Long Range Crop instead of downscale

jakaskerl

Thanks for the help, sorry for the slow response

I did what you suggested but now I'm getting the following error:
Left input image stride ('3840') should be equal to its width ('1280'). Skipping frame!

Code:

#!/usr/bin/env python3
from collections import deque
import cv2
import numpy as np
import depthai as dai
import time

fps = 30

# Create pipeline
pipeline = dai.Pipeline()
device = dai.Device()
queueNames = []

# Define sources and outputs
camRgb = pipeline.create(dai.node.Camera)
left = pipeline.create(dai.node.ColorCamera)
right = pipeline.create(dai.node.ColorCamera)
stereo = pipeline.create(dai.node.StereoDepth)

rgbOut = pipeline.create(dai.node.XLinkOut)
disparityOut = pipeline.create(dai.node.XLinkOut)

rgbOut.setStreamName("rgb")
queueNames.append("rgb")
disparityOut.setStreamName("depth")
queueNames.append("depth")

#Properties
rgbCamSocket = dai.CameraBoardSocket.CAM_A
camRgb.setBoardSocket(rgbCamSocket)
camRgb.setSize(1280, 800)
camRgb.setFps(fps)
camRgb.setPreviewSize(1280, 800)

# For now, RGB needs fixed focus to properly align with depth.
# This value was used during calibration
try:
    calibData = device.readCalibration2()
    lensPosition = calibData.getLensPosition(rgbCamSocket)
    if lensPosition:
        camRgb.initialControl.setManualFocus(lensPosition)
except:
    raise
left.setResolution(dai.ColorCameraProperties.SensorResolution.THE_1200_P)
left.setCamera("left")
left.setFps(fps)
left.setPreviewSize(1280, 800)

right.setResolution(dai.ColorCameraProperties.SensorResolution.THE_1200_P)
right.setCamera("right")
right.setFps(fps)
right.setPreviewSize(1280, 800)

stereo.setDefaultProfilePreset(dai.node.StereoDepth.PresetMode.HIGH_DENSITY)
stereo.setOutputSize(1280, 800)
stereo.setLeftRightCheck(True)
stereo.setSubpixel(True)
stereo.setSubpixelFractionalBits(5)

stereo.setDepthAlign(rgbCamSocket)

camRgb.preview.link(rgbOut.input)
left.preview.link(stereo.left)
right.preview.link(stereo.right)
stereo.disparity.link(disparityOut.input)

maxDisp = stereo.initialConfig.getMaxDisparity()

prev_frame_time = 0
new_frame_time = 0
average_fps = deque(maxlen=30)
with device:
    device.startPipeline(pipeline)

    while not device.isClosed():
        queueNames = device.getQueueEvents()
        for q in queueNames:
            message = device.getOutputQueue(q).get()
            # Display arrived frames
            if type(message) == dai.ImgFrame:
                frame = message.getCvFrame()
                if 'depth' in q:
                    disp = (frame * (255.0 / maxDisp)).astype(np.uint8)
                    disp = cv2.applyColorMap(disp, cv2.COLORMAP_JET)
                    cv2.imshow(q, disp)
                else:
                    new_frame_time = time.time()
                    average_fps.append(1/(new_frame_time-prev_frame_time))
                    prev_frame_time = new_frame_time
                    cv2.putText(frame, str(np.mean(average_fps)), (7, 70), cv2.FONT_HERSHEY_SIMPLEX , 3, (100, 255, 0), 3, cv2.LINE_AA)
                    cv2.imshow(q, frame)
        if cv2.waitKey(1) == ord('q'):
            break

jakaskerl Thanks, but that does not crop the image. If I just link to isp instead of preview then that is not cropped so the stereo module gives: Maximum supported input image width for stereo is 1280. Skipping frame! and if I use setIspScale then that downscales to the correct resolution but I want a cropped image.

To reiterate I want to crop the stereo pair images to 1280x800 so that it works with the stereo node and I want to crop instead of downscale because I want to reduce the fov to get better long range accuracy.

jakaskerl Thanks, that solves the problem!