From 8b64b730a9514542811546173fbf13678fd63104 Mon Sep 17 00:00:00 2001 From: bluepulsar Date: Sun, 1 Feb 2026 10:46:21 +0800 Subject: [PATCH] add script and data for debug speed --- benchmark.py | 113 +++++++++++++++++++++++++++++++++++++++++++ src/orisys/sensor.py | 24 +++++++-- 2 files changed, 134 insertions(+), 3 deletions(-) create mode 100644 benchmark.py diff --git a/benchmark.py b/benchmark.py new file mode 100644 index 0000000..9396414 --- /dev/null +++ b/benchmark.py @@ -0,0 +1,113 @@ + +import os +os.environ["OPENCV_VIDEOIO_MSMF_ENABLE_HW_TRANSFORMS"] = "0" +import cv2 +import orisys +import numpy as np +import time + +def main(): + # 步骤1：创建传感器对象 (必要步骤)， + # config_name为配置文件名（内置配置文件名）或路径（自定义配置文件路径） + # cal_path为校准文件路径, 每一个设备需要创建/使用一个校准文件 + + # 方式1: 从视频文件读取 + sensor = orisys.Sensor(0, config_name="./config/prototype_of3.json", cal_path= "./config/dd01a_0.npy", isstitch=True, cuda=True, verbose=True) + + # 方式2: 从摄像头读取（取消注释使用） + # sensor = orisys.Sensor(0, config_name="./config/prototype_of3.json", cal_path= "./config/cal.npy", isstitch=True, cuda=True, verbose=True) + + print("\n按 'q' 退出程序，按 'r' 重置光流追踪器") + + + while True: + # 步骤2：信息提取 + # 步骤 2.1: 获取并拼接图像 (必要步骤) + start_time = time.perf_counter() + sensor.get_img() + end_time = time.perf_counter() + get_img_time = end_time - start_time + # if sensor.detect_motion(threshold=2): + # print("motion detected") + # 步骤 2.2: 计算形变场 (必要步骤) + start_time = time.perf_counter() + sensor.compute_deformation() + end_time = time.perf_counter() + compute_deformation_time = end_time - start_time + # 步骤 2.3: 计算触碰点 + start_time = time.perf_counter() + sensor.compute_contact() + end_time = time.perf_counter() + compute_contact_time = end_time - start_time + + + #步骤 2.4: 读取信息 + start_time = time.perf_counter() + fps, flow, vnormal, img, contour, centroid, depth_map = sensor.read_info( + sensor.info.FPS, + sensor.info.VRAW, + sensor.info.VNORMAL, + sensor.info.IMG, + sensor.info.CONTOUR, + sensor.info.CENTROID, + sensor.info.DEPTH + ) + end_time = time.perf_counter() + read_info_time = end_time - start_time + #步骤 3: 可视化显示 + # 步骤 3.1: 绘制形变场 + start_time = time.perf_counter() + # arrows = orisys.util.draw_arrows( + # img, + # flow, + # threshold=4, + # grid_spacing=40, + # arrow_scale=1.0 + # ) + # cv2.imshow("Deformation", arrows) + + # # 步骤 3.2: 在图像上触碰点和区域 + # image_with_foe = orisys.util.draw_contact(img, contour, centroid) + # cv2.imshow("Contact Centroid", image_with_foe) + + # # 步骤 3.3: 绘制深度图 + # # 归一化散度值到0-255范围（使用绝对值，然后归一化） + # div_abs = depth_map + # if div_abs.max() > div_abs.min(): + # div_normalized = ((div_abs - div_abs.min()) / (div_abs.max() - div_abs.min()) * 255).astype(np.uint8) + # else: + # div_normalized = np.zeros_like(div_abs, dtype=np.uint8) + + # # 应用颜色映射（使用JET：蓝色=低散度，红色=高散度） + # cv2.imshow("Depth Map", cv2.applyColorMap(div_normalized, cv2.COLORMAP_JET)) + + #步骤 3.4: 打印资料 + # print(f"fps: {fps}") + end_time = time.perf_counter() + display_time = end_time - start_time + + # Get detailed timing from sensor (capture vs processing) + capture_time = sensor._last_capture_time + processing_time = sensor._last_processing_time + + total_time = get_img_time + compute_deformation_time + compute_contact_time + read_info_time + display_time + print(f"get_img: {get_img_time*1000:.2f}ms (capture: {capture_time*1000:.2f}ms, processing: {processing_time*1000:.2f}ms), deformation: {compute_deformation_time*1000:.2f}ms, contact: {compute_contact_time*1000:.2f}ms, read: {read_info_time*1000:.2f}ms, display: {display_time*1000:.2f}ms, total: {total_time*1000:.2f}ms") + + #步骤 3.5: 键盘控制 + key = cv2.waitKey(1) & 0xFF + if key == ord("q"): + print("\n退出程序...") + break + if key == ord("r"): + # 重置光流追踪器（当追踪漂移时使用） + sensor.reset() + print("传感器已重置") + + # ========================================================================= + # 4. 清理资源 + # ========================================================================= + sensor.disconnect() # 释放视频源和相关资源 + + +if __name__ == '__main__': + main() diff --git a/src/orisys/sensor.py b/src/orisys/sensor.py index 1076cb3..8fd0857 100644 --- a/src/orisys/sensor.py +++ b/src/orisys/sensor.py @@ -121,6 +121,8 @@ class Sensor: # Timing self.last_log_time = time.perf_counter() self._fps = 0 + self._last_capture_time = 0.0 + self._last_processing_time = 0.0 # Async pipeline state # self._async_enabled = False @@ -161,10 +163,20 @@ class Sensor: Returns: np.ndarray: Processed grayscale image """ + # Simple direct read - buffer size of 1 should minimize frame accumulation + # Timing measurement helps identify if capture or processing is the bottleneck + t_capture_start = time.perf_counter() ret, frame = self.cap.read() + t_capture_end = time.perf_counter() + self._last_capture_time = t_capture_end - t_capture_start + if not ret: print("Failed to get_img from video source.") return None + + # Start processing timing + t_processing_start = time.perf_counter() + self.frame = frame.copy() # Extract and stitch tiles @@ -181,7 +193,9 @@ class Sensor: img = cv2.flip(img, 1) self.img = img.astype(np.uint8) - + + t_processing_end = time.perf_counter() + self._last_processing_time = t_processing_end - t_processing_start # FPS calculation self._fps = 1.0 / (time.perf_counter() - self.last_log_time) @@ -414,9 +428,13 @@ class Sensor: self.cap = cv2.VideoCapture(self.vid_src) # Set buffer size to 1 to reduce latency and avoid frame accumulation - self.cap.set(cv2.CAP_PROP_BUFFERSIZE, 1) + # Note: Not all backends support this, but it helps when available + try: + self.cap.set(cv2.CAP_PROP_BUFFERSIZE, 1) + except: + pass # Some backends don't support this property self.cap.set(cv2.CAP_PROP_FOURCC, cv2.VideoWriter.fourcc('M', 'J', 'P', 'G')) - self.cap.set(cv2.CAP_PROP_FPS, self._config_fps) + # self.cap.set(cv2.CAP_PROP_FPS, self._config_fps) self.cap.set(cv2.CAP_PROP_FRAME_WIDTH, self._config_resolution_raw[0]) self.cap.set(cv2.CAP_PROP_FRAME_HEIGHT, self._config_resolution_raw[1])