如何从nvarguscamerasrc的3840x2160流获取1280x1280帧且不缩放？

问题核心：你的相机源是3840x2160（16:9宽高比），nvarguscamerasrc默认会按原比例缩放，指定输出1280宽度时，高度只能对应到720，剩余区域填充黑边。要得到无黑边的1280x1280帧，需先将源帧裁剪为正方形，再缩放到目标分辨率。

修改gstreamer pipeline配置 #

在JetsonCamera.py的gstreamer_pipeline函数中，添加裁剪和强制缩放逻辑，具体代码如下：

def gstreamer_pipeline(
    capture_width=3840,
    capture_height=2160,
    display_width=1280,
    display_height=1280,
    framerate=30,
    flip_method=0,
):
    # 从3840x2160源帧裁剪中间2160x2160的正方形区域
    crop_left = (capture_width - capture_height) // 2
    return (
        "nvarguscamerasrc ! "
        f"video/x-raw(memory:NVMM), width=(int){capture_width}, height=(int){capture_height}, framerate=(fraction){framerate}/1 ! "
        f"nvvidconv flip-method={flip_method} ! "
        f"video/x-raw(memory:NVMM), width=(int){capture_height}, height=(int){capture_height}, "
        f"cropleft={crop_left}, croptop=0, cropwidth={capture_height}, cropheight={capture_height} ! "
        f"nvvidconv ! "
        f"video/x-raw, width=(int){display_width}, height=(int){display_height}, format=(string)BGRx ! "
        "videoconvert ! "
        "video/x-raw, format=(string)BGR ! "
        "appsink"
    )

调整Camera类初始化逻辑 #

确保Camera类使用修改后的pipeline，示例代码：

class Camera:
    def __init__(self, width=1280, height=1280, framerate=30, flip_method=0):
        self.width = width
        self.height = height
        self.pipeline = gstreamer_pipeline(
            capture_width=3840,
            capture_height=2160,
            display_width=width,
            display_height=height,
            framerate=framerate,
            flip_method=flip_method
        )
        self.cap = cv2.VideoCapture(self.pipeline, cv2.CAP_GSTREAMER)
        if not self.cap.isOpened():
            raise RuntimeError("Failed to open camera")
        self.frame_reader = FrameReader(self.cap)
        self.frame_reader.start()

FrameReader线程类无需额外修改 #

只要Camera类的pipeline正确输出1280x1280的BGR帧，线程读取逻辑保持原样即可：

class FrameReader(threading.Thread):
    def __init__(self, cap):
        super().__init__()
        self.cap = cap
        self.frame = None
        self.running = True

    def run(self):
        while self.running:
            ret, frame = self.cap.read()
            if ret:
                self.frame = frame

    def stop(self):
        self.running = False
        self.join()

main.py调用逻辑保持匹配 #

确保请求的分辨率与Camera初始化参数一致：

from JetsonCamera import Camera

if __name__ == "__main__":
    try:
        camera = Camera(width=1280, height=1280)
        while True:
            frame = camera.frame_reader.frame
            if frame is not None:
                # 处理1280x1280帧
                cv2.imshow("Camera Output", frame)
                if cv2.waitKey(1) & 0xFF == ord('q'):
                    break
    finally:
        camera.frame_reader.stop()
        camera.cap.release()
        cv2.destroyAllWindows()

关键逻辑说明 #

1. 裁剪正方形区域：从3840x2160源帧中裁剪中间2160x2160的正方形（左右各裁840像素），得到1:1比例的原始帧。
2. 强制缩放：将裁剪后的2160x2160帧缩放到1280x1280，因源与目标均为1:1比例，不会出现黑边。
3. 硬件加速：全程使用memory:NVMM格式，利用Jetson硬件加速避免性能损耗。

内容的提问来源于stack exchange，提问作者Jason Rich Darmawan