From df234f1d13ed01af3ca847920a6af842caa15298 Mon Sep 17 00:00:00 2001
From: Eyuep Sueyruege <eyuep.sueyruege@noreply.ais-lab.eu>
Date: Sat, 24 Feb 2024 12:24:08 +0100
Subject: [PATCH] Add docs/research_ImageRecognition.md

---
 docs/research_ImageRecognition.md | 76 +++++++++++++++++++++++++++++++
 1 file changed, 76 insertions(+)
 create mode 100644 docs/research_ImageRecognition.md

diff --git a/docs/research_ImageRecognition.md b/docs/research_ImageRecognition.md
new file mode 100644
index 0000000..1591527
--- /dev/null
+++ b/docs/research_ImageRecognition.md
@@ -0,0 +1,76 @@
+### Coordination between Arduino and Image Recognition
+
+#### Feasibility and Frameworks
+- Integration between Arduino and image recognition is feasible.
+- Various small cameras like OV7670 are compatible with Arduino.
+- Available image processing libraries:
+  - OpenCV: Resource-intensive but widely used.
+  - Alternatives: Cameramatcs or Pixy2, tailored for microcontrollers.
+- Object recognition:
+  - Pre-trained models or frameworks like TensorFlow Lite for Microcontrollers or EdgeImpulse suitable for microcontrollers.
+  
+#### Processing and Output
+- Data processing on Arduino:
+  - Camera detects cards, sends raw data to Arduino.
+  - Arduino analyzes data and interprets recognized cards.
+- Decision logic:
+  - Implementation of decision logic on Arduino defining actions.
+  - Possible speech output based on card recognition.
+
+#### Energy Consumption
+- Low power consumption or consistent power supply is desirable.
+
+*Note*: Image processing on Arduino can be challenging due to limited resources and performance capabilities. Finding compromises may be necessary.
+
+#### Frameworks
+- TensorFlow Lite for Microcontrollers:
+  - Executes machine learning models on resource-constrained devices.
+- Edge Impulse:
+  - Simplifies application development for microcontrollers, facilitates data collection, model training, and deployment.
+- OpenMV:
+  - Platform for computer vision on microcontrollers, supports basic image processing tasks.
+- Arduino Libraries:
+  - Example: Arduino Computer Vision Library.
+
+#### Considerations
+- Processing 10-15 playing cards simultaneously on Arduino can be challenging, considering resource limitations.
+- Possibilities:
+  1. Algorithm optimization.
+  2. External processing unit.
+  3. Prototype and test.
+  
+#### Example Code Collection
+##### Image Recognition with Color Detection (Arduino)
+```cpp
+#include <Arduino.h>
+#include <Wire.h>
+#include <Adafruit_Sensor.h>
+#include <Adafruit_TCS34725.h>
+
+#define RED_THRESHOLD 1000 // Threshold for red color detection
+
+Adafruit_TCS34725 tcs = Adafruit_TCS34725(TCS34725_INTEGRATIONTIME_50MS, TCS34725_GAIN_4X);
+
+void setup() {
+  Serial.begin(9600);
+  if (tcs.begin()) {
+    Serial.println("Color sensor found!");
+  } else {
+    Serial.println("Color sensor not found. Check connection.");
+    while (1);
+  }
+}
+
+void loop() {
+  uint16_t clear, red, green, blue;
+  tcs.getRawData(&red, &green, &blue, &clear);
+  
+  if (red > RED_THRESHOLD) {
+    Serial.println("Red color detected!");
+    // Further actions for the detected card can be implemented here
+  } else {
+    Serial.println("No red color detected.");
+  }
+  
+  delay(1000); // Delay between detection loops
+}