_pip_单缺口_多缺口滑块验证码识别，pip安装使用.zip

简体中文 | [English](https://github.com/chenwei-zhao/captcha-recognizer/blob/main/README_en.md) # Captcha-Recognizer Captcha-Recognizer是一个易用的通用滑块验证码识别库，通过深度学习训练通用的缺口检测模型，基于训练的结果，识别出验证码中的滑块缺口位置，并返回缺口的坐标与可信度。 # 支持的验证码类型 - 单缺口验证码背景图 - 多缺口验证码背景图 - 验证码截图（包含滑块和背景图） # 在线演示 - [在线演示](http://47.94.198.97/) <p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/online-demo.gif" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/online-demo.gif" > </p> # 版本要求 * ``Python`` >= 3.6.0 * ``opencv-python`` * ``shapely`` * ``onnxruntime`` * Works on Linux, Windows, MacOS # 使用方式 - Pypi - [HTTP API](https://github.com/chenwei-zhao/captcha-api) ## Pypi ### 从 Pypi 安装 ```bash pip install captcha-recognizer ``` ## HTTP API 请移步: [captcha-api](https://github.com/chenwei-zhao/captcha-api) # 使用示例 ## V2 增强版 V2增强版，增强了对多缺口复杂验证码的识别效果 V2支持以下类型验证码的识别 - 单缺口验证码背景图（不含滑块的背景图） - 多缺口验证码截图或合成图（含滑块和背景图） ### V2 Beta 增强版使用示例 ```shell pip install captcha-recognizer --upgrade ``` ```python3 # V2增强版 from captcha_recognizer.slider import SliderV2 box, confidence = SliderV2().identify(source=f'images_example/example8.png', show=True) print(f'缺口坐标: {box}') print('置信度', confidence) ``` ## V1 版本 ### 基于单缺口/多缺口验证码背景图识别滑块缺口 ```Python from captcha_recognizer.recognizer import Recognizer # source传入图片， 支持 Union[str, Path, bytes, np.ndarray] # is_single 默认为False表示支持多缺口背景图识别，is_single=True表示仅支持单缺口背景图识别（指定is_single为True时，对单缺口背景图识别准确度更高） recognizer = Recognizer() box, confidence = recognizer.identify_gap(source='your_example_image.png') print(f'缺口坐标: {box}') print(f'可信度: {confidence}') """ 打印结果如下: 缺口方框坐标: [331.72052001953125, 55.96122741699219, 422.079345703125, 161.7498779296875] 可信度: 0.9513089656829834 坐标原点：图片左上角 缺口方框坐标为缺口方框左上角和右下角距离坐标原点的距离 """ ``` ### 基于单缺口/多缺口验证码背景图识别滑块缺口的示例图片 包括且不限于以下类型、尺寸的滑块图片检测 <p>示例图 1</p> <p>尺寸 552*344</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example1.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example1.png" > <p>识别效果示例图 1</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict1.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict1.png" > <p>示例图 2</p> <p>尺寸 260*160</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example2.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example2.png" > <p>识别效果示例图 2</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict2.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict2.png" > <p>示例图 3</p> <p>尺寸 400*200</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example3.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example1.png" > <p>识别效果示例图3</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict3.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict3.png" > <p>示例图 4</p> <p>尺寸 672*390</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example4.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example4.png" > <p>识别效果示例图4</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict4.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict4.png" > <p>示例图 5</p> <p>尺寸 280*155</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example5.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example5.png" > <p>识别效果示例图 5</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict5.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict5.png" > <p>示例图 6</p> <p>尺寸 590*360</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example6.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example6.png" > <p>识别效果示例图 6</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict6.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict6.png" > <p>示例图 7</p> <p>尺寸 320*160</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example7.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example7.png" > <p>识别效果示例图 7</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict7.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict7.png" > ### 基于验证码截图的识别滑块缺口 ```Python from captcha_recognizer.recognizer import Recognizer # source传入图片，支持 Union[str, Path, bytes, np.ndarray] recognizer = Recognizer() box, confidence = recognizer.identify_screenshot(source='your_example_image.png') print(f'缺口坐标: {box}') print(f'可信度: {confidence}') """ 打印结果如下: 缺口方框坐标: [331.72052001953125, 55.96122741699219, 422.079345703125, 161.7498779296875] 可信度: 0.9513089656829834 坐标原点：图片左上角 缺口方框坐标为缺口方框左上角和右下角距离坐标原点的距离 """ ``` ### 基于验证码截图的滑块识别滑块缺口示例 包括且不限于以下类型、尺寸的滑块验证码截图 <p>示例图 8</p> <p>尺寸 305*156</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/example8.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/example8.png" > <p>识别效果示例图 8</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_predict/predict8.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/predict8.png" > ### 基于验证码截图识别滑块距离 ```python3 from captcha_recognizer.recognizer import Recognizer # source传入图片, 支持Union[str, Path, bytes, np.ndarray] recognizer = Recognizer() distance = recognizer.identify_distance_by_screenshot(source='your_screenshot.jpg') print('滑块距离', distance) ``` # 注意事项 ## 偏移量 某些种类的滑块验证码，滑块初始位置存在一定偏移，以下面图中的滑块初始位置为例： <p>示例图 9</p> <img src="https://raw.githubusercontent.com/chenwei-zhao/captcha-recognizer/main/images_example/offset2.png" alt="https://captcha-slider.oss-cn-beijing.aliyuncs.com/slider/offset2.png" > 如示例图9中： - 第一条黑线位置为滑块初始位置，距离图片边框有大概有8个像素的偏移量（offset为8） - 识别结果的缺口坐标为 [x1, y1, x2, y2] 对应�