This is an answer to a two year old question, but I wanted to answer it anyway.
If you could fill none to all squares on that lattice (that is any combination (any possible binary number)) the answer would have been some number consisting of 31 digits (in decimal), which would have been a very large number.
Since you are using only two colors (green and white), in this case the colors does not matter. i.e. if you only have one color and another background color: you can define (the color) green as '1', and white (background color) as '0'. In computer terminology we know these are called bits (binary numbers). So by this there is another solution to this problem involving some bruteforcing. I know more programming than math, so bear with me because I can't answer with a very detailed mathematical vocabulary. The answer is very simple if you can bruteforce and check if the hamming weight is equal to n. The only problem using this technique would be the speed of the computer-registers that compute this. We are talking about 100 bits here, and I dont think mainstream cpus have 128-bit registers yet and not even an popcount-instruction to compute the hamming weight. However, the algorithm should work despite all of this.
'Hamming weights' aka 'population count' works this way: It counts the number of on-bits in a binary number representation (Was invented (or atleast named after) by a guy called Richard Hamming). You could make a function that bruteforces all the possible combinations of hamming weight = n (where n is the number of filled (green) squares / or on-bits in computer terminology).
It boils down to this: You have a natural number z that always add +1 when the hamming weight equals n, and so the result (all the sums) of z equals the number of possible permutations.
Pseudocode
Initial condition:
n=6 (number of filled squares).
z=0 (some natural number or integer).
Iterative loop:
For i = 0 to (2^100)-1:
If (HammingWeight(i) == n) z = z + 1
Next i
z equals the solution after the loop has finished.