基于ncnn模型实现android车辆识别车牌角度矫正车牌文本识别车牌颜色识别源码+模型+项目说明.zip

// Tencent is pleased to support the open source community by making ncnn available. // // Copyright (C) 2020 THL A29 Limited, a Tencent company. All rights reserved. // // Licensed under the BSD 3-Clause License (the "License"); you may not use this file except // in compliance with the License. You may obtain a copy of the License at // // https://opensource.org/licenses/BSD-3-Clause // // Unless required by applicable law or agreed to in writing, software distributed // under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR // CONDITIONS OF ANY KIND, either express or implied. See the License for the // specific language governing permissions and limitations under the License. #include <android/asset_manager_jni.h> #include <android/bitmap.h> #include <android/log.h> #include <jni.h> #include <string> #include <vector> #include <iostream> using namespace std; // ncnn #include "layer.h" #include "net.h" #include "benchmark.h" #include <opencv2/core/core.hpp> #include <opencv2/highgui/highgui.hpp> #include <opencv2/imgproc.hpp> using namespace cv; static ncnn::UnlockedPoolAllocator g_blob_pool_allocator; static ncnn::PoolAllocator g_workspace_pool_allocator; #define ASSERT(status, ret) if (!(status)) { return ret; } #define ASSERT_FALSE(status) ASSERT(status, false) static ncnn::Net yolov5; static ncnn::Net crnn; static ncnn::Net color_net; //static string plate_chars[68] = { "京", "沪", "津", "渝", "冀", "晋", "蒙", "辽", "吉", "黑", // "苏", "浙", "皖", "闽", "赣", "鲁", "豫", "鄂", "湘", "粤", // "桂", "琼", "川", "贵", "云", "藏", "陕", "甘", "青", "宁", // "新", // "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", // "A", "B", "C", "D", "E", "F", "G", "H", "J", "K", // "L", "M", "N", "P", "Q", "R", "S", "T", "U", "V", // "W", "X", "Y", "Z", "I", "O", "-" //}; // crnn使用 string plate_chars[78] = { "#","京", "沪", "津", "渝", "冀", "晋", "蒙", "辽", "吉", "黑", "苏", "浙", "皖", "闽", "赣", "鲁", "豫", "鄂", "湘", "粤", "桂", "琼", "川", "贵", "云", "藏", "陕", "甘", "青", "宁", "新", "学", "警", "港", "澳", "挂", "使", "领", "民", "航", "危", "0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "A", "B", "C", "D", "E", "F", "G", "H", "J", "K", "L", "M", "N", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z", "险", "品"}; #define ASSERT(status, ret) if (!(status)) { return ret; } #define ASSERT_FALSE(status) ASSERT(status, false) static vector<string> crnn_rec(const cv::Mat& bgr){ cv::Mat img = bgr; //获取图片的宽 int w = img.cols; //获取图片的高 int h = img.rows; ncnn::Mat in = ncnn::Mat::from_pixels_resize(img.data, ncnn::Mat::PIXEL_BGR, w, h, 168, 48); float mean[3] = { 149.94, 149.94, 149.94 }; float norm[3] = { 0.020319,0.020319,0.020319 }; //对图片进行归一化,将像素归一化到-1~1之间 in.substract_mean_normalize(mean, norm); ncnn::Extractor ex = crnn.create_extractor(); ex.set_light_mode(true); //设置线程个数 ex.set_num_threads(1); //将图片放入到网络中,进行前向推理 ex.input("images", in); ncnn::Mat feat_plate; ncnn::Mat feat_color; // color = ['黑色', '蓝色', '绿色', '白色', '黄色'] //获取网络的输出结果 ex.extract("129", feat_plate); ex.extract("output_2", feat_color); // int color_index = max_element(feat_color + 0, feat_color + 78) - feat_color; cout<<feat_color.c<<endl; cout<<feat_color.h<<endl; cout<<feat_color.w<<endl; // pretty_print(feat_color); ncnn::Mat plate_mat = feat_plate; ncnn::Mat color_mat = feat_color; vector<string> final_plate_str{}; string finale_plate; for (int q = 0; q < plate_mat.c; q++) { float prebs[21]; for (int x = 0; x < plate_mat.w; x++) //遍历十八个车牌位置 { const float* ptr = plate_mat.channel(q); float preb[78]; for (int y = 0; y < plate_mat.h; y++) //遍历68个字符串位置 { preb[y] = ptr[x]; //将18个 ptr += plate_mat.w; } int max_num_index = max_element(preb + 0, preb + 78) - preb; // cout<<"max_num_index"<<max_num_index<<endl; prebs[x] = max_num_index; } //去重复、去空白a vector<int> no_repeat_blank_label{}; int pre_c = prebs[0]; cout<<"pre_c"<<pre_c<<endl; if (pre_c != 0) { no_repeat_blank_label.push_back(pre_c); } for (int value : prebs) { if (value == 0 or value==pre_c) { if (value == 0 or value == pre_c) { pre_c = value; } continue; } no_repeat_blank_label.push_back(value); pre_c = value; } // 下面进行车牌lable按照字典进行转化为字符串 string no_repeat_blank_c = ""; for (int hh : no_repeat_blank_label) { no_repeat_blank_c += plate_chars[hh]; } cout << "单个车牌:" << no_repeat_blank_c << endl; final_plate_str.push_back(no_repeat_blank_c); for (string plate_char : final_plate_str) { cout << "所有车牌:" << plate_char << endl; finale_plate += plate_char; } } string str = finale_plate; cout << str << endl; float color_result[5]; for (int q = 0; q < color_mat.c; q++) { const float* ptr = color_mat.channel(q); for (int y = 0; y < color_mat.h; y++) { for (int x = 0; x < color_mat.w; x++) { // printf("%f ", ptr[x]); //cout << "1111:" << ptr[x]; color_result[x] = ptr[x]; } ptr += color_mat.w; // printf("\n"); } printf("------------------------\n"); } int color_code = max_element(color_result, color_result + 5) - color_result; string color_names[5] = { "黑色", "蓝色", "绿色", "白色", "黄色" }; vector<string> plate_color(2); plate_color[0] = str; plate_color[1] = color_names[color_code]; return plate_color; } //static int color_rec_1(const cv::Mat& bgr){ //// ncnn::Net color_net; //// //// color_net.load_param("color-sim.param"); //// color_net.load_model("color-sim.bin"); // //获取图片的宽 // int w = bgr.cols; // //获取图片的高 // int h = bgr.rows; // // ncnn::Mat in = ncnn::Mat::from_pixels_resize(bgr.data, ncnn::Mat::PIXEL_BGR2RGB, w, h, 34, 9); // // const float norm_vals[3] = { 1.f / (0.2569 * 255), 1.f / (0.2478 * 255), 1.f / (0.2174 * 255)}; // // const float mean_vals[3] = { 0.4243f * 255.f, 0.4947f * 255.f, 0.434f * 255.f}; // // in.substract_mean_normalize(mean_vals, norm_vals); // // ncnn::Extractor ex = color_net.create_extractor(); // //将图片放入到网络中,进行前向推理 // ex.input("input.1", in); // ncnn::Mat feat_color; // ex.extract("14", feat_color); // ncnn::Mat m = feat_color; // float color_result[3]; // for (int q = 0; q < m.c; q++) // { // const float* ptr = m.channel(q); // for (int y = 0; y < m.h; y++) // { // // for (int x = 0; x < m.w; x++) // { // printf("%f ", ptr[x]); // //cout << "1111:" << ptr[x]; // color_result[x] = ptr[x]; // } // ptr += m.w; // prin