ai控制鼠标生成刀路系统环境搭建尝试7 lsd识别刀路线段2

python 复制代码
import cv2
import numpy as np
from tkinter import Tk
from tkinter.filedialog import askopenfilename
import matplotlib.pyplot as plt

# 隐藏Tk窗口
Tk().withdraw()

print("请选择一张刀路图（S形，从下往上走）...")
file_path = askopenfilename(
    title="选择刀路图像",
    filetypes=[("Image files", "*.jpg *.jpeg *.png *.bmp")]
)

if not file_path:
    print("未选择图片，程序退出。")
    exit()

src = cv2.imread(file_path)
if src is None:
    print("无法读取图片。")
    exit()

height, width = src.shape[:2]

# ===================================================================
# === 方法1：宽松 BGR 范围（推荐先用这个）===
# 青色（Cyan）：允许偏蓝、偏绿、轻微泛白或泛红
lower_cyan = np.array([150, 150, 0])    # B≥150, G≥150, R≤80
upper_cyan = np.array([255, 255, 80])

# 黄色（Yellow）：允许偏橙、偏暗、轻微泛白
lower_yellow = np.array([0, 140, 140])  # B≤80, G≥140, R≥140
upper_yellow = np.array([80, 255, 255])

binary_lit = cv2.inRange(src, lower_cyan, upper_cyan)
binary_unlit = cv2.inRange(src, lower_yellow, upper_yellow)

# ===================================================================
# === 方法2：HSV 更鲁棒（如BGR效果仍不佳，取消注释以下块，注释上面）===
# hsv = cv2.cvtColor(src, cv2.COLOR_BGR2HSV)
# 
# # 青色：H=80~120（覆盖蓝绿到青）
# binary_lit = cv2.inRange(hsv, np.array([80, 50, 80]), np.array([120, 255, 255]))
# 
# # 黄色：H=15~45（覆盖橙黄到黄）
# binary_unlit = cv2.inRange(hsv, np.array([15, 50, 80]), np.array([45, 255, 255]))
# ===================================================================

# 可选：轻微膨胀连接断点（提升短线检测）
kernel = np.ones((2, 2), np.uint8)
binary_lit = cv2.morphologyEx(binary_lit, cv2.MORPH_CLOSE, kernel)
binary_unlit = cv2.morphologyEx(binary_unlit, cv2.MORPH_CLOSE, kernel)

# === LSD 检测线段（在灰度图上）===
gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
lsd = cv2.createLineSegmentDetector(refine=cv2.LSD_REFINE_ADV)
all_lines, _, _, _ = lsd.detect(gray)

if all_lines is None or len(all_lines) == 0:
    print("未检测到任何线段。")
    exit()

# === 按行分组 ===
row_height = 25  # 根据你的图调整（当前图约25px/行）
rows = {}
for line in all_lines:
    x1, y1, x2, y2 = line[0]
    mid_y = (y1 + y2) / 2
    row_id = int(mid_y // row_height)
    if row_id not in rows:
        rows[row_id] = []
    rows[row_id].append((x1, y1, x2, y2))

# === 从下往上找"当前行"（部分点亮+部分未点亮）===
sorted_row_ids = sorted(rows.keys())
current_row_info = None

for row_id in sorted_row_ids:
    y_min = row_id * row_height
    y_max = (row_id + 1) * row_height

    has_lit = False
    has_unlit = False
    lit_side = None

    for (x1, y1, x2, y2) in rows[row_id]:
        seg_mask = np.zeros_like(gray, dtype=np.uint8)
        cv2.line(seg_mask, (int(x1), int(y1)), (int(x2), int(y2)), 255, thickness=2)

        total = cv2.countNonZero(seg_mask)
        if total == 0:
            continue

        lit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_lit))
        unlit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_unlit))

        if lit_px / total > 0.3:
            has_lit = True
            lit_side = 'left' if (x1 + x2) / 2 < width / 2 else 'right'
        if unlit_px / total > 0.3:
            has_unlit = True

    if has_lit and has_unlit:
        current_row_info = {
            'row_id': row_id,
            'y_range': (y_min, y_max),
            'lit_side': lit_side
        }
        break

# === 处理起始/完成状态 ===
if current_row_info is None:
    total_lit_area = cv2.countNonZero(binary_lit)
    if total_lit_area / (height * width) > 0.9:
        print("\n✅ 刀路已完成：所有区域均已点亮。")
    elif sorted_row_ids:
        bottom_row_id = sorted_row_ids[0]
        y_min = bottom_row_id * row_height
        y_max = (bottom_row_id + 1) * row_height
        current_row_info = {
            'row_id': bottom_row_id,
            'y_range': (y_min, y_max),
            'lit_side': None
        }
        print("\nℹ️ 刀路尚未开始，将从最下行左侧开始。")

# === 输出分析结果 ===
if current_row_info:
    y_min, y_max = current_row_info['y_range']
    lit_side = current_row_info['lit_side']

    if lit_side is None:
        print("\n✅ 当前刀路状态分析:")
        print(f"- 当前行 Y 范围: {y_min:.1f} ~ {y_max:.1f}")
        print("- 当前行状态: 全未点亮（起始状态）")
        print("- 下一刀方向: 向右切削（从左侧开始）")
    else:
        side_str = "左边" if lit_side == 'left' else "右边"
        print("\n✅ 当前刀路状态分析:")
        print(f"- 当前行 Y 范围: {y_min:.1f} ~ {y_max:.1f}")
        print(f"- 当前行状态: 既有已点亮线段，也有未点亮线段")
        print(f"- 已点亮侧: {side_str}")
        print("- 下一刀方向:", "向右切削" if lit_side == 'left' else "向左切削")

# === 可视化：三图并列 ===
fig, axes = plt.subplots(1, 3, figsize=(21, 6))

# 图1：原始图
axes[0].imshow(cv2.cvtColor(src, cv2.COLOR_BGR2RGB))
axes[0].set_title('原始刀路图', fontsize=12)
axes[0].axis('off')

# 图2：LSD线段 + 状态颜色
status_canvas = np.ones_like(src) * 255  # 白底
for line in all_lines:
    x1, y1, x2, y2 = line[0]
    seg_mask = np.zeros_like(gray, dtype=np.uint8)
    cv2.line(seg_mask, (int(x1), int(y1)), (int(x2), int(y2)), 255, 2)
    
    total = cv2.countNonZero(seg_mask)
    if total == 0:
        color = (180, 180, 180)  # 灰色
    else:
        lit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_lit))
        unlit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_unlit))
        if lit_px / total > 0.3:
            color = (0, 255, 0)      # 绿色：点亮（原青色）
        elif unlit_px / total > 0.3:
            color = (0, 100, 255)    # 橙色：未点亮（原黄色）
        else:
            color = (180, 180, 180)  # 灰色：不确定

    cv2.line(status_canvas, (int(x1), int(y1)), (int(x2), int(y2)), color, 2)

axes[1].imshow(cv2.cvtColor(status_canvas, cv2.COLOR_BGR2RGB))
axes[1].set_title('LSD线段（绿=已点亮，橙=未点亮）', fontsize=12)
axes[1].axis('off')

# 图3：分析结果（当前行高亮）
result_img = src.copy()
if current_row_info:
    y_min, y_max = current_row_info['y_range']
    cv2.rectangle(result_img, (0, int(y_min)), (width, int(y_max)), (255, 0, 0), 3)

    row_id = current_row_info['row_id']
    if row_id in rows:
        for (x1, y1, x2, y2) in rows[row_id]:
            mid_y = (y1 + y2) / 2
            if y_min <= mid_y <= y_max:
                seg_mask = np.zeros_like(gray, dtype=np.uint8)
                cv2.line(seg_mask, (int(x1), int(y1)), (int(x2), int(y2)), 255, 2)
                lit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_lit))
                unlit_px = cv2.countNonZero(cv2.bitwise_and(seg_mask, binary_unlit))
                color = (0, 255, 0) if lit_px > unlit_px else (0, 100, 255)
                cv2.line(result_img, (int(x1), int(y1)), (int(x2), int(y2)), color, 2)

    text = f"当前行: {'左' if current_row_info['lit_side'] == 'left' else '右'}已点亮" \
           if current_row_info['lit_side'] else "起始行"
    cv2.putText(result_img, text, (10, int(y_min) + 20),
                cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2)

axes[2].imshow(cv2.cvtColor(result_img, cv2.COLOR_BGR2RGB))
axes[2].set_title('刀路状态分析（蓝框=当前行）', fontsize=12)
axes[2].axis('off')

plt.tight_layout()
plt.show()

# === 调试信息 ===
print(f"\n📊 调试信息:")
print(f"- 图像尺寸: {width} x {height}")
print(f"- 检测线段数: {len(all_lines)}")
print(f"- 分组行数: {len(rows)}")
print(f"- 行高度: {row_height} 像素")
ai控制鼠标生成刀路系统 环境搭建尝试7 lsd识别刀路线段2

ai控制鼠标生成刀路系统环境搭建尝试7 lsd识别刀路线段2
