2)row = row.reshape(grid_h, grid_w, 1, 1).repeat(3, axis=-2)grid = np.concatenate((col, row), axis=-1)box_xy += gridbox_xy *= int(IMG_SIZE/grid_h)box_wh = pow(sigmoid(input[..., 2:4])*2, 2)box_wh = box_wh * anchorsbox = np.concatenate((box_xy, box_wh), axis=-1)return box, box_confidence, box_class_probsdef filter_boxes(boxes, box_confidences, box_class_probs):"""Filter boxes with box threshold. It's a bit different with origin yolov5 post process!# Argumentsboxes: ndarray, boxes of objects.box_confidences: ndarray, confidences of objects.box_class_probs: ndarray, class_probs of objects.# Returnsboxes: ndarray, filtered boxes.classes: ndarray, classes for boxes.scores: ndarray, scores for boxes."""boxes = boxes.reshape(-1, 4)box_confidences = box_confidences.reshape(-1)box_class_probs = box_class_probs.reshape(-1, box_class_probs.shape[-1])_box_pos = np.where(box_confidences >= OBJ_THRESH)boxes = boxes[_box_pos]box_confidences = box_confidences[_box_pos]box_class_probs = box_class_probs[_box_pos]class_max_score = np.max(box_class_probs, axis=-1)classes = np.argmax(box_class_probs, axis=-1)_class_pos = np.where(class_max_score >= OBJ_THRESH)boxes = boxes[_class_pos]classes = classes[_class_pos]scores = (class_max_score* box_confidences)[_class_pos]return boxes, classes, scoresdef nms_boxes(boxes, scores):"""Suppress non-maximal boxes.# Argumentsboxes: ndarray, boxes of objects.scores: ndarray, scores of objects.# Returnskeep: ndarray, index of effective boxes."""x = boxes[:, 0]y = boxes[:, 1]w = boxes[:, 2] - boxes[:, 0]h = boxes[:, 3] - boxes[:, 1]areas = w * horder = scores.argsort()[::-1]keep = []while order.size > 0:i = order[0]keep.append(i)xx1 = np.maximum(x[i], x[order[1:]])yy1 = np.maximum(y[i], y[order[1:]])xx2 = np.minimum(x[i] + w[i], x[order[1:]] + w[order[1:]])yy2 = np.minimum(y[i] + h[i], y[order[1:]] + h[order[1:]])w1 = np.maximum(0.0, xx2 - xx1 + 0.00001)h1 = np.maximum(0.0, yy2 - yy1 + 0.00001)inter = w1 * h1ovr = inter / (areas[i] + areas[order[1:]] - inter)inds = np.where(ovr <= NMS_THRESH)[0]order = order[inds + 1]keep = np.array(keep)return keepdef yolov5_post_process(input_data):masks = [[0, 1, 2], [3, 4, 5], [6, 7, 8]]anchors = [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45],[59, 119], [116, 90], [156, 198], [373, 326]]boxes, classes, scores = [], [], []for input, mask in zip(input_data, masks):b, c, s = process(input, mask, anchors)b, c, s = filter_boxes(b, c, s)boxes.append(b)classes.append(c)scores.append(s)boxes = np.concatenate(boxes)boxes = xywh2xyxy(boxes)classes = np.concatenate(classes)scores = np.concatenate(scores)nboxes, nclasses, nscores = [], [], []for c in set(classes):inds = np.where(classes == c)b = boxes[inds]c = classes[inds]s = scores[inds]keep = nms_boxes(b, s)nboxes.append(b[keep])nclasses.append(c[keep])nscores.append(s[keep])if not nclasses and not nscores:return None, None, Noneboxes = np.concatenate(nboxes)classes = np.concatenate(nclasses)scores = np.concatenate(nscores)return boxes, classes, scoresdef draw(image, boxes, scores, classes):"""Draw the boxes on the image.# Argument:image: original image.boxes: ndarray, boxes of objects.classes: ndarray, classes of objects.scores: ndarray, scores of objects.all_classes: all classes name."""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))top = int(top)left = int(left)right = int(right)bottom = int(bottom)cv2.rectangle(image, (top, left), (right, bottom), (255, 0, 0), 2)cv2.putText(image, '{0} {1:.2f}'.format(CLASSES[cl], score),(top, left - 6),cv2.FONT_HERSHEY_SIMPLEX,0.6, (0, 0, 255), 2)def letterbox(im, new_shape=(640, 640), color=(0, 0, 0)):# Resize and pad image while meeting stride-multiple constraintsshape = im.shape[:2]  # current shape [height, width]if isinstance(new_shape, int):new_shape = (new_shape, new_shape)# Scale ratio (new / old)r = min(new_shape[0] / shape[0], new_shape[1] / shape[1])# Compute paddingratio = r, r  # width, height ratiosnew_unpad = int(round(shape[1] * r)), int(round(shape[0] * r))dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1]  # wh paddingdw /= 2  # divide padding into 2 sidesdh /= 2if shape[::-1] != new_unpad:  # resizeim = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR)top, bottom = int(round(dh - 0.1)), int(round(dh + 0.1))left, right = int(round(dw - 0.1)), int(round(dw + 0.1))im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color)  # add borderreturn im, ratio, (dw, dh)if __name__ == '__main__':host_name = get_host()if host_name == 'RK3566_RK3568':rknn_model = RK3566_RK3568_RKNN_MODELelif host_name == 'RK3562':rknn_model = RK3562_RKNN_MODELelif host_name == 'RK3588':rknn_model = RK3588_RKNN_MODELelse:print("This demo cannot run on the current platform: {}".format(host_name))exit(-1)# Create RKNN objectrknn_lite = RKNNLite()# load RKNN modelprint('--> Load RKNN model')ret = rknn_lite.load_rknn(rknn_model)if ret != 0:print('Load RKNN model failed')exit(ret)print('done')# Init runtime environmentprint('--> Init runtime environment')# run on RK356x/RK3588 with Debian OS, do not need specify target.if host_name == 'RK3588':ret = rknn_lite.init_runtime(core_mask=RKNNLite.NPU_CORE_0)else:ret = rknn_lite.init_runtime()if ret != 0:print('Init runtime environment failed!')exit(ret)print('done')# Set inputsimg = cv2.imread(IMG_PATH)#img, ratio, (dw, dh) = letterbox(img, new_shape=(IMG_SIZE, IMG_SIZE))img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)img = cv2.resize(img, (IMG_SIZE, IMG_SIZE))# Inferenceprint('--> Running model')outputs = rknn_lite.inference(inputs=[img])#np.save('./onnx_yolov5_0.npy', outputs[0])#np.save('./onnx_yolov5_1.npy', outputs[1])#np.save('./onnx_yolov5_2.npy', outputs[2])print('done')# post processinput0_data = outputs[0]input1_data = outputs[1]input2_data = outputs[2]input0_data = input0_data.reshape([3, -1]+list(input0_data.shape[-2:]))input1_data = input1_data.reshape([3, -1]+list(input1_data.shape[-2:]))input2_data = input2_data.reshape([3, -1]+list(input2_data.shape[-2:]))input_data = list()input_data.append(np.transpose(input0_data, (2, 3, 0, 1)))input_data.append(np.transpose(input1_data, (2, 3, 0, 1)))input_data.append(np.transpose(input2_data, (2, 3, 0, 1)))boxes, classes, scores = yolov5_post_process(input_data)img_1 = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)if boxes is not None:draw(img_1, boxes, scores, classes)# show outputcv2.imwrite("out.jpg", img_1)#cv2.imshow("post process result", img_1)#cv2.waitKey(0)#cv2.destroyAllWindows()rknn_lite.release()

5、运行测试程序

# 板卡端执行
cd lubancat_ai_manual_code/dev_env/rknn_toolkit_lite2/examples/yolov5_inference
python3 test.py

查看最后生成的out.jpg：

6、程序拆解

1. 创建rknnlite对象

rknn_lite = RKNNLite()

2. 加载rknn模型

rknn_lite.load_rknn(rknn_model)

3. 初始化运行环境

rknn_lite.init_runtime()

4. 模型推理（Inference）

outputs = rknn.inference(inputs=[img])

5. 后处理（Post-process）

# post process
input0_data = outputs[0]
input1_data = outputs[1]
input2_data = outputs[2]input0_data = input0_data.reshape([3, -1]+list(input0_data.shape[-2:]))
input1_data = input1_data.reshape([3, -1]+list(input1_data.shape[-2:]))
input2_data = input2_data.reshape([3, -1]+list(input2_data.shape[-2:]))input_data = list()
input_data.append(np.transpose(input0_data, (2, 3, 0, 1)))
input_data.append(np.transpose(input1_data, (2, 3, 0, 1)))
input_data.append(np.transpose(input2_data, (2, 3, 0, 1)))boxes, classes, scores = yolov5_post_process(input_data)img_1 = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
if boxes is not None:draw(img_1, boxes, scores, classes)# show output
cv2.imwrite("out.jpg", img_1)
#cv2.imshow("post process result", img_1)
#cv2.waitKey(0)
#cv2.destroyAllWindows()

7、总结

参考文章：

https://doc.embedfire.com/linux/rk356x/Ai/zh/latest/lubancat_ai/env/toolkit_lite2.html#id3