Milestone 4

The goal of this milestone was to detect targets; their presence or lack thereof, their color, and their shape.

In order to determine whether there was actually a treasure as opposed to a few stray blue or red pixels, we used a threshold value. If the number of pixels of a certain color exceded this threshold, we assumed that there was a treasure and then began the shape detection process. Once done, we send the color and shape to the arduino, which displays this information on the serial monitor. If the threshold is not reached for either blue or red, "NULL" is sent to the ardunio. Below is a demonstration of this code.

Because of the code we used inside DEO_NANO to display color, detecting color was made simpler. Inside of DEO_NANO, whenever we saw a pixel that was above a certain threshold of blue or red, we set that pixel to pure blue or red. In code, we set the pixels equal to 8'b000_000_11 and 8"b111_000_00. We are able to exploit this in our color detection. Firstly, we only detect color if it is within a certain region we have specified. Inside of this region, we examine each pixel and determine whether it is red, blue, or neither. If it is red or blue, we increment a variable that counts the amount of pixels of each color. Finally, if either count is above a specified threshold, we assume that one, we seeing a treasure and two, that it is that color. If this criteria is not met, we display "NULL." An overview of the code used as well as a demonstration are shown below.

    	if(VGA_PIXEL_X>((`SCREEN_WIDTH/2)-30)&& VGA_PIXEL_X<((`SCREEN_WIDTH/2)+30) && VGA_PIXEL_Y<((`SCREEN_HEIGHT/2)+50)&& VGA_PIXEL_Y>((`SCREEN_HEIGHT/2)-50)) begin
		if(PIXEL_IN == BLUE) begin
			countBLUE = countBLUE + 10'd1;
		end
		else if(PIXEL_IN == RED) begin
			countRED = countRED +10'd1;
		end		
	if(VGA_VSYNC_NEG == 1'b1 && lastsync == 1'b0) begin
		if(countRED >= R_CNT_THRESHOLD) begin
			begin red shape detection                                           
		end
		else if(countBLUE >= B_CNT_THRESHOLD) begin 
			begin blue shape detection
		end
		else begin
			display NULL																				
		end
      

We used a unique strategy to detect our shapes by counting pixels in segments of our figure. We first located the first pixel and last pixel in the image with the right color. We use these to create three little sigments where we get the total number of color pixels. We then compare and average them to determine what shape the image has. Below is a code snippet and a video to show the outcome.

         if(VGA_PIXEL_X>((`SCREEN_WIDTH/2)-30)&& VGA_PIXEL_X<((`SCREEN_WIDTH/2)+30) && VGA_PIXEL_Y<((`SCREEN_HEIGHT/2)+50)&& VGA_PIXEL_Y>((`SCREEN_HEIGHT/2)-50)) begin // only focus on a segment of the screen
		
		        if(PIXEL_IN == BLUE) begin     // this is for blue pixel. similar for red
			          countBLUE = countBLUE + 10'd1; // count blue
			          lastBLUE = VGA_PIXEL_Y;        // get last pixel
			          if(countBLUE==10'd1) firstBLUE= VGA_PIXEL_Y; // get first pixel
            end
            if(VGA_PIXEL_Y==firstRED+((lastRED-firstRED)*(1/3)) || VGA_PIXEL_Y ==((lastBLUE-firstBLUE)*(1/3)))begin  // at one third of the segment of whole shape
              blue1 = countBLUE; // take blue pixel count
              red1=countRED;     // or red pixel count
            end
            ... // take two more values for comparison
     	 

			// redi, for i = 1, 2... is a segment pixel count for segment i.
			if(red1<(red2-red1)&& (red2-red1)<(red3-red2)) begin  // analysis of segment pixels to determine shape
				RESULT = 3'b010; end         // triangle
			else if(red1<(red2-red1) && (red3-red2)<(red2-red1)) begin
				RESULT= 3'b001; end          // diamond
			else if((red1+red2)-(red2+red3)<= 10'd40 ||(red3+red2)-(red1+red2)<= 10'd40) RESULT = 3'b011; //Square
			//else RESULT = 3'b111;