perf: batch=1 优化减少延迟

- settings: batch_size=41
- tensorrt_engine: BATCH_SIZE=41
- preprocessor: 移除 padding 逻辑，直接 batch=1
- 预处理延迟从 17ms  5ms

config/settings.py

@@ -75,7 +75,7 @@ class InferenceConfig:
     model_path: str = "./models/yolo11n.engine"
     input_width: int = 480
     input_height: int = 480
-    batch_size: int = 4
+    batch_size: int = 1
     conf_threshold: float = 0.5
     nms_threshold: float = 0.45
     device_id: int = 0