keypoints

# Object tracking with keypoints example.
# Show the camera an object and then run the script. A set of keypoints will be extracted
# once and then tracked in the following frames. If you want a new set of keypoints re-run
# the script. NOTE: see the docs for arguments to tune find_keypoints and match_keypoints.

from media.camera import *
from media.display import *
from media.media import *
import time, os, gc, sys

DISPLAY_WIDTH = ALIGN_UP(1920, 16)
DISPLAY_HEIGHT = 1080
SCALE = 4
DETECT_WIDTH = DISPLAY_WIDTH // SCALE
DETECT_HEIGHT = DISPLAY_HEIGHT // SCALE

def camera_init():
    # use hdmi for display
    display.init(LT9611_1920X1080_30FPS)
    # config vb for osd layer
    config = k_vb_config()
    config.max_pool_cnt = 1
    config.comm_pool[0].blk_size = 4*DISPLAY_WIDTH*DISPLAY_HEIGHT
    config.comm_pool[0].blk_cnt = 1
    config.comm_pool[0].mode = VB_REMAP_MODE_NOCACHE
    # meida buffer config
    media.buffer_config(config)
    # init default sensor
    camera.sensor_init(CAM_DEV_ID_0, CAM_DEFAULT_SENSOR)
    # set chn0 output size
    camera.set_outsize(CAM_DEV_ID_0, CAM_CHN_ID_0, DISPLAY_WIDTH, DISPLAY_HEIGHT)
    # set chn0 output format
    camera.set_outfmt(CAM_DEV_ID_0, CAM_CHN_ID_0, PIXEL_FORMAT_YUV_SEMIPLANAR_420)
    # create meida source device
    globals()["meida_source"] = media_device(CAMERA_MOD_ID, CAM_DEV_ID_0, CAM_CHN_ID_0)
    # create meida sink device
    globals()["meida_sink"] = media_device(DISPLAY_MOD_ID, DISPLAY_DEV_ID, DISPLAY_CHN_VIDEO1)
    # create meida link
    media.create_link(meida_source, meida_sink)
    # set display plane with video channel
    display.set_plane(0, 0, DISPLAY_WIDTH, DISPLAY_HEIGHT, PIXEL_FORMAT_YVU_PLANAR_420, DISPLAY_MIRROR_NONE, DISPLAY_CHN_VIDEO1)
    # set chn1 output nv12
    camera.set_outsize(CAM_DEV_ID_0, CAM_CHN_ID_1, DETECT_WIDTH, DETECT_HEIGHT)
    camera.set_outfmt(CAM_DEV_ID_0, CAM_CHN_ID_1, PIXEL_FORMAT_RGB_888)
    # media buffer init
    media.buffer_init()
    # request media buffer for osd image
    globals()["buffer"] = media.request_buffer(4 * DISPLAY_WIDTH * DISPLAY_HEIGHT)
    # start stream for camera device0
    camera.start_stream(CAM_DEV_ID_0)

def camera_deinit():
    # stop stream for camera device0
    camera.stop_stream(CAM_DEV_ID_0)
    # deinit display
    display.deinit()
    os.exitpoint(os.EXITPOINT_ENABLE_SLEEP)
    time.sleep_ms(100)
    # release media buffer
    media.release_buffer(globals()["buffer"])
    # destroy media link
    media.destroy_link(globals()["meida_source"], globals()["meida_sink"])
    # deinit media buffer
    media.buffer_deinit()

def camera_drop(frame):
    for i in range(frame):
        os.exitpoint()
        img = camera.capture_image(CAM_DEV_ID_0, CAM_CHN_ID_1)
        camera.release_image(CAM_DEV_ID_0, CAM_CHN_ID_1, img)

def capture_picture():
    # create image for drawing
    draw_img = image.Image(DISPLAY_WIDTH, DISPLAY_HEIGHT, image.ARGB8888)
    # create image for osd
    buffer = globals()["buffer"]
    osd_img = image.Image(DISPLAY_WIDTH, DISPLAY_HEIGHT, image.ARGB8888, alloc=image.ALLOC_VB, phyaddr=buffer.phys_addr, virtaddr=buffer.virt_addr, poolid=buffer.pool_id)
    osd_img.clear()
    display.show_image(osd_img, 0, 0, DISPLAY_CHN_OSD0)
    kpts1 = None
    # NOTE: uncomment to load a keypoints descriptor from file
    #kpts1 = image.load_descriptor("/desc.orb")
    camera_drop(60)
    fps = time.clock()
    while True:
        fps.tick()
        try:
            os.exitpoint()
            rgb888_img = camera.capture_image(CAM_DEV_ID_0, CAM_CHN_ID_1)
            img = rgb888_img.to_grayscale()
            camera.release_image(CAM_DEV_ID_0, CAM_CHN_ID_1, rgb888_img)
            draw_img.clear()
            if kpts1 == None:
                # NOTE: By default find_keypoints returns multi-scale keypoints extracted from an image pyramid.
                kpts1 = img.find_keypoints(max_keypoints=150, threshold=10, scale_factor=1.2)
                if kpts1:
                    if SCALE == 1:
                        draw_img.draw_keypoints(kpts1)
                        draw_img.copy_to(osd_img)
                        time.sleep(2)
                    fps.reset()
            else:
                # NOTE: When extracting keypoints to match the first descriptor, we use normalized=True to extract
                # keypoints from the first scale only, which will match one of the scales in the first descriptor.
                kpts2 = img.find_keypoints(max_keypoints=150, threshold=10, normalized=True)
                if kpts2:
                    match = image.match_descriptor(kpts1, kpts2, threshold=85)
                    if (match.count()>10):
                        # If we have at least n "good matches"
                        # Draw bounding rectangle and cross.
                        draw_img.draw_rectangle([v*SCALE for v in match.rect()])
                        draw_img.draw_cross(match.cx()*SCALE, match.cy()*SCALE, size=10)

                    print(kpts2, "matched:%d dt:%d"%(match.count(), match.theta()))
                    # NOTE: uncomment if you want to draw the keypoints
                    #img.draw_keypoints(kpts2, size=KEYPOINTS_SIZE, matched=True)
            draw_img.copy_to(osd_img)
            del img
            gc.collect()
            print(fps.fps())
        except KeyboardInterrupt as e:
            print("user stop: ", e)
            break
        except BaseException as e:
            sys.print_exception(e)
            break

def main():
    os.exitpoint(os.EXITPOINT_ENABLE)
    camera_is_init = False
    try:
        print("camera init")
        camera_init()
        camera_is_init = True
        print("camera capture")
        capture_picture()
    except Exception as e:
        sys.print_exception(e)
    finally:
        if camera_is_init:
            print("camera deinit")
            camera_deinit()

if __name__ == "__main__":
    main()

具体接口使用请参考相关文档说明：

• camere

• display

• media

• image