如何对街景图像的对象进行地理定位？
我已经在Python中开发了一种基于Yolo的算法，以检测Google Maps图像的对象。现在，我已经完成了将该对象分解为图像中的任务。我试图访问图像的EXIF信息并将其调整到对象上，但是它会有很高的位置错误。
是否有任何方法或技术可以将检测到的对象进行地理位置？
I have developed a YoLo based algorithm in Python to detect objects from Google Maps images. Now, i have the task to GEOLOCATE that object in the image. I have tried to access the EXIF information of the image and adapt it to the object, but it gets high position errors.
Is there any way or technique to acces geolocation of an object detected of my image?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
除非 Google 提供 API 将街景图像像素映射到 North-East-Down 坐标中的向量，否则您通常可以忘记这样做。
街景图像是非常复杂的图像处理管道的结果，其结果是立方体全景图。在最终的 Web 客户端视图中，您可以访问视图的位置（纬度、经度和海拔）以及视图的水平方向（相对于北的角度），但仅此而已。为了对场景中的对象进行地理定位，您需要知道它与相机的距离以及您看到它的方向。
Unless Google provides an API to map Street View image pixels to vectors in North-East-Down coordinates, you can forget about doing it in general.
Street View images are the result of a very complex image processing pipeline, whose result is a cubic panorama. In the final web client view you have access to the location of the view (as latitude, longitude and elevation), and the horizontal direction of view (as an angle with respect to North), but that's about it. In order to geolocate an object in the scene you'd need to know its distance from the camera and the direction under which you see it.