Support both esp_video and esp32_camera (#1671)

* Update project version to 2.2.1 and refactor camera component handling

- Incremented project version from 2.2.0 to 2.2.1 in CMakeLists.txt.
- Removed legacy esp32_camera component and replaced it with esp_video for ESP32-S3 and ESP32-P4 boards.
- Updated board implementations to utilize the new esp_video component, ensuring compatibility and improved functionality.
- Cleaned up Kconfig options related to camera selection, streamlining the configuration process.
- Enhanced camera initialization logic across various board files to support the new component structure.

* Refactor camera handling in AtomS3R CAM/M12 EchoBase board

- Replaced the legacy EspVideo component with the new Esp32Camera class for improved camera functionality.
- Updated camera initialization logic to utilize a more structured configuration approach, enhancing clarity and maintainability.
- Removed outdated comments and code related to the previous camera implementation in the README file.

* Update camera configuration for atoms3r-cam-m12-echo-base

- Removed outdated camera configuration options from config.json to streamline the setup.
- Retained essential partition table configuration for improved clarity.

* Enhance Esp32Camera functionality and memory management

- Added esp_timer.h for improved timing functionality.
- Streamlined camera initialization by removing redundant frame buffer setup and logging.
- Improved memory allocation for JPEG encoding and added error handling for unsupported pixel formats.
- Updated comments for clarity and consistency, ensuring better understanding of the code flow.

main/boards/common/esp_video.h

@@ -0,0 +1,53 @@
+#pragma once
+#include "sdkconfig.h"
+
+#include <lvgl.h>
+#include <thread>
+#include <memory>
+#include <vector>
+
+#include <freertos/FreeRTOS.h>
+#include <freertos/queue.h>
+
+#include "camera.h"
+#include "jpg/image_to_jpeg.h"
+#include "esp_video_init.h"
+
+struct JpegChunk {
+    uint8_t* data;
+    size_t len;
+};
+
+class EspVideo : public Camera {
+private:
+    struct FrameBuffer {
+        uint8_t *data = nullptr;
+        size_t len = 0;
+        uint16_t width = 0;
+        uint16_t height = 0;
+        v4l2_pix_fmt_t format = 0;
+    } frame_;
+    v4l2_pix_fmt_t sensor_format_ = 0;
+#ifdef CONFIG_XIAOZHI_ENABLE_ROTATE_CAMERA_IMAGE
+    uint16_t sensor_width_ = 0;
+    uint16_t sensor_height_ = 0;
+#endif  // CONFIG_XIAOZHI_ENABLE_ROTATE_CAMERA_IMAGE
+    int video_fd_ = -1;
+    bool streaming_on_ = false;
+    struct MmapBuffer { void *start = nullptr; size_t length = 0; };
+    std::vector<MmapBuffer> mmap_buffers_;
+    std::string explain_url_;
+    std::string explain_token_;
+    std::thread encoder_thread_;
+
+public:
+    EspVideo(const esp_video_init_config_t& config);
+    ~EspVideo();
+
+    virtual void SetExplainUrl(const std::string& url, const std::string& token);
+    virtual bool Capture();
+    // 翻转控制函数
+    virtual bool SetHMirror(bool enabled) override;
+    virtual bool SetVFlip(bool enabled) override;
+    virtual std::string Explain(const std::string& question);
+};