Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/AI-FanGe/OpenAIglasses_for_Navigation
/ functions
Functions
403 in github.com/AI-FanGe/OpenAIglasses_for_Navigation
⨍
Functions
403
◇
Types & classes
31
↳
Endpoints
3
↓ 1 callers
Method
_detect_sharp_corner
检测急转弯
workflow_blindpath.py:1829
↓ 1 callers
Method
_detect_traffic_light_by_color
基于 HSV 颜色空间检测红绿灯
workflow_blindpath.py:300
↓ 1 callers
Method
_draw_command_button
绘制底部中央的指令按钮(类似yolomedia风格)
workflow_crossstreet.py:920
↓ 1 callers
Method
_draw_command_button
绘制底部中央的指令按钮(与斑马线模式统一)
workflow_blindpath.py:2556
↓ 1 callers
Method
_draw_dashed_line_vertical_angle
同样以竖直为0°,画 through center 的虚线。
workflow_crossstreet.py:490
↓ 1 callers
Method
_draw_data_panel_no_bg
绘制数据面板(无黑底,描边文字),与 blindpath 一致
workflow_crossstreet.py:1007
↓ 1 callers
Method
_draw_data_panel_no_bg
绘制数据面板(无黑底版本)
workflow_blindpath.py:2659
↓ 1 callers
Method
_draw_line_vertical_angle
以“竖直方向”为0°基准,angle_deg>0 表示左偏,<0 表示右偏。 在 center 处画一条通过点的直线。
workflow_crossstreet.py:473
↓ 1 callers
Method
_draw_visualizations
增强的可视化绘制方法(与 blindpath 一致)
workflow_crossstreet.py:1049
↓ 1 callers
Method
_draw_visualizations
增强的可视化绘制方法
workflow_blindpath.py:2715
↓ 1 callers
Method
_estimate_angle_by_stripes
基于掩码内条纹(霍夫线)估计角度和可视化(放宽参数 + 鲁棒聚类): 返回 dict: { 'angle_deg': float, # 相对竖直方向偏角([-45,45]),正=左偏,负=右偏 'lines': L
workflow_crossstreet.py:570
↓ 1 callers
Method
_execute_state_machine
执行状态机逻辑
workflow_blindpath.py:1367
↓ 1 callers
Function
_extract_sentence
(event_obj: Any)
asr_core.py:36
↓ 1 callers
Method
_find_sharp_turn
查找急转弯点
workflow_blindpath.py:1803
↓ 1 callers
Method
_generate_guidance
生成引导语音
crosswalk_awareness.py:129
↓ 1 callers
Method
_generate_navigation_guidance
生成导航指引
workflow_blindpath.py:2112
↓ 1 callers
Method
_get_edge_mask
获取掩码的内边缘区域,用于特征点检测
workflow_crossstreet.py:740
↓ 1 callers
Method
_get_edge_mask
获取掩码的内边缘区域,用于特征点检测
workflow_blindpath.py:1062
↓ 1 callers
Method
_get_vanishing_point_features
提取灭点特征
workflow_blindpath.py:1685
↓ 1 callers
Method
_get_voice_priority
获取语音指令的优先级 优先级:障碍物(100) > 转向/平移(50) > 保持直行(10)
workflow_blindpath.py:375
↓ 1 callers
Method
_get_width_at_y
获取特定y坐标的路径宽度
workflow_blindpath.py:1923
↓ 1 callers
Method
_handle_avoiding_obstacle
处理避障状态
workflow_blindpath.py:1616
↓ 1 callers
Method
_handle_crosswalk_approaching
处理接近斑马线的情况
workflow_blindpath.py:1290
↓ 1 callers
Method
_handle_locking_on
处理锁定状态
workflow_blindpath.py:1596
↓ 1 callers
Method
_handle_maneuvering_turn
处理转弯状态
workflow_blindpath.py:1514
↓ 1 callers
Method
_handle_navigating
处理常规导航状态
workflow_blindpath.py:1444
↓ 1 callers
Method
_handle_onboarding
处理上盲道状态
workflow_blindpath.py:1387
↓ 1 callers
Method
_handle_pixel_domain_onboarding
处理像素域的上盲道引导
workflow_blindpath.py:2166
↓ 1 callers
Method
_has_hotword
(self, text: str)
asr_core.py:137
↓ 1 callers
Method
_hull_mask_from_points
从一组点的凸包生成二值掩码
workflow_crossstreet.py:324
↓ 1 callers
Function
_init_font
()
trafficlight_detection.py:110
↓ 1 callers
Function
_init_font
()
yolomedia.py:242
↓ 1 callers
Method
_inner_offset_edge
对二值掩码做内收后提边缘,便于在目标内部打光流特征点
workflow_crossstreet.py:310
↓ 1 callers
Method
_is_area_stable
检查面积是否稳定(避免抖动触发)
crosswalk_awareness.py:251
↓ 1 callers
Method
_is_crosswalk_almost_done
判断斑马线是否"快消失"(斑马线在画面底部且面积很小)- 更严格的判定
workflow_crossstreet.py:424
↓ 1 callers
Method
_is_crosswalk_near
判断斑马线是否"很近"(到用户跟前)- 更严格的判定条件
workflow_crossstreet.py:398
↓ 1 callers
Method
_largest_contour
(mask_bin: np.ndarray)
workflow_crossstreet.py:382
↓ 1 callers
Method
_linear_to_ulaw
16位线性PCM转μ-law
audio_compressor.py:49
↓ 1 callers
Function
_looks_like_blind_path
几何+互斥检查,过滤'横条纹/路牙'伪盲道
workflow_crossstreet.py:151
↓ 1 callers
Function
_make_client
()
qwen_extractor.py:20
↓ 1 callers
Method
_mask_center
用图像矩计算掩码质心;失败返回 None
workflow_crossstreet.py:389
↓ 1 callers
Function
_mask_iou
计算两个mask的IoU
workflow_crossstreet.py:141
↓ 1 callers
Function
_merge_voice_map
读取 voice/map.zh-CN.json 并合并到 AUDIO_MAP
audio_player.py:112
↓ 1 callers
Method
_offset_from_centerline
基于“青色法线中央直线”计算左右偏移: - angle_vertical_deg: 以“竖直方向为0°”的角(与 _draw_line_vertical_angle 相同坐标系) - center_pt: 掩码质心 (cx, cy)
workflow_crossstreet.py:515
↓ 1 callers
Method
_postprocess_mask
形态学净化 + 移除小碎片,缓解毛边与噪点
workflow_crossstreet.py:356
↓ 1 callers
Method
_predict_mask_with_flow
使用 Lucas-Kanade 光流预测掩码位置(与 blindpath 一致)
workflow_crossstreet.py:751
↓ 1 callers
Method
_reset_if_needed
重置状态(斑马线消失时)
crosswalk_awareness.py:265
↓ 1 callers
Method
_reset_turn_tracker
重置转弯追踪器
workflow_blindpath.py:1905
↓ 1 callers
Function
_runner
()
app_main.py:613
↓ 1 callers
Method
_seed_points_from_mask
基于掩码的内收边界,播种 LK 光流特征点
workflow_crossstreet.py:336
↓ 1 callers
Function
_shorten
(s: str, limit: int = 200)
asr_core.py:8
↓ 1 callers
Method
_smooth_centerline
平滑中心线数据,减少抖动
workflow_blindpath.py:1080
↓ 1 callers
Method
_speech_for_obstacle
生成障碍物语音提示
workflow_crossstreet.py:1187
↓ 1 callers
Method
_stabilize_obstacle_list
稳定障碍物检测结果,避免重复叠加(与 blindpath 一致)
workflow_crossstreet.py:812
↓ 1 callers
Method
_stabilize_obstacle_list
稳定障碍物检测结果,避免重复叠加
workflow_blindpath.py:3174
↓ 1 callers
Method
_stabilize_seg_mask
使用 Lucas-Kanade 光流的掩码稳定化实现
workflow_blindpath.py:905
↓ 1 callers
Method
_tensor_to_mask
将张量掩码转换为numpy数组
workflow_blindpath.py:838
↓ 1 callers
Function
_to_proto
(hand_lms)
yolomedia.py:179
↓ 1 callers
Method
_try_backend
尝试调用 yoloe_backend 风格的接口。由于各项目实现不同,这里做“宽容地调用”: - 优先尝试 backend.detect(image, target_classes=['traffic light']) - 次选 backend.in
navigation_master.py:79
↓ 1 callers
Function
_ui_next_y_top
返回右上角下一行的y坐标
trafficlight_detection.py:98
↓ 1 callers
Method
_ulaw_to_linear
μ-law转16位线性PCM
audio_compressor.py:86
↓ 1 callers
Method
_update_obstacle_properties
更新障碍物的派生属性
workflow_crossstreet.py:852
↓ 1 callers
Method
_update_obstacle_properties
更新障碍物的派生属性
workflow_blindpath.py:3243
↓ 1 callers
Method
_update_turn_tracker
更新转弯追踪器
workflow_blindpath.py:1892
↓ 1 callers
Function
_wav_header_unknown_size
(sr=16000, ch=1, sw=2)
audio_stream.py:45
↓ 1 callers
Function
addTimestamp
()
static/main.js:161
↓ 1 callers
Method
add_audio
添加音频数据(PCM 16bit) :param pcm_data: PCM格式音频数据 :param text: 语音文本(用于日志)
sync_recorder.py:164
↓ 1 callers
Method
add_frame
添加一帧视频(原始JPEG数据) :param jpeg_data: JPEG格式的图像数据
sync_recorder.py:98
↓ 1 callers
Method
adpcm_to_pcm16
将4位ADPCM转换回16位PCM
audio_compressor.py:211
↓ 1 callers
Function
cancel_current_ai
取消当前大模型语音任务,并等待其退出。
audio_stream.py:18
↓ 1 callers
Function
cleanup_on_exit
程序退出时的清理工作
app_main.py:237
↓ 1 callers
Method
clearCanvas
()
static/vision_renderer.js:95
↓ 1 callers
Function
compute_approach_metrics
计算障碍物的接近度量 :param prev_obstacles: 前一帧障碍物列表 :param curr_obstacles: 当前帧障碍物列表 :param M: 仿射变换矩阵 :param H: 图像高度 :param W: 图像宽度
utils.py:183
↓ 1 callers
Method
connect
()
static/vision_renderer.js:48
↓ 1 callers
Method
connectVisionWS
()
static/vision.js:98
↓ 1 callers
Method
connectWebSocket
()
static/visualizer.js:49
↓ 1 callers
Function
createDataPanel
()
static/main.js:511
↓ 1 callers
Method
createDataPanel
()
static/vision.js:74
↓ 1 callers
Function
createDirectionLabels
()
static/main.js:495
↓ 1 callers
Method
createProgressBars
()
static/vision.js:48
↓ 1 callers
Function
createScale
()
static/main.js:482
↓ 1 callers
Method
createStatusIndicator
()
static/vision.js:38
↓ 1 callers
Method
decompress
解压音频数据
audio_compressor.py:384
↓ 1 callers
Method
detect
利用白名单作为提示词寻找障碍物。 如果提供了 path_mask,则执行与路径相关的空间过滤。 如果 path_mask 为 None,则进行全局检测。
obstacle_detector_client.py:92
↓ 1 callers
Method
detect_traffic_light
检测红绿灯状态 返回: 'red', 'green', 'yellow', 'unknown'
workflow_blindpath.py:273
↓ 1 callers
Method
drawAreaRatio
(ratio, guidance)
static/vision_renderer.js:401
↓ 1 callers
Function
drawBlob
(buf)
static/main.js:231
↓ 1 callers
Method
drawBox
(box, color, lineWidth = 2)
static/vision_renderer.js:326
↓ 1 callers
Method
drawCountdown
(seconds)
static/vision_renderer.js:243
↓ 1 callers
Method
drawFPS
(fps)
static/vision_renderer.js:418
↓ 1 callers
Method
drawFlashMode
(width, height)
static/visualizer.js:218
↓ 1 callers
Method
drawFrame
(frameData)
static/vision.js:204
↓ 1 callers
Method
drawGrid
(width, height)
static/visualizer.js:100
↓ 1 callers
Method
drawHUD
(width, height)
static/visualizer.js:120
↓ 1 callers
Method
drawHand
(handData)
static/vision.js:215
↓ 1 callers
Method
drawHand
(handData, width, height)
static/visualizer.js:321
↓ 1 callers
Method
drawHandSkeleton
(landmarks)
static/vision_renderer.js:186
↓ 1 callers
Method
drawLockAnimation
(x, y, progress)
static/visualizer.js:499
↓ 1 callers
Method
drawMeasureArrow
(p1, p2)
static/vision_renderer.js:359
↓ 1 callers
Method
drawObjects
(objects)
static/vision.js:252
← previous
next →
101–200 of 403, ranked by callers