H7 Plus Opensource Visual Module OV5640 5MP Smart Color Camera Module without Distortion for OpenMV4

H7 Plus Opensource Visual Module OV5640 5MP Smart Color Camera Module without Distortion for OpenMV4

Former Problems:
- Many users have reported that it is particularly prone to burning, manifested by the high temperature of the microcontroller and the 3.3V voltage being pulled down. Through analysis of multiple chips, the reason is that the IO port of the microcontroller is broken down and short circuited to ground.
- The Micro USB interface is prone to dropping, often during car debugging, when the car moves, the USB port drops directly.
- The original designed VIN input range was 3.6V ~ 5.2V, and batteries with higher voltages used on some platforms require secondary voltage reduction.
- Many users have asked why the program has no issues during debugging, but there are problems when running offline. After analyzing from multiple aspects, the reason is that the platform's power supply current is not enough. Although the static input is 5V, the voltage drops during operation due to the consumption of current. The microcontroller appears to be running under voltage, but the program is already disordered.

Improvement:
- All external IO ports have been equipped with separate ESD tubes.
- Replace the Micro USB port with a Type-C interface.
- Expanding the range of Vin to 3.6V-17V with minimal need for secondary voltage reduction.
- PG indicator has been added to the board. As long as the PG light is on, 3.3V is guaranteed to be within a 5% error range, providing a little convenience for troubleshooting.

Features:
- OV5640 sensor module added, compatible with OpenMV4.
- Track the firmware source code, submit the code to the firmware, and also design some easy-to-use tips
- The board upgraded to be with genuine key, which eliminates the trouble of opening IDE reminders
- SDRAM and FLASH have been expanded outside the module. Say goodbye to out of memory. At the same time, the code is stored in the qspi flash to deal with high-intensity vibration

Watch Tutorials:
https://docs.edgeimpulse.com/docs/running-your-impulse-mbed

Functions:
- STM32H7xxII for ARM Cortex M7 processor runs at 480 MHz and has 1MB of RAM and 2MB of flash memory. All I/O pins output 3.3V voltage and can withstand 5V voltage. The processor has the following I/O interfaces: Full-speed USB (12Mbs) interface with computer. When plugged in, your open source camera module will appear as a virtual COM port and USB flash drive
- SD card slot has a read/write capacity of 100Mbs, allowing your open source camera module to record video and easily unplug machine vision assets from a SD card (not included)
- A SPI bus that can run up to 100Mbs allows you to easily stream image data from the system to LCD Shield, WiFi Shield or other microcontrollers
- I2C bus, CAN bus and asynchronous serial bus (TX/RX) are used to interface with other microcontrollers and sensors
- A 12-bit ADC and a 12-bit DAC
- Two I/O pins are used for servo control
- Interrupt and PWM on all I/O pins (there are 10 I/O pins on the board)
- And, one RGB LED and two high-power 850nm infrared LEDs.
- 32MB external 32-bit SDRAM with a clock frequency of 100 MHz and a bandwidth of 400 MB/s
- In 4-bit DDR mode, 32 MB external Quadspi flash memory (16 MB available now) is provided at a frequency of 100 MHz with a bandwidth of 100 MB/s

A movable camera module system that allows to interface with different sensors:
- The PLUS Version has an OV5640 image sensor, capable of shooting 2592x1944 (5MP) images. The simple algorithm will run at 30 FPS or more at QVGA (320x240) resolution and below. Your image sensor comes with a 2.8mm lens on a standard M12 lens mount. If you want to use more professional lenses in the image sensor, you can buy and install them by yourselves.
- For professional machine vision applications, you can purchase our global shutter camera module.
- For thermomechanical vision applications, you can purchase our adapter module for FLIR Lepton.

Applications:
- Frame difference: Detect movement in the scene by looking at the changes. The frame difference allows you to use it for security applications.
- Color tracking: Detect 16 more colors in the image at once (in fact, you never want to find more than 4 colors), and each color can have any number of different spots. Then, tell you the location, size, center of mass and direction of each spot. Using color tracking, you can program to perform tasks such as tracking the sun, tracking a straight line, and target tracking.
- Tag tracking: You can detect color groups instead of individual colors. This allows you to create color makers (2 or more color labels) that can be placed on the object so that you can understand what the marked object is.
- Face Detection: You can (or any general object) detect human faces. It can handle Haar cascade for regular object detection, and comes with built-in Front Face Face Cascade and Eye Haar Cascade to detect faces and eyes.
- Eye tracking: You can use eye tracking together to detect someone's line of sight. Then you can use it to control the robot. Eye tracking can detect where the pupils are, instead of detecting whether there are eyes in the image.
- People detection: You can use our built-in person detector TensorFlow Lite model to detect whether there is a person in the field of view. The video demo is here.
- Light flow: You can use optical flow detection. For example, you can use optical flow on a quadcopter to determine its stability in the air.
- QR code detection/decoding: You can read the QR code in its field of view. With QR code detection/decoding, you can create an intelligent robot that can read tags in the environment.
- Data matrix detection/decoding: It can also detect and decode data matrix two-dimensional barcodes. You can view a video about this feature here.
- Linear barcode decoding: It can also decode one-dimensional linear barcodes. In particular, it can decode EAN2, EAN5, EAN8, UPCE, ISBN10, UPCA, EAN13, ISBN13, I25, DATABAR, DARABAR_EXP, CODABAR, CODE39, CODE93 and CODE128 barcodes.
- April tag: Better than the above QR code, it can also track AprilTags. AprilTags are tags for rotation, scaling, clipping, and lighting unchanged.
- Line detection: Infinite line detection can be completed quickly. And, you can also find line segments of non-infinite length. You can watch our video about this feature here. In addition, we support running linear regression on the image for subsequent applications such as DIY Robot car.
- Circle detection: You can easily detect the circles in the image.
- Rectangle detection: You can also use our AprilTag library's quaternary detector code to detect rectangles.
- Template matching: You can use templates together to detect when to display translated pre-saved images. For example, template matching can be used to find a reference point on the PCB or to read a known number on the display.
- Image collection: You can collect grayscale/RGB565 BMP/JPG/PPM/PGM images. You can directly control how to get images in scripts for Python. The important thing is that you can perform machine vision functions and/or draw on the frame before saving.
- Video recording: You can record more grayscale/RGB565 MJPEG video or GIF image (or RAW video). You can directly control how to record each frame of the video in a Python script, and you can completely control the start and end of the video recording. And, just like capturing an image, you can perform machine vision functions and/or draw the video frame before saving it.
- TensorFlow Lite supports microcontrollers: TensorFlow Lite supports enabling you to run custom image classification and segmentation models. With TensorFlow Lite support, you can easily view the complex areas of interest in the view and control the I/O pins based on what you see. For more information, see TensorFlow module.
- All the above functions can be mixed and matched together with I/O pin control in your own custom application to talk to the real world.

Package Included:
- 1 x Camera Module
- 1 x Data Cable for Android
- 1 x DuPont Cable
- 1 x Storage Box